2005 - Acadia05 - Book Smart Homes For Rural

ACADIA05: Smart Architecture 1
Copyright © 2005 by The Association for Computer-Aided

Design in Architecture
All rights reserved by the individual paper authors who are

solely responsible for their content.
Cover design by Osman Ataman.

Graphics design by Nathan Charris
Printed by Martin Publishing, Champaign, IL.
No part of this work covered by the copyright hereon my be

reproduced or used in any form or by any means-graphic,
electronic, or mechanical, including photocopying,
recording, taping, or information storage and retrieval
systems without the prior permission of the copyright owner.
Library of Congress Catalog Card Number: 2005934032

ISBN: 0-9772832-0-8
2 ACADIA05: Smart Architecture

ACADIA: ASSOCIATION FOR COMPUTER-AIDED DESIGN IN ARCHITECTURE
President, Kevin Klinger

Vice-President, Wassim Jabi
Secretary, Carmina Sanchez
Treasurer, Volker Mueller
Membership, Aron Temkin
Newsletter, Mark Clayton
Awards, Douglas Noble
Outreach, Kevin Klinger
Competition, Peter Anders
Steering Committee
Alfredo Andia
Aron Temkin
Brian Johnson
Carmina Sanchez
Huy Sinh Ngo
Julio Bermudez
Osman Ataman
Peter Anders
Philip Beesley
Theodore W. Hall
Volker Mueller
Wassim Jabi
ACADIA’05 Technical Chair, Osman Ataman

ACADIA’05 Site Coordinator, Huy Sinh Ngo

ACADIA 2005 Technical Review Committee
Anders, Peter Johnson, Brian

Mindspace.net University of Washington
ptr@mindspace.net brj@u.washington.edu
Ataman, Osman Johnson, Scott

University of Illinois at Urbana-Champaign University of Michigan
oataman@uiuc.edu sven@umich.edu
Bermudez, Julio Kim, Micheal

University of Utah University of Illinois at Urbana-Champaign
bermudez@arch.utah.edu mkkim1@uiuc.edu
Castillo, Tim Klinger, Kevin

University of New Mexico Ball State University
timc@unm.edu krklinger@bsu.edu
Chan, Chiu-Shui Kolarevic, Branko

Iowa State University University of Pennsylvania
cschan@iastate.edu branko@pobox.upenn.edu
Clayton, Mark Kvan, Thomas

Texas A&M University of Hong Kong
mark-clayton@tamu.edu tkvan@arch.hku.hk
Elvin, George Luhan, Gregory

Ball State University University of Kentucky
elvin@uiuc.edu galuhan@uky.edu
Fowler, Tom Martens, Bob

Cal Poly Vienna University of Technology
tfowler@calpoly.edu b.martens@tuwien.ac.at
Guzowski, Mary McCullough, Malcolm

University of Minnesota University of Michigan
guzow001@umn.edu mmmc@umich.edu
Harfmann, Anton Muller, Voelker

University of Cincinnati NBBJ
anton.harfmann@uc.edu vmueller@nbbj.com
Noble, Doug
Huy, Ngo University of Southern California
Savannah College of Art and Design dnoble@usc.edu
hngo@scad.edu
Sanchez, Carmina
Jabi, Wassim Hampton University
New Jersey Institute of Technology carmina.sanchez@hamptonu.edu
wassim.jabi@njit.edu

ACADIA 2005 Technical Review Committee
Temkin, Aron
Florida Atlantic University
atemkin@fau.edu
Vasquez de Velasco,
Guillermo Texas A&M
vasquez@archone.tamu.edu

Table of Contents
ACADIA’05 Officers and Anne Marie Due Schmidt

Steering Committee 3 Navigating Towards Digital Tectonic Tools 114
ACADIA’05 Tecnical Review Committee 4 Carmina Sanchez-del-Valle

Adaptive Kinetic Architecture:
Osman Ataman A Portal To Digital Prototyping 128
Editor’s Preface 8
Chapter 1: Main Framework and Theoretical Chapter 2: Smart Systems, Techniques and
Arguments Environments
Sachin Anshuman Jaewook Lee, Yehuda E. Kalay

Responsiveness and Social Expression: Collaborative Design Approach to
Seeking Human Embodiment in Intelligent Environments 142
Intelligent Façades 12
Mark Cabrinha
Anijo Mathew From Bézier to NURBS: Integrating
Smart Homes for the Rural Population: Material and Digital Techniques through
New Challenges and Opportunities 24 a Plywood Shell 156
Mahesh Senagala Daniel Barker, Andy Dong

Building is a Network for Living in: A Representation Language for
Toward New Architectures 36 a Prototype CAD Tool for
Intelligent Rooms 170
George Katodrytis
Poiesis(*) and Autopoiesis in Architecture 48 Matthew G. Fineout
The Tower of Babel:
José Daveiga, Paulo Ferreira Bridging Diverse Languages with
Smart and Nano Materials in Architecture 58 Information Technologies 184
Michael Silver Thomas Modeen, Christine Pasquire,

Discrete Space: Automason Ver. 1.0 68 Rupert Soar
Design Ground - An Iconic Tactile Surface 192
Michael Fox, Catherine Hu
Starting From The Micro: Tsou Jin Yeu, Chan Yi Lee, Mak Kwok Pui, Ru
A Pedagogical Approach to Xu Du, Liang Jian, Yeung Kim
Designing Interactive Architecture 78 Applying Scientific Simulation to Integrate
Thermoelectric Conductor Module into
Mark J. Clayton Architectural Design – Smart Wall for
How I Stopped Worrying and Thermal Comfort 200
Learned to Love AutoCAD 94
Peter Anders, Werner Lonsing

AmbiViewer: A Tool for Creating
Architectural Mixed Reality 104

Chapter 3: Applications
Gregory A. Luhan
Modern Translations, Contemporary
Methods: DL-1_Resonance House® 212
Nimish Biloria, Kas Oosterhuis, Cas Aalbers

Design Informatics 226
Jordan Brandt
Skin That Fits:
Designing and Constructing Cladding
Systems with As-built Structural Data 236
Madalina Wierzbicki-Neagu
Unfolding Architecture:
Study, Development and Application of
New Kinetic Structure Topologies 246
Bob Sheil, Chris Leung

‘Kielder Probes’ – Bespoke Tools for
an Indeterminate Design Process 254

Editor’s Preface general; digital and building technologies in
particular. By doing that, it also provides a
structure for discussing some issues with
Smart Architecture developing and integrating digital technologies
into architecture that can evolve to better meet the
opportunities and challenges of the future.
Osman Ataman
University of Illinois at Urbana-Champaign Consequently, as the technical chair of the
conference, I invited papers that may suggest some
answers to the theoretical as well as practical issues
Recent developments in digital technologies and addressing this subject. These papers leave many
smart materials have created new opportunities questions unanswered, and many also unasked.
and are suggesting significant changes in the way But, they do indicate the diverse demands and
we design and build architecture. Traditionally, implications to be studied. This book examines a
however, there has always been a gap between the sample of recent work in the “Smart Architecture”
new technologies and their applications into other field to help determine what kind of problems,
areas. Even though, most technological issues and paradigms are emerging, how and why.
innovations hold the promise to transform the There are three major chapters in this volume: (1)
building industry and the architecture within, and Main Framework and Theoretical arguments; (2)
although, there have been some limited attempts Smart Systems, Techniques and Environments;
in this area recently; to date architecture has failed and (3) Applications.
to utilize the vast amount of accumulated
technological knowledge and innovations to In the first chapter, we see quite a few surveys
significantly transform the industry. Consequently, and pedagogical arguments about Smart
the applications of new technologies to Architecture. In his paper, “Responsiveness and
architecture remain remote and inadequate. Social Expression,” Anshuman surveys twenty-
Although, there have been some adaptations in this six smart building facades and sixteen large media-
area recently, the improvements in architecture facades to demonstrate how architects have used
reflect only incremental progress, not the the technology to turn building surfaces into
significant discoveries needed to transform the socially engaging architectural elements and at the
industry. end he proposes a conceptual model as a possible
method to address the emergent issues. In another
However, architectural innovations and paper, “Smart Homes for the Rural Population:
movements have often been generated by the New Challenges and Opportunities,” Mathew
advances of building technologies, such as the surveys and examines the challenges and
impact of steel in the last and reinforced concrete opportunities in the design of smart technologies
in this century. This relationship –between new for preventive healthcare in rural homes. He later
technologies and ‘new architecture’ is very summarizes findings from current ethnographic
significant and has always played a significant role and demographic studies; and examines other
in architectural field so that architecture in modern research in the field of ubiquitous computing and
times is characterized by its capacity to take smart homes. Later, Senagala looks at the big
advantage of the scientific developments and picture in “Building is a Network for Living in”
technological innovations (Morales 1997). and outlines a series of recent developments in
digital technologies that would enable architecture
It is precisely the main goal of this conference to to become sensate, supple and globally networked.
address this situation by focusing on how digital Along the same line, in his paper, “Poiesis and
and building technologies can be integrated with Autopoiesis in Architecture,” Katodrytis discusses
architecture to satisfy the emerging and future the scope and depth of technological advancement
needs of people. This conference tried to focus on to architecture and introduces some of his
all levels of architecture and different dimensions hypothetical, unbuilt projects as his poetic
of design and the ways they relate to digital response. Finally, in their collaborative paper,
technology and buildings. It is my hope that this “Smart and Nano Materials in Architecture,”
conference provides an opportunity to establish a Daveiga and Ferreira describe and analyze the
link between architecture and technology in fields of Smart and Nano Materials and their

potential impact on architectural design and architecture, generated from an examination of
building fabrication. research in engineering and architecture. Her
characterization introduces the challenges
Later, three papers in the first section focus on the presented both by modeling form and
pedagogical issues. In his paper, “Discrete Space,” environment, and simulating their interaction.
Silver attempts to demonstrate the power of digital Finally, she surveys digital prototyping software
technology by encouraging architects to become for engineering applications that can be transferred
more involved in the creation of small scale, ad- to architecture, and identifies some of the
hoc and ask specific software tools. Silver argues unresolved issues.
that rather than appropriating code originally
designed to solve the visualization problems faced Papers in the second section offer interesting
in other fields, designers should instead develop viewpoints on Smart Systems, Techniques and
their own programming culture. Fox and Hu Environments. In “Collaborating Design
outline a pedagogical approach in their paper, Approach to Intelligent Environments”, Lee and
“Starting From the Micro,” whereby a number of Kalay propose an approach on dynamic,
technology-intensive skills can be quickly learned collaborative design by viewing the built
to a level of useful practicality through a series of environment. Their approach views the
discrete, yet cumulative explorations with the environment as comprised of multiple independent
design goal of creating intelligently responsive object-agents, each of which is responsible for one
architectural systems. At the end, the authors small aspect of the environment. In “From Bezier
summarize certain processes and tools that we can to NURBS”, Cabrinha introduces his research on
utilize in creating and demonstrating of such developing digital and material techniques toward
systems and the implications of adopting a more intelligent construction based on the
active role in directing the development of this correspondence between digitally driven surface
new area of design. At the end of this sub-section, and digitally driven material processes. Later, in,
from an interesting viewpoint, in his paper, “How “A Representation Language for a Prototype CAD
I Stopped Worrying and Learned to Love Tool for Intelligent Rooms,” Barker and Dong
AutoCAD,” Clayton looks at the history of present a Hardware as Agents Description
computing as smart design tools and analyzes the Language for Intelligent Rooms (HADLIR) to
pedagogy behind them. model hardware in an intelligent room as
“hardware agents” having sensor and/or effector
In the final sub-section of the first chapter, three modalities with rules and goals. Their approach
papers mostly discuss Smart Systems and the illustrates how the HADLIR representation assists
relationship between tool and technology. In their in the design, simulation and implementation of
paper, Anders and Lonsing present a new mixed an intelligent room and provides a foundation
reality system for architecture: Ambiviewer. The technology for design tools for the creation of
system employs digital video, onboard modeler intelligent rooms.
and global positioning to merge physical and
simulated entities on the screen. The authors In the second part of this section, Fineout examines
discuss the use of this system in creating cybrids, the integration of building project and identify
compositions of virtual and material reality. Along opportunities for the deployment of information
the same line, in her paper, “Navigating Towards technologies to bridge boundaries, ultimately
Digital Tectonic Tools”, Schmidt investigates what providing for and delivering architectural projects.
a digital tectonic tool could be and what Fineout attempts to demonstrate how information
relationship with technology it should represent. technologies can be implemented within the
She argues that an understanding of this multifaceted framework of conventional building
relationship can help us not only to understand projects to yield a project of unprecedented form.
the conflicts in architecture and the building Later, in their paper, “Design Ground,” Modeen
industry but also bring us further into a discussion et al. summarize a project which aims to suggest
of how architecture can use digital tools. The final an alternate methodology for utilizing additive
paper of this section belongs to Sanchez-del-Valle. Rapid Manufacturing for the conceptualization
In her successful paper, “Adaptive Kinetic and fabrication of design and architecture. At the
Architecture,” Sanchez-del-Valle presents a end of this section, in their paper,
definition of adaptive kinetic structures in “ApplyingConductor Module into Architectural

Design,” Tsou et al. present their architectural Acknowledgments
design concept which is to integrate the new
material and technology into the building design I should like to express my gratitude to all those
to achieve the thermal comfort and at the same who have contributed in some way to the ACADIA
time reduce the energy consumption of the 2005 conference on Smart Architecture, and to the
building by making use of renewable energy. production of the present proceedings volume.
First, I would like to thank my site coordinator
In the last section of this book, various applications Huy Ngo for his dedicated work and tremendous
in various fields are reported. In his paper, efforts to make this conference well-organized,
“Modern Translations, Contemporary Methods,” pleasant and painless. Second, I would like thank
Luhan reports the procedure and the results of this all the authors that responded with their work to
Design-built studio. Luhan employs several smart the Call for Participation. Without their effort and
systems for design and application and trust, this conference would not have been
consequently argues that the resulting effect has possible. I also would like to acknowledge the
been a series of virtual studies with real-world technical review committee for their committed,
applications and an increased role for students in professional and hard work, particularly because
shaping a new reality for practice through most of them reviewed at least 7 papers and
advances in technology. In another paper, Biloria provided long thoughtful feedback to the authors.
et al. report a design technique developed at ONL Special recognition goes to Wassim Jabi for his
(Oosterhuis and Lenard), Netherlands, specifically technical contribution and consistently long
appropriated for envisaging complex geometrical phone-conversations throughout submission
forms. The “informed design technique,” they period. I also appreciate ACADIA Steering
argue, aart from being highly instrumental in Committee’s support and encouragement during
conceptualizing and generating the geometric this past year of work. The desktop publishing
component constituting architectural form in a work for this book done by graduate student Ajla
parametric manner, is also efficiently utilized for Zisko also deserves a special acknowledgment. I
precise computer-aided manufacturing and also would like extend my gratitude to several
construction of the speculated form. UIUC and SCAD students that helped us in diverse
ways before, during and after the conference.
In the second part of this section, Brandt provides Finally, I would like to thank ACADIA for giving
a model for implementing a new non-mechanical me the opportunity to contribute to advancement
technology to facilitate a real-time parametric architectural knowledge in general, Smart
Building Information Modeling (BIM) to support Architecture as an emerging field in particular.
the fabrication, erection and quality control of
cladding envelopes. Finally, there are two short
papers in the Work-in-Progress category. First Osman Ataman, Ph.D.
paper in this category, “Unfolding Architecture,” ACADIA ’05 Technical Chair
Wierzbicki-Neagu introduces foldable systems
that can be deployed off dedicated, motorized
platform. In general, the author argues,
configurable structures will provide lesser
environmental impact than traditional technologies
as they are better suited for re-using, modifications
and re-location. In the last paper of this category,
“Kielder Probes,” Sheil and Leung report their
ongoing work with smart systems and production
tools.

Chapter 1
Main Framework and Theoretical Arguments
Sachin Anshuman
Responsiveness and Social Expression; Seeking Human Embodiment in Intelligent Façades
Anijo Mathew
Smart Homes for the Rural Population: New Challenges and Opportunities
Mahesh Senagala
Building is a Network for Living in: Toward New Architectures
George Katodrytis
Poiesis(*) and Autopoiesis in Architecture
José Daveiga, Paulo Ferreira

Smart and Nano Materials in Architecture
Michael Silver
Discrete Space: Automason Ver. 1.0
Michael Fox, Catherine Hu

Starting From The Micro: A Pedagogical Approach to Designing Interactive Architecture
Mark J. Clayton
How I Stopped Worrying and Learned to Love AutoCAD
Peter Anders, Werner Lonsing

AmbiViewer: A Tool for Creating Architectural Mixed Reality
Anne Marie Due Schmidt

Navigating Towards Digital Tectonic Tools
Carmina Sanchez-del-Valle
Adaptive Kinetic Architecture: A Portal To Digital Prototyping

S. Anshuman / Responsiveness and Social Expression
Responsiveness and Social Expression; Seeking

Human Embodiment in Intelligent Façades
Sachin Anshuman1
1
Glasgow Caledonian University
Abstract
This paper is based on a comparative analysis of some twenty-six intelligent building facades and sixteen
large media-facades from a socio-psychological perspective. It is not difficult to observe how deployment of
computational technologies have engendered new possibilities for architectural production to which surface-
centeredness lies at that heart of spatial production during design, fabrication and envelope automation
processes. While surfaces play a critical role in contemporary social production (information display,
communication and interaction), it is important to understand how the relationships between augmented building
surfaces and its subjects unfold. We target double-skin automated facades as a distinct field within building-
services and automation industry, and discuss how the developments within this area are over-occupied with
seamless climate control and energy efficiency themes, resulting into socially inert mechanical membranes.
Our thesis is that at the core of the development of automated façade lies the industrial automation attitude
that renders the eventual product socially less engaging and machinic. We illustrate examples of interactive
media-façades to demonstrate how architects and interaction designers have used similar technology to turn
building surfaces into socially engaging architectural elements. We seek opportunities to extend performative
aspects of otherwise function driven double-skin façades for public expression, informal social engagement
and context embodiment. Towards the end of the paper, we propose a conceptual model as a possible method
to address the emergent issues.
Through this paper we intend to bring forth emergent concerns to designing building membrane where
technology and performance are addressed through a broader cultural position, establishing a continual dialogue
between the surface, function and its larger human context.

Introduction Augmenting such space would require

understandings of embodied existence, socio-
Since sometime, relationships between the human psychological aspects of space and its operative
subject and the built environment have been principles.
central in architectural discourse again. After
decades of the machine’s influence through the Intelligent Building
industrial revolution until wake of the digital era,
we have come to focus again on the importance Concepts of intelligent building and smart space,
of humane space, urban responsibilities of built however, address technologies for built spaces
structures and its relationships to the end user. with a distinct approach that is often utilitarian.
While it is no news that built environments have Energy efficiency continues to be a top priority in
lasting effects on human psychology, shaping of what constitutes an intelligent building [Szell
collective identities and social behaviour 2003], [Sharples 1999]. Early approaches
[Norberg-Schulz 1971], [Rapoport 1969], the essentially focused on two basic objectives -
subject is re-visited recently through studies in human comfort through climate control, and cost-
spatial psychology [Vidler 2000], environmental effective energy management. Although there
cognition and architectural theories in accordance exists no fixed opinion or definition of the term
with emergent digital revolution [Zellner 1999], “Intelligent Building”, it is important to note how
[Oosterhuis 2002], [McCullough 2004]. Building even the earliest definitions address the notion of
surfaces or façades form an important area of responsiveness through electronic system controls,
investigation for how they represent an initial time based programmability of building services
interface to our interaction with built entities and and integrated operation [Wigginton 2002]. Unlike
urban environments in general. Notion of response the use of digital technologies in architectural
and media augmented expression are increasingly design, which focuses on augmenting human
addressed in contemporary architectural design experience through creative technological means,
through incorporating new digital techniques in monotony of energy efficiency theme remains
an attempt to equip architectural surfaces with new central in design and production of intelligent
forms of expression [Rahim 2002], [Krueger building façades. This is apparent in Intelligent
1996]. Studies in consumer product psychology Building definitions, which define the Intelligent
repetitively suggest that the value of products are Building as the one that maximises efficiency
often determined against the expression they while allowing effective management with
release for how they signify abstract persona and minimum costs [IBC, Intelligent Building
allow us to project ourselves through their International], [Kim 1993], anticipates conditions
ownership [Reeves 1996][Norman 1993]. and active forces [Kroner 1997], optimizes
Symbolic exchange-value of the commodity structure, systems, services and management to
positions the possessor within a class where the help business owners, property managers and
consensual meaning associated with the objects occupants to realise their goals in cost, comfort,
finds manifestation. What the object signifies (its convenience, safety, long term flexibility and
sign value) often is more important in social terms marketability [Kim 1996], or be able to respond
than what the object does (its utility value), and to organisational change and adapt to new tasks
therefore, it is not only the product that we [Harrison 1994]. Basic objectives of IB systems
consume, but the idea of the product, and what can be divided in four categories, namely
the product will allow us to become. Places are Increased environmental comfort, Energy
defined similarly - less by location, building optimisation, Security, and Automated building
attributes, technical efficiency or landscape, and maintenance procedures. Façade engineering
communities; but by the focusing of collective plays a major role in implementing these
experiences and subjective embodiment into objectives. However for architecture, in its long
particular spatial conditions [Dourish 2001]. history, façade was a social frame first and became
Architecture is a notion of space inextricably operable equipment (automated envelope) only
linked with, and defined by, the wealth of human later [McCullough 2004]; unlike developments in
experience and use occurring within it. Invested computing – an operation based system first;
by subjective values, philosophical projections, whose developmental trajectory is lately being
and cultural influences, architectural space is questioned altogether for being an orthodox
defined far beyond the physical envelope. extension of industrial automation slogan that

focused around process automation and not the matters and escalating social consequences, the
human subject. With recent explosion of path towards such connection involves a shift from
ubiquitous computing and the shifting focus of foreground objects to background experiences.
interface design to what is now termed interaction [McCullough 2004] It is in this underlying attitude
design denote popular reaction to a general that these projects differ significantly from
realisation to move away from idiot-proof, button- conventional function driven intelligent façades.
based interfaces that perpetuate such industrial
legacy and rethink the idea of interaction from a Most automated façade systems invite little bodily
deeper perceptual and social level. Recent interest participation both from the occupants’ perspective
of HCI in Heidegger’s phenomenology and its within the building and larger urban participation
application to the design of computational system from the outside. Sense of belonging in spatial
architecture denote such shift [Dourish settings involves a perception of insideness and
2001],[Coyne 1999]. Façades designed and not the empirical attributes of physical construct.
manufactured with industrial automation outlook Foundations of embodiment are rooted in such
significantly lack sociability, expression and thus perception, which is largely governed by trust,
engagement alike, for how the human relationships support networks, collective experiences, interests,
with the building, its elements and the social types of transactions, history, landmark, proximity
surrounding of the building are considered less etc. [McCullough 2004] Allocation of interaction
important from the start. This reflects not only in systems acknowledging such interventions lack
the assigned functions and aesthetic qualities of significantly in the design of automated facades.
façade, but also in the interfaces developed to For example, bay wise louver control for optimal
control them, which form primary means of lighting in the systems we studied usually involved
operation and exercising offered affordance. the user participation in two forms. One, through
Systems that eventually develop around the provision of a push-button based system-override
energy/cost/climate parameters hence remain panel and two, in some cases, through pyro-electric
limited in terms of how human subject identifies infrared occupancy sensors mounted at room-level
with them. resolution to interpolate the control of artificial
lighting. The former denies physical contact and
How Intelligent Facades are Socially Inert? hence reduces the spectrum of control through
limited push buttons to control a few aspects of
In recent years, interaction designers and architects what are otherwise elaborate architectural
have demonstrated potential in physical computing elements, and the latter comes through as overtly
and embedded systems in large interactive impersonal and dissociative for its popular
surfaces. As the need to connect architecture and psychological association with voyeur and silent-
interaction design comes from overlapping subject surveillance. Such automated provisions do not
Figure 1. BIX communicative display skin embodies 930 fluorescent lamps integrated in acrylic glass façade of the biomorphic
structure

efficiently replace the complex ritual of interacting Benedikt, we orient spatially not only through eyes
with a window, for example, for view and/or light but also with tactile body [Benedikt 1992];
and/or breeze but rather reduce the scope of the repetitive standardised curtain-wall elements
architectural elements to singular functions – confiscate unique identification of places and
which in this case is, seamless maintenance of locations along the surface and render the
internal illumination level. spectator-body, in turn, as the object of operation
rather than subject of action [Bloomer 1977].
While many intelligent facades are now equipped Perception of locations and places are rooted in
with sophisticated Artificial Neural Network spatial relationship networks, which are at the
(ANN) based control systems, to learn occupant heart of our conceptual system [Lakoff 1980].
activity patterns and climatic variations, perceived
sense of response (and so façade as an active In case of operable components encased between
entity) remain low to the occupants. Most louver two transparent sheet-glass layers of double-skin
control systems, for example, set re-adjustment façades, or automated doors, air-inlets, shading
timings to 30mins or more to avoid continuous elements etc, invoke tactile detachment. Such
movement and thereby reducing disturbance. detachment from tactile contact, and therefore
Studies in social psychology [Brown 1979], from direct control, eliminates a variety of
[Kolers 1972] show that lack of visible motion activities related to traditional building elements
results in perceived inertness for how cognition emerging from physical contact [leaning on a
fails to assess activeness of an entity [Sternberg surface, peeping through a hole, sitting on
1986, 1981], [Livingstone 1988], [Bolivar 1995]. window-sill, opening/closing doors/windows etc].
Therefore even though actual intelligence/ Lack of social engagement, personal projection
responsivity is significant, perceived intelligence/ and tactile contact, seemingly forbids personal
response remain low. We will later see how in affiliation with these systems. Such humane
interactive media-façades, this quality is interaction with conventional building-elements
significant and there fore results in increased social forms the bases for embodiment in traditional built
participation. Collective motion of façade environments, which remain low in mechanized
elements, closely tied up with environmental intelligence driven building surfaces.
sensors and weather stations, does not respond to
the user except for local override through wall While it may seem that machinic appearance of
mount control panels. automated surfaces may account for alienation too,
there is evidence in psychological literature that
It is well known that industrial standardisation this condition does not quite prevent us from
processes exert unity at a mass level by extracting identifying with technological objects. Popular
individual identities [Jencks 1998]. One of the examples of such theory come from research
well-criticised downsides of modernist showing how we assign personalities to
architecture has been standardisation. This technological objects such as televisions, mobile
continues to reflect, for obvious reasons, in phones and other new media objects [Reeves
contemporary construction practices too. 1996]. It is not in the way automated façades look
Personalisation of space in traditional spatial but what they signify (through the level of our
settings came through authorship, display and engagement with them) is important in social
control of personal objects, attributes, and their terms. It is at this level that the design of operable
organisation in space. (What makes something my façades lack affordance.
place?) In modern settings these practices are
reduced to personal notes, pictures, greeting cards While primary agendas of Intelligent building
etc stuck on interior partition surfaces, PC- membranes remain rooted in climatic adaptation,
wallpapers, demographies of personal objects on security and cost-effective energy optimization,
work desks and other surfaces etc. Standardisation we notice very little significance of these self-
of surface-elements both within façades, and in contained surfaces socially in urban settings. This
other interior surfaces, with their authorships unresponsiveness is especially noticeable when
assigned to automated systems, reduces the compared to highly expressive surfaces of
potential of personal projection, authorship and interactive media-facades.
interaction with these elements. Moreover, to

Response and Interaction in Media Façades display skin draw on these specificities of
perception and invite engagement, despite low
BIX façade at Kunsthaus Graz, for example, deals resolution. Less definitive abstract movements
with the issue of communication and embodiment resembling to those of real-world objects and
through the deployment of low-resolution display people thus characterize these surfaces and invite
technologies enabling the building surface to engagement.
display video-like images through controlling
brightness of the individual fluorescent tubes. The Approaches to interaction in media façades vary
issue of embodiment is addressed through devising significantly. Hyposurface (Fig 2) responds to
an additional web based service allowing artists ambient sounds and user position/motion by
to compose and post video impressions to the undergoing topological transformations. Kinetic
surface, online; expanding this way, the wall (Fig 3) similarly engages passer by on the
interactivity of the Kunsthaus beyond geographic street by collective motion of Whisks like robotic
proximities. At a local level, the institution projects façade members reacting in real time to human
itself through soft yet dynamic expression released movement. Instant physical response, however
by the built form. In terms of image resolution, abstract, induces a sense of live presence and
the BIX façade (930 pixels) is smaller than the attention, making the façade an active part of urban
screen of a common mobile phone. Such low life. Although majority of these installations serve
resolution stretched across complex doubly curved no utilitarian function except for inviting active
structure pushes the image content to being highly participation, their appropriateness in social
suggestive and abstract. Although most media context is significant. Moreover, although the level
façades involve such low to medium image of computational intelligence or logic-complexity
resolution, success of these schemes lies in the in of such systems is low [Anshuman 2003];
the way foveal and peripheral vision works in perceived intelligence and responsiveness remains
cognitive processes. Studies in image size, significantly high.
resolution and motion [Reeves 1993], [Nystrom
1992], [Detenber 1999] in television-screens and It is only recently that surface has gained a new
cinema suggest that level of arousal and thereby focus in design as a ground for display,
involvement is directly proportional to image size communication and interaction that now requires
but does not depend on resolution [Hochberg re-appropriation of otherwise distinct technologies
1986]. Images that extend to peripheral boundary that form integral part of the surface formation. A
are more involving. Also, eyes are more sensitive more integrative outlook to surface design and
to motion in peripheral vision than the definitive integration of digital or electronic technologies at
attributes of the image [Nystrom 1992], [Hatada material level is now beginning to be demonstrated
1986]. Façades such as the BIX communicative in façade systems. While conventional intelligent
Figure 2. Hyposurface is pneumatically driven animate membrane developed by dECOi

façades demonstrate low embodiment from and electricity, while remaining distinctly
occupants’ perspective, media façades lack real unrelated to neither spatial configuration nor
function. There are clear analogies to be drawn surface design. The notion that the distribution
from media-façades that could inform the design services should be architecturally unrelated,
conventions of intelligent façades. usually reflects modernist agendas of industrial
optimization through segregation, and modularity,
Discussion as much as the often-stated practical assertions of
convenience and ease of construction [Kennedy
Service Segregation and Architectural Surface 2004]. Persistence of this logic has produced a set
of ideas and conventions that continue to influence
We think that the reasons for the automated façade contemporary practice that include façade
engineering to have developed solely around engineering as part of intelligent building
energy efficiency and climate control parameters approaches. This segregation resulted into
are rooted in the historical kinship of this field to conceptual and physical segregation of service
building services and not architectural design. provision infrastructures and their material
Modernist agendas of seeking efficiency through mediums (i.e. lighting fixtures, wall mount
service segregation and efficient function delivery switches, HAVC infrastructure being distinct from
negate the higher-level design issues concerning spatial design). With the proliferation of
embodiment, expression or situatedness in social information technologies, these systems were
contexts. This holds true from the times when further divided into four operating areas namely,
electrical or other technologies were beginning to Energy efficiency, Life-safety systems,
be integrated in architectural production. Within telecommunications systems, and workplace
the first decade of the twentieth century, methods automation [the National Academy of Sciences’
for delivering electrical distribution within committee for “electronically-enhanced”
buildings were developed at an industrial scale. buildings], which over time merged into two
With increasing demand, electricity soon became broader areas: Facilities Management (FM)
integrated into spatial production without formal (energy and security systems) and Information
intrusion as a part of architectural design. Cavity Systems (ITC)(telecommunications and
walls later emerged as a new component type that workplace automation systems), forming two
radically restructured the conventional materiality primary cores of contemporary building services’
of architectural surface, and quickly became the architecture [Szell 2003] (Figure 4).
site for integrating unified building service
infrastructures such as plumbing, heating, cooling
Figure 3. Kinetic wall is prototype surface developed by Kinetic Design Group at MIT

This division is especially evident in the intelligent interpersonal and building’s interaction with
façade examples from the last decade. In general, occupants and appliances. With the spread of
FM deals with the physical structure itself and how context-aware computing [Weiser 91] new
it is operated. It finds manifestation through façade channels of connecting human activities to
elements and floor-concealed HAVC, security and information are now opening with various sensing,
fire systems. Various IB standards (appendix) have processing and networking technologies. These
since emerged to automate FM (Fig 7). ITC on developments have demonstrated potential for
the other hand, refers to the way information is architectural spaces to perform in previously
handled within the building - it operates unimaginable manner, that is, through augmented
independently from façade, structure or intrinsic experience, expression and interaction [Gage
building elements. ITC systems dealing with web, 1998], [Krueger 1996]. Unfortunately, material
email, LAN, WAN networks, workstations, IO manifestation of current ITC systems in IB systems
peripherals, telephone systems, and in some cases (desktop PCs, telephones, printers, video-
personal ubiquitous devices, is manifested only conferencing facilities etc.) form a detached inner-
through object-level entities distributed within layer consisting of distinct objects and services
building envelopes. which are not commonly perceived as parts of
building anatomy. FM, through automation,
An Alternative Model continues to detach definitive formal elements of
built entities from its occupants; to which, façade
What strikes one are the underlying organizing and other surfaces are operative grounds. For it
principles in the current IB systems model. While is these material surfaces, that form essential
FM continues to decrease human intervention mediums for occupant’s social embodiment
through typical strategies for programmed start/ through interaction. Optimizing built surfaces for
stop, duty cycling, set-point resets, sensor based expression, display and interaction holds the
adaptive control and optimal energy routing for potential to produce a heightened presence
vital building functions, ITC continues to increase characterized by the intensity of experiential
Figure 4. Primary model of contemporary building services’ architecture
Figure 5. Diagram showing the current model of IB with unrelated FM and ITC (Left), and the proposed model (Right) for
increased user embodiment

responsiveness beyond mechanical automation, functionally and socially, require an outlook

and therefore embodiedness within built towards design that in fact increases the affordance
structures. The alternative model (Figure 5) takes offered by such automated elements, rather than
integrated approach to environmental handling reducing its participative spectrum. We have
together with occupant information thought ITC demonstrated this limitation by comparing the
system instead of FM and ITC performing button based louver control with the manual
independently. ITC could therefore be optimized control of conventional window and the ease of
for it’s newly found potential to integrate objects, operation and variety of engagement afforded by
occupants and environmental inputs, and organize such element. We propose for widening such
heterogeneous output at a wider material level. engagement spectrum through better interaction
This approach most importantly allows for input design where technology does not reduce human
from human actions (explicit keyed instructions dimension otherwise inherent in an entity.
and implicit sensor data) and occupant attributes Although it was out of the scope of this paper to
that otherwise remain limited to interconnected discuss technologies driving such systems, it was
ITC components. As ubiquitous computing mark observed during the study that both media-façades
the third wave in computing, interaction between and double skin membranes use very similar
computational objects and mediated interpersonal technologies to achieve their respective goals.
information exchange will become more human While expression, experience and user
centred and context aware. Routing building participation form primary agendas for media
management procedures via ITC systems this way façades, energy efficiency and seamless control
allows for the response actions to be linked to continue to predominate developments in
architectural surfaces and built objects; making automated double skin membranes. Lack of
them expressive or responsive to their human expression and interaction preclude these rather
contexts. The increasing number of materials machinic systems from being sociable and
through with digital technology can be distributed engaging. We have also argued that the reasons
and information could be displayed for the lack of such objectives in their design are
(Electroluminescent materials, FoLED, SMA and rooted in the relatedness of this field to the building
piezoelectric materials) expands the array of services that throughout history operated alongside
integrating such materials into spatial production; architectural design but were never an active
extending this way, responsive qualities both via component of architecture. While the growth of
robotic and audio-visual means. ubiquitous computing now raises new issues to
infuse computational properties and media in
Conclusions everyday objects and surfaces, this continuum
pushes the building industry to re-organize and
In this paper, we have tried to demonstrate two adapt to new integrated processes of material
distinct approaches to façade design. Keeping design and their spatial manifestation. Architecture
automated double skin façade in focus, we have must then operate to site these properties within
illustrated variety of media installations developed the discipline and argue for the value of their
by architects, interaction designers and media projective effects that could create new concepts
artists. In such comparative perspective, it is not for energy distribution, spatial experiences, and
hard to notice that public expression and expression; giving the discipline a more central
interactivity of large media façades result in role among other contemporary modes of cultural
heightened social participation and informal production. Current form of ITC systems holds
engagement that differ significantly from video- this potential, and must therefore be optimized to
screen and large LED displays conventionally used act as a central core connecting future
in urban settings. Real time responsiveness of infrastructural, spatial and human components. For
these systems to their user and/or larger social it is in doing so, that the architect would be able
contexts, provide these buildings with an added to engage processes of technical and imaginative
social dimension that usually lacks in automated engineering required for adapting and projecting
façades. We have proposed that the reasons for spatial concepts for the material character of a
such inertness are rooted in the manner in which technology that has remained historically outside
such systems are conceptualised and the of architecture, but is now central to our culture.
foundations that they are based on. Control and
engagement afforded to the end user, both

Appendices
Table 1: User interaction and climatic input in intelligent façades

Table 2: Existing Intelligent Building Standards and Attributes

References
Hochberg, J.E. (1986). Representation of Motion
Anshuman, S., & Kumar, B. (2003). A Framework and Space in Video and Cinematic Displays. In
for Designing the Responsive Skin. In Ciftcioglu, K.R. Boff, L. Kaufman, & J.P. Thomas (Eds.),
O., & Dado, E. (eds.) Intelligent Computing in Handbook of perception and human performance
Engineering. Delft Netherlands: Foundation for (Chapter 22), New York, NY: Wiley
Design Research SOON.
Jenks, C. (1995). Visual Culture. New York, NY:
Benedikt, M. (1992). Cyberspace, First Steps. Routledge.
Cambridge, MA: MIT Press.
Kim, Jong Jin. (1996). Intelligent building
Bloomer, K.C. & Moore, C.W., (1977). Body technologies; a case of Japanese Buildings. The
Memory and Architecture. London: Yale journal of architecture, Vol.1, No.2, 148-164.
University Press.
Kim, Jong Jin. & Jones, J. (1993), A Conceptual
Bolivar, V.J., & Barresi, J. (1995). Giving Meaning Framework for Dynamic Control of Daylighting
to Movement; A Developmental Study. Ecological and Electric Lighting Systems. IEEE Industry
Psychology, 7(2), 71-97. Applicantions Society, Ontario, 2-8.
Brown, E.L., & Defenbacher, K. (1979). Kolers, P.A. (1972). Aspects of Motion Perception.
Perception and the Senses. New York: Oxford Elmsford, NY: Pergamon Press.
University Press.
Kroner, Walter M. (1997). An Intelligent and
Colomina, B. (1992). Sexuality and Space; Responsive Architecture. Automation in
Princeton Papers on Architecture. MA: Princeton Construction(200).
Architectural Press.
Krueger, T. (1996). Like a second skin, living
Coyne, R. (1999). Technoromanticism, Digital machines. Architectural Design(66), 9-10.
Narrative, Holism and the Romance of the Real.
Cambridge, MA: MIT Press. Lakoff, G. & Johnson, M. (1980). Metaphors We
Live By. IL: University of Chicago Press.
Detenber, B., & Reeves, B. (1999). Motion and
Image Size Effects on Viewer Responses to Livingstone, M., & Hubel, D. (1988). Segregation
Pictures; Application of a Bio-Information Theory of Form, Colour, Movement, and Depth; Anatomy,
of Emotion. Journal of Communication(1), 23-34. physiology, and Movement, Science(240), 740-
749.
Dourish, P. (2001). Where the Action is.
Cambridge MA: MIT Press124-125 Meyrowitz, J. (1985). No Sense of Place; The
Impact of Electronic Media on Social Behaviour.
Gage, S. (1998). Intelligent interactive New York, NY: Oxford University Press.
architecture. Architectural Design 68(11-12), 80-
85. Norberg-Schulz, C. (1971). Existence, Space and
Architecture. New York, NY: Praeger.
Harrison, A. (1994). Intelligence quotient; smart
tips for smart buildings. DEGW, architecture Norman, D. (1993). Things that Make Us Smart,
today(46). Defending Human Attributes in the Age of the
Machine. Perseusbooks.
Hatada, T., Sakata, H., &Kusaka, H. (1980).
Psychological Analysis of the Sensation of Reality Nystrom, B.D. (1992). Perceived Image Quality
Induced by a Visual Wide Field Display. SMPTE of 16:9 and 4:3 Aspect Ratio Video Displays.
Journal, 89(8), 560-569. Journal of Electronic Imaging, 1(1), 99-103.

Oosterhuis, K. (2002). E-motive Architecture.

Rotterdam: 010 publishers.
Rahim, A. (2000). Contemporary processes in

Architecture (AD) 70, No. 3, Wiley Academy
publications.
Rahim, A. (2000). Contemporary Techniques in

Architecture (AD) 71, No.4, Wiley Academy
publications.
Rapoport, A. (1969). House Form and Culture.

Englewood Cliffs, N.J: Prentice-Hall.
Reeve, B., Detenber, B., & Steuer, J.S. (1993).

New Televisions; The Effects of Big Pictures and
Big Sound on Viewer Responses to the Screen.
paper presented to the information of systems
division of the international communication
association, Washington, DC
Reeves, B. & Nass, C., (1996). The Media

Equation, How People Treat Computers,
Television, and Newmedia Like Real People and
Places. Cambridge MA: Cambridge University
Press.
Sharples, S., Callaghan, V., & Clarke G. (1999). A

Multi-Agent Architecture for Intelligent Building,
Sensing and Control. International Sensor Review
Journal, 132.
Szell, M. (2003). Intelligent Building, Integrating

Design. Periodica Polytechnical
Ser.CIV.ENG(47), No.1, 49.
Vidler, A. (2000). Warped Space. Cambridge, MA:

MIT Press,176.
Wigginton, M., & Harris, J. (2002). Intelligent

Skins. Oxford: Architectural Press,171
Zellner, P. (1999). Hybrid Space, New forms in

digital architecture. London: Thames and Hudson
Publications.

A. Mathew / Smart Homes for the Rural Population
Smart Homes for the Rural Population: New

Challenges and Opportunities
Anijo Mathew1
1
Mississippi State University
Abstract
“Smart” Homes (domestic environments in which we are surrounded by interconnected technologies that are
more or less responsive to our presence and actions) seem increasingly plausible with the emergence of
powerful mobile computing devices and real time context aware computing (Edwards and Grinter, 2001).
Research at premier technology universities have given birth to home “labs” that experiment with sensors,
cameras and monitors to study physical, behavioral and social consequences of such technologies on occupants
of such homes. One of the most important problems that “smart” homes will eventually help to address is that
of spiraling costs of healthcare. Using ubiquitous technologies to motivate healthy decisions can help prevent
the onset of myriad medical problems (Intille, 2004). Moving the focus of attention from the health centers
and hospitals to the working home through such technology interventions would eventually lead to decreased
financial pressure on the traditional healthcare system.
This paper examines the challenges and opportunities in the design of “smart” technologies for preventive
healthcare in rural homes. It summarizes findings from current ethnographic and demographic studies; and
examines other contemporary research in the field of ubiquitous computing and “smart” homes. With the help
of these studies, the paper lists different technical, social and functional challenges that we as designers may
have to consider before designing “smart” homes for rural populations.

Introduction many computers available throughout the physical

environment, while making them effectively
invisible to the user (Morris and Lundell, 2003).
The problem of the house
Before we start talking of how incredulous (or not)
“The problem of our epoch is the problem of the such a claim seems, let us first take a look at two
house” – Le Corbusier pervasive technologies that were once unique but
is invisible to us today. The first is the written
word; the birth of the writing system coincides
Architects of the 20th Century imagined that their
with the transition of hunter-gatherer societies to
new tools – electricity, steel, concrete, plate glass,
more permanent agrarian encampments when it
mass production and fresh ideas about design –
became necessary to count one’s property, animals
could be used to transform society for the better
or measures of cultivation. A natural result of the
(Larson et al., 2004). Perhaps it did, perhaps it
cultivation and storage of grain was the production
didn’t. The problem of the house still remains; the
of beer. It is not surprising, therefore, that some
only difference is that the house has changed.
of the very oldest written inscriptions concern the
Today’s homes are not just well designed; they
celebration of beer and the daily rationing allotted
are also “intelligent”. Intelligent thermostats adjust
to each citizen by the elite of these communities.
the interior temperature based on the temperature
It took thousands of years for man to develop the
outside; the digital video recorder knows when
written language and thousands more before it
your favorite TV program is on and records it so
became accessible to the common man. The
that you can watch it later; the microwave can
second technology is electricity- at the turn of the
connect to the internet to calculate how long you
century, the homes of the wealthy were often
should defrost your dinner. However, with new
outfitted with electrically-conducting rails in the
intelligence, comes new problems. Eventually we
floors; “electricity girls” equipped with metal
will live in so-called “smart” homes and the
shoes and wearable light fixtures, would entertain
architects of the 21st century have to be concerned
party guests by moving from room to room,
about more than the just the “chassis” of the house.
carrying their own illumination (Edwards and
These “smart” homes offer new opportunities to
Grinter, 2001).
augment people’s lives not just with good design,
but also ubiquitous computing technologies that
Today, the written word and electricity is so
will provide users with increased communications,
pervasive that we have difficulty imagining that
awareness, and functionality (Weiser, 1996). With
they were once considered novel technologies
new tools such as inexpensive computing, almost-
used only by the elite rich for vanity and control.
free electronics, wireless communication, high
Computer scientists believe that the same
performance materials, and new design,
pervasiveness will one day be true of computers.
fabrication, and supply-chain technologies
The signs are already here - the same chip that
(Larson et al., 2004), we are today equipped with
landed the Eagle on the moon is running your
perhaps more than the architects of Corbusier’s
laptop today, what one day used to take a whole
times. But as with Corbusier’s architects, are we
room to fit, can be held in the palm of your hand.
perhaps in awe of these new tools available to us?
As architects, we should be ready to design for a
Are we oblivious to the change that is happening
future where affordable sensing and computation
in the pervasive computing realm and do we as
will find its way into nearly everything man-made,
architects need to understand the use of these tools
including building components (Larson et al.,
as more than just material for design?
2004).
The computer is taking over my house

As “smart” homes become realities in urban
centers, it is inevitable that such technologies will
Weiser (Weiser, 1996) talks of three waves of eventually trickle down to the rural home. The
computing – the first wave was mainframe question is, as we begin to design for rural centers,
computers; the second wave, personal computers; do we work with the same constraints as those of
and the third wave is referred to as ubiquitous our urban counterparts or do we develop a new
computing. Ubiquitous computing can be defined guidelines and standards? Do we design for the
as a method of enhancing computer use by making

same ethnographic issues as those in urban centers home help or pay for privately provided
or do we evaluate the different settings to inform assisted living
our designs? 3. less well-off return-migrants at or near
retirement age, returning to a family
Background home, to be near relatives and friends
4. better-off return-migrants
The people of my village 5. incoming retiree migrants without local
connections who are generally more
educated and materially better off.
The image of the agrarian village is no longer the
(Bryden, 2002)
American dream. Demographic changes and
economic re-structuring create significant
challenges for the rural population of America. No doctor in the house
Couple this with new lifestyles and comforts - the
myth of the extended farm family breaks down.
The problem with such a changing demography
Bryden (Bryden, 2002) explains how longer life
is that the healthcare system is not adequate to
expectancies and lower rates of natural population
adjust to this change. Aging results in decreased
growth, have changed the demography of
performance of abilities - physical: ability to
America’s villages:
provide for one’s need; psycho-cognitive: ability
to act and participate; and social: fulfill one’s role
1. The rural population is “older” than the
in the society.
urban population in the USA. 11.9% of
the total population is aged 65 or over,
while 13.9% of the non-metropolitan
population is 65 or over.
2. The non-metro population is older than
the metro population with a median age
of 36 in 1998, compared to 34 for the
metro population.
3. The rural elderly are poorer than their
urban counterparts, and the incidence of
poverty increases with age, ethnic group,
and geographical remoteness.
This demographic imbalance is a significant

problem in many rural states, especially since these
states are also actively encouraging policies to
attract retirees. Attracting retirees, claims Bryden
(Bryden, 2002), is a good economic development
policy but there are both costs and benefits. Not
only is the aged urban population now migrating
to the villages, there is also a universal tendency
for many young people to leave their rural
homelands to gain education, training and
experience.
Figure 1. Proportion of Population (per age) needing
He categorizes people living in rural United States Assistance
into five types:
1. indigenous poor who lack the resources However, rural healthcare systems work
in the form of employment or differently from urban healthcare. Small, rural
employment pensions hospitals in addition to being the only source of
2. relatively well-off indigenous people, emergency care are often a community’s only
who have sufficient resources to hire resource for health care services such as long term

care, home health services and out-patient services situation, the Medicare reimbursement system
(Bryden, 2002). assumes that it is cheaper to deliver health care in
rural areas, and its reimbursement rates for rural
Problem of lack of health care providers: Rural health providers is correspondingly lower than it
healthcare systems are burdened by the cost of is in cities (Bailey and Nemet, 2002).
healthcare services and the lack of health care
providers. Mississippi for example has one of the Problems of access to healthcare: In a study
lowest active physicians to population ratio in the (Bailey and Nemet, 2002) of ‘elderly health care
country. utilization in New Orleans County”, 69% of the
respondent’s physicians were located in the two
main population centers. Those traveling to the
nearest hospital faced a round trip of 160 miles,
often necessitating an overnight stay.
There is also a lack of access to private and public

transportation by the rural elderly and the poor.
Many small rural centers do not have public
transportation systems and often not even taxi
services. This impacts the elderly and the poor who
do not own a car or are unable to drive one.
I don’t want to leave my house
One solution to the problem is to move the elderly

to assisted living centers and nursing homes; while
other solutions include the diversification of
Figure 2. Unnecessary Hospitalizations hospital services. The “swing bed” program allows
patients to stay in the hospital beyond the end of
Problem of cost: Medicare and Medicaid systems their acute stay and receive nursing services they
are often the only sources of revenue for rural need. However, in the following study conducted
hospitals. Yet a recent report shows that Medicare by Forrester Research in 2003, it was found that
spending for rural counties was significantly lower although nursing homes house almost 1.5 million
than for urban counties. To further aggravate the seniors and one-third of those have been there for
TOTAL TOTAL
ACTIVE PER 100K
GEOGRAPHIC AREA PHYSICIANS POPULATION
UNITED STATES 727,573 257.9
NORTHEAST 185,439 345.7
SOUTH 237,788 236.5
EAST SOUTH CENTRAL 36,899 216.4
MISSISSIPPI 4,931 173.1
WEST SOUTH CENTRAL 65,952 209.1
NORTH CENTRAL 152,573 236.6
EAST NORTH CENTRAL 108,711 240.5
WEST NORTH CENTRAL 43,862 227.7
WEST 151,773 239.2
Table 1. Active Physicians per 100K population

Source: American Medical Association. Physician Characteristics and Distribution in the U.S.,2002-2003 edition. Chicago, 2002.

more than three years, 77% of American 1. The rapid adoption of powerful mobile
consumers say that nursing homes are a last resort computing devices. These devices can be
for themselves and their family members (Boehm used to collect and process sensor data
et al., 2004). from wearable wireless sensors and
convey this information to the user.
The above statistics echo concerns that the current Future devices will be smaller, lighter,
clinical model of healthcare is not working at an inexpensive, and available in a variety of
optimal level. As a result, it is an ideal time to form factors.
initiate changes within people’s lifestyles and 2. The emergence of real-time context
homes to help augment the current model of aware computing. A context aware
healthcare. One suggested alternative is a computer can automatically infer what a
prevention rather than crisis management person is doing from sensor data.
approach. Research has shown the significance
of moving the focus of attention from the health Morris and Lundell at Intel Research identify
centers and hospitals to the working home through different ubiquitous computing technologies that
technology interventions (Intille et al., 2003). could potentially be used in homes (Morris and
Lundell, 2003):
Ubiquitous computing for proactive healthcare • Wireless broadband/networks - to allow
communication between devices
Technology • Biosensors, Activity sensors, and bodily
diagnostics – non intrusive real time
Computer scientists envision that one of the first monitoring of users and occupants
uses of ubiquitous computing at home will be in
the field of proactive healthcare. Intille (Intille,
• Information fusion and inference engines
2004) claims that the using technology in homes – that must ferret out inherent noise and
would be one way of potentially reducing the correctly infer high level behaviors from
spiraling medical costs. He mentions two trends low level sensor data
in computing will allow the introduction of • Ambient displays and actuator displays
computing within the architecture of a home: – to enable communication between the
user and the system
Figure 3. Preferences of healthcare

• Natural interfaces – that enable adoption of new technologies these areas. In our
communication through natural media own ethnographic studies we charted the following
built into the lifestyle of the user biases that rural populations carry:
Eventually these technologies will be integrated 1. Economic bias: Most of the current rural
with architecture and design; used not just to cure population has little or no access to any
sickness and help people with cognitive decline, sort of computing, leave alone ubiquitous
but also to promote wellness throughout all stages devices. Much of the population does not
of life. Currently, MIT (MIT-TIAX PlaceLab, even have access to common household
2004) and other premier institutions are devices. Even if they did have the facility
developing “living laboratories” – places where to buy these devices and were willing to
the concepts of ubiquitous computing (sensors, do, they may not have the economic
cameras, monitors) can be studied with real-life resources to do so.
users. These laboratories are developing and 2. Cultural bias: Culturally rural areas are
testing several products that can be introduced into not accustomed to the use of technology
the architecture of a house, to track, monitor and solutions to solve problems. Due to
provide feedback to the resident. Such problems of availability and accessibility,
technologies can successfully motivate long-term a rural person would consider technology
healthy decision making, and thus delay or even as his/her last resort.
prevent the onset of medical problems such as 3. Lifestyle bias: The lifestyle of rural
obesity and chronic illnesses, alleviating the America is significantly different from
pressure on the traditional healthcare system that of the urban. Current designs are
(Intille, 2004). usually created with the largest user base
in mind– the urban buyer. Hence, some
Adoption of these solutions may not be suitable for
use in rural areas or even if they are, may
Several independent studies show us that the not be used in the same manner by rural
American population is more receptive of users.
technology solutions and health monitoring
systems than ever before. A major reason for this Ethnographic studies by Intel Research (Morris
shift is perhaps the wide acceptance of congruent and Lundell, 2003) however shows that in spite
technology (internet, cable television, cell phones) of these cultural, ethnic and socioeconomic
by the American consumer. In one such study, barriers to adoption of new technologies; there is
Forrester Research shows: more enthusiasm for hi-tech solutions among
extremely low income caregivers with very little
Since we are primarily looking at a rural previous exposure to computing than with those
population, it is also important to consider with significantly higher financial resources.
Figure 4. Adoption of technology by US consumers

Smart Homes for the Rural Population face-to-face mode. Forrester Researcher estimates
that only beyond 2010 will third party payments
be initiated for preventive healthcare technologies
In their paper, Edwards and Grinter (Edwards and in the house (Boehm et al., 2004). The challenge
Grinter, 2001) present seven challenges from for us designers is to work with government
technical, social, and pragmatic domains for using agencies in order to introduce these technologies
ubiquitous computing at home. Although these into rural architecture. The initiative may begin
challenges are true for most homes, we believe with population centers and finally trickle down
that the rural home is different. As we move to the rural poor. As more people begin to use these
towards designing in rural areas, it is paramount technologies, the government will realize the
to develop similar constraints that are unique to importance of such designs and mandate
rural homes. These constraints would then form legislations to make them insurable.
the framework for the design and ultimately the
implementation and maintenance of the “I don’t know how to use it”
technology. In the following part of this paper, we
present the challenges that we believe will be the In early days of the computer, system
challenges of the rural “smart” home. In order to administrators were an important part of the
initiate a discussion on what constitutes a “smart” everyday running of the system. As computer
home for rural America, it is important to think systems became more ubiquitous the role of the
like the rural person. Thus, we present our system administrator shifted to the user. But rural
challenges in the form of concerns a rural person America is still not tech-savvy enough. More often
may have when first introduced to “smart” home than not technologies, common in the cities, have
technology: not reached these rural communities. When they
do, people don’t always know or want to know
“I can’t afford it” how to use them. It is important thus to make these
systems as uncomplicated as possible. One method
The limiting factor for computing initiatives for is to shift the intelligence from the device to the
healthcare in rural America will not be lack of network. Traditional appliances, like telephones
need, lack of interest, or immature technology; it or televisions are commonly accepted because the
will be the lack of funding (Boehm et al., 2004). intelligence of the system lies in the network and
not the device. The home only contains the most
Ubiquitous computing initiatives are available simple and minimal “front end” functionality
today – for those who can pay out of their own needed to access the network (Edwards and
pockets. The rural population is significantly Grinter, 2001). It is paramount that future designs
poorer than its urban counterpart. The irony of the also be able to scale as well as degrade gracefully.
situation is that it is the rural population that A component that fails should not bring the rest
perhaps requires these technologies more than of the system down. In addition to the workability
those in the cities; poor healthcare centers, poor of these systems, insurance companies may also
outreach of medical facilities, and even lack of demand to see certain levels of safety (meeting
access to health providers. One way that these regulations and codes, seismic tolerance etc.). In
technologies can reach the rural poor is if the short, the “smart” home of the future must also be
government mandates legislation to include them a reliable home of the future.
as part of the medical insurance system or
subsidize the technology to enable initial “Where do I buy it?”
penetration. Similar subsidies were available to
the rural public with new technologies like the While new homes may eventually be built for such
telephone and electricity. Initial subsidies allowed smart applications, existing homes are not
the rural owner to purchase these technologies; designed as such (Edwards and Grinter, 2001). It
drive the price down; thus make the technology would be also presumptuous to assume that
an essential commodity. However, only five states someone who does not want to move to a nursing
– California, Kentucky, Louisiana, Oklahoma, and home will consider moving out of a house that
Texas – have passed legislation mandating holds their past. An acceptable alternative is to
reimbursement of tele-medical consults that would develop another (retiree) home that in some way
be covered if treatment occurred in the traditional resonates with their earlier home. In both situations

it is important for us to understand that most questions of whether it will work with the lifestyle
“smart” technologies used in the homes will be of the user or not.
bought piecemeal from a local convenience store
or a large specialty store. The challenge for the “Will it take over my home?”
architect is to anticipate change in the designs of
their spaces when these technologies are brought Science fiction movies and books have ingrained
together gradually. It is also important that into the minds of people that when computers
affordances are made for old devices to be reach our homes, they will be in control. Real
removed and new devices to be added without the research however takes a very different route.
inconvenience of changing the basic design. What People feel strongly when a computer takes all
designers must also assure is the impromptu the decision; it makes them feel insecure and out
operability as well as interoperability of these of control. The best method is to abandon the “shut
technologies. Architects must eventually design up and eat your mush” approach and take a more
spaces that will not end up as islands of “here are your options” approach. Leave all control
functionality but connect seamlessly as a whole. for decision making to with the user but instead
One way to ensure such a design is to draw on the provide just-in-time information highlighting the
way people use spaces and designs currently; benefits of engaging in particular behaviors.
another is to study the behavior of people in natural Instead of trying to wrench control from the user,
settings informing us the best configuration of the computer should reward a behavior using
space and technology. powerful motivational strategy of positive
reinforcement operant conditioning (Intille, 2004).
“Will it interfere with my life?” Such long term rewards to the user will ensure
that the behavior change should be sustained even
One of the most important challenges of stitching on the removal of the interface. The designer must
ubiquitous computing into architecture is to create make the choice between persuasive (indicating
a non-disruptive environment. The problem of that it is time to take medication) and coercive
today’s technology is that it conflicts with the real mechanisms (forcing the person to take
world which is a “highly analog” environment medications). Most “smart” home technologies
(Edwards and Grinter, 2001), presenting a great also have severe privacy implications. Even the
deal of ambiguity and unpredictability. These most subtle design may be too intrusive for a rural
technologies must not only accommodate family. Using sensors and cameras to monitor a
differences across individuals but also differences user may not be acceptable by certain
across households. communities, even though such a solution would
be the most apt for the problem they face.
Rural “smart” homes must present information
attuned to the lifestyle of the user. The challenge If “smart” technologies have to be accepted by
is to design for the rural lifestyle: what works in the larger rural community, it must eventually be
the city may not work in the village. Current designed for them and with them. Without the
emergency response systems work fine if you fall involvement of the end user, these technologies
in the house, but if you fall in the garden while may never find use off the store shelf.
watering your tomatoes, you may have a problem
(Coughlin, 2001). Elderly people also live “I don’t want to be isolated”
differently from younger people. Morris and
Lundell’s study showed that elders seemed to The last thing a designer wants is to isolate the
designate “command centers” (a kitchen chair, a inhabitants of his/her designs from the real world.
bed) that served as a base for entertainment, eating, Any technology should aspire to catalyze rather
work, and socializing. Unless the technologies are than replace human interactions (Morris and
designed within easy reach of these command Lundell, 2003) and social connections. Any system
centers, they may not be used at all (Morris and that remotely offers the idea of isolation faces the
Lundell, 2003). Designers must understand that danger of immediate rejection. Morris and Lundell
in the rural home, cultural and social biases will (Morris and Lundell, 2003) claim that socializing
take precedence over technical and aesthetic is a strong motivation for participation in healthy
biases. It is important then to re-analyze the behaviors and that socially isolated members of
designs built for rural households; asking their group were often less satisfied, less optimistic

and in poorer states of health than their active situations? In such cases, it is important that the
counterparts. Our studies show that there is a large architect, instead of physically visiting the site and
increase in the number of rural elderly that live changing design, take on the role of a
alone and a decrease in the number of married computational critic (Larson et al., 2004) –
couples and couples with children. Most rural someone who, using the communication
elders are retirees whose children have moved technologies, discusses with the users on how they
away to a large city or to another part of the country can change the design to better suit the change in
or the globe. Many of these children still have their lives.
backward ties with their parents and grand parents
– often concerned for their health and well being. Designers of technology solutions also assume that
Successful aging requires the elder to reach a the users will use the technology in a particular
balance between self-confidence on one hand and manner; but users don’t. They will always find
comfort with increasing reliance on others on the novel ways to circumvent conventional thinking
other (Morris and Lundell, 2003). Design of of the designer and carve out new uses for the
“smart” homes should eventually address the solution. Predicting these circumventions may be
issues of a rural population that is living alone, difficult, but it is important that we pay attention
yet looking for connections to their children (and/ to the use of these systems by actual rural users to
or other relatives) and vice versa. understand not only the working but also the scope
Not only is it important to design for social of use afforded by the solution. In case of elderly
interaction and connectivity, but also for shifts in homes, a further dilemma is created by the fact
social dynamics within real life situations. Edward that although the end users of such technology may
and Grinter (Edwards and Grinter, 2001) mention be the elderly; the actual running of it may be done
that the challenge for us as designers is to be aware by younger stakeholders (children, grandchildren,
of the broader effects of our work, and to realize caregivers etc.). Thus, designs have to not only
that even simple technologies (washing machines, account for front end interfaces but also back end
telephones, televisions) can have broad changes interfaces that are easy to understand and fix.
on the dynamics of the home and society.
The Role of the Architect
“Can I fix it myself?”
Products developed and tested in laboratories often
It is a challenge for designers to design systems
fail when introduced into natural settings such as
that will ensure that the users understand the
homes because they are often designed without
pragmatics of sensors, interpretation, and machine
understanding that human behavior in natural
actions as well as they understand the pragmatics
settings are strongly tailored to the settings
of devices in their homes now (Edwards and
themselves and to the behavior of the people
Grinter, 2001). The rural user has several
nearby (Intille et al., 2003). One common trend
constraints to the use of a design that is not evident
we find among the research centers currently
in urban areas. Most rural homes often have
working on a new “lab” or a “home” is that the
accessibility issues – it may not be possible for a
locus of attention is on technology. Many solutions
technician to make a round every fortnight to
from these labs are often based on technical rather
check on the components of the “smart” home. It
than aesthetic or usable values; the architect can
is also contradictory to assume that users who do
help by understanding the idiosyncrasies of the
not have proper access to health centers will have
user. As a designer, he/she is capable of changing
access to large stores that can service these
the product to meet the cultural, ethnic and the
technologies and devices. The plausible solution
socioeconomic requirements of the user. But
to these problems is to create devices as simple as
research needed to tap the potential of new
the ones they use today. Unlike current internet
architecture is fragmented and out-of context
technology, which relies on devices at the edge of
claims the Open Source Building Alliance:
the network; future networks should be
independent of the devices in the home.
Computer Scientists attend pervasive computing
conferences to present visions of the future, but
The issue poses a pertinent question, what if it’s
rarely does an architect attend. Architects gather
not the device but the space itself that needs
to debate the latest design ideologies without
improvement – to adjust to new lifestyles, new

including those who actually make the systems homes of the future and to work together to find a
and materials they will use. Health researchers common solution or set of solutions.
propose visions of the “smart” medical home of
the future, but without careful study of how the
behavioral and non-medical needs of their patients
will change when they leave the hospital for the
home. Devices are prototyped for the home
without evaluating their use in the complex mix
of everyday activity (Larson et al., 2004).
Conclusion
The fact that this paper has remained a report of

existing research and study of current conditions
is intentional. We have used prevalent conditions
and research to frame a set of unique constraints
that will help us to develop designs in the context
of rural America. We believe that such basic study
of conditions is necessary to develop the problem
(ask the question properly) before finding a
solution (answer the question). At Mississippi
State University’s Design Research & Informatics
Lab (DRIL), we are working towards research in
the areas of actually using technology within
homes for proactive healthcare. As we progress
towards the next wave of architecture, it is
important that the architects be aware and
conscious of the change in the way we design
homes. Through live projects and user studies, the
DRIL works with undergraduate and graduate
students to develop designs that use these
technologies within the architectural framework
of the rural home. Besides teaching the DRIL is
also involved in multiple research projects
exploring the acceptance of technology by
Mississippi’s population; the design and
deployment of contextual devices and designs; and
developing designs with rural/quasi-rural
communities that can be acceptable and sustained
by the people within or migrating to these
communities.
It is our belief that only through the careful

mediation of technical aspects of design with the
phenomenological and the aesthetic, can we dream
of using ubiquitous computing in our homes.
Elegant and usable design will be achieved only
through conversations between researchers in
various disciplines and through widespread
awareness of current literature and research in this
realm. It is our hope that the above discussion will
help us initiate a dialogue– to understand both the
problems and the opportunities of designing rural

References
Larson, K., Intille, S.S, McLeish, T. J., Beaudin,
Bailey, Adrian J., Nemet, Gregory F. (2000). J., and Williams, R. E. (2004). Open source
Distance and Health Care Utilization among the building-reinventing places of living. BT
Rural Elderly. Social Science and Medicine, Vol. Technology Journal, vol. 22, pp. 187-200.
50, Pergamon Press, Oxford.
MIT-TIAX PlaceLab. (2004). A Residential
Bryden, John M. (2002). The Elderly Outside the Research Facility to Study Supportive
Metropolis: Myths and Realities. Lecture for the Technologies for the Home http://www.tiax.biz/
Haskell Master Class October 25, 2002; Arizona images/spotlights/place_lab_032803.pdf.
State University.
Morris, M., and Lundell, J. (2003). Ubiquitous
Coughlin, J.F. (2001). “Technology and the Future computing for cognitive decline: Findings from
of Aging”. Proceeding: Technologies for Intel’s proactive health research. Alzheimer’s
Successful Aging. Supplement to Journal of Association website.
Rehabilitation Research and Development; Vol. h t t p : / / w w w. a l z . o r g / R e s e a r c h e r s / R G P /
38: 40-42. Intel_UbiquitousComputing.pdf
Edwards, W. K; Grinter, R. E. (2001). At Home Weiser, Mark. (1996). Open House - Xerox PARC.
with Ubiquitous Computing: Seven Challenges. h t t p : / / w w w. u b i q . c o m / h y p e r t e x t / w e i s e r /
UBICOMP 2001: Ubiquitous Computing wholehouse.doc
International Conference; 2001 September 30-
October 2; Atlanta, GA. Berlin: Springer Verlag;
LNCS 2201: 256-272.
Boehm, E.W, with Holmes, B.J., Brown, E.G,

Bishop, L., McAulay, S.E., Gaudet, J. (2004).
Forrester Big Idea: Who Pays For Healthcare
Unbound (The $34 Billion Market For Personal
Medical Monitoring).
Intille, S.S., Tapia E.M, Rondoni, J., Beaudin, J.,

Kukla, C., Agarwal, S., and Bao, L. (2003). Tools
for studying behavior and technology in natural
settings. Proceedings of UBICOMP 2003:
Ubiquitous Computing, vol. LNCS 2864, A.K.
Dey, A. Schmidt, and J.F. McCarthy, Eds. Berlin
Heidelberg: Springer, pp. 157-174.
Intille, S. S., Larson, K., and Tapia, E. M. (2003).

Designing and evaluating technology for
independent aging in the home. Proceedings of
the International Conference on Aging, Disability
and Independence.
Intille, S. S. (2004). A New Research Challenge:

Persuasive Technology to Motivate Healthy
Aging. Transactions on Information Technology
in Biomedicine, vol. 8(3), pp. 235-237.


M. Senagala / Building is a Network for Living in
Building is a Network for Living in: Toward New

Architectures
Mahesh Senagala1
1
University of Texas
Abstract
Our societies today are beginning to think, communicate, interact and live differently as everything in the
human world is beginning to be networked wirelessly at the speed of light with everything else in the world
(including architecture). This article looks at the big picture and outlines a series of recent developments in
digital technologies that would enable architecture to become sensate, supple and globally networked at the
speed of light. New thinking, new commerce, new polity, and new architectures are emerging out of the
apparently disparate yet closely related design and technological inventions. We are on the verge of moving
from the outmoded notions of space and time to the post-spatial notion of sensate and supple space-time.

Introduction toward a global coalition of an interconnected

world (de Chardin, 1961). He called such a world
“Architecture of intelligence is the architecture of noosphere (the sphere of interconnected human
connectivity. It is the architecture that brings consciousness). A recent CNN story chronicles
together the three main spatial environments that how the new WiMAX technology buoys these
we live in and with today: mind, world and possibilities by preparing to introduce cell phone-
networks.” like ubiquitous and wide-ranging network
Derrick de Kerckhove (2001) coverage for laptops and other computational
devices. A WiMAX-equipped laptop or
It is the author’s contention that something computational device can stay connected to the
phenomenal and previously unforeseen is Internet all the time without any wiresi. Intel is
emerging from a series of small yet important planning, according to this report, to package
developments in architecture and related fields. WiMAX into laptop chips by 2006. Such a
By infusing computational abilities into technological network would be a crucial turning
architecture, by enabling architecture to become point in the journey toward the emerging
sensate, intelligent, responsive and adaptive, a noosphere.
number of apparently disparate developments are
resulting in the emergence of a new way of At a time when even the toasters and refrigerators
thinking and being. These new developments are are being hooked up to the Internet, architecture
pointing toward a paradigmatic shift in the is not going to be left too far behind, despite the
evolution of architecture as a discipline. Our best efforts to resist the evolution by the
thinking is beginning to shift from the notion of profession’s conservative core. Connection is the
computers in architecture or computers affecting keyword, the buzzword and the overarching
architectural design to the notion of architecture concept of how buildings could become networked
as the computer. Our notions of architectural computer systems. General Motors’ OnStar ®
practice are shifting from the model of a lone vehicle security system has already transformed
architect standing in front of a drawing board to our automobiles, which are now controlled by
architect as a conductor of events and resources dozens of embedded computers and networked via
that are globally distributed yet digitally connected satellites, into GPS-powered real-time network
into a networked practice. This paper is an attempt nodes. An OnStar operator can access most of the
to bring together seemingly far-flung critical systems of an automobile remotely and
developments in various fields into a coherent suggest or coordinate a course of action at the
framework to reveal the emergence of this touch of a remote button. The biggest evolutionary
paradigmatic shift. As de Kerckhove has pointed jump for automobiles is not in their growing
out, connectivity is the keyword for the engine size or seductive body shape; it is in the
architecture of intelligence (de Kerckhove, 2001). pervasive computerization and wireless digital
We are at the dawn of the era of fully networked networking. The mobile space of the automobile
buildings. has been transformed into an interactive real-time
network node capable of keeping us connected to
A Building is a Network for Living In the rest of the world. Automobiles are already a
part of the emerging noosphere. Later on we will
A network is an interconnected system with a discuss the OnStar At Home® pilot program,
certain structure of relationships. Kevin Kelly, which extends the OnStar’s automobile network
executive editor of WIRED magazine summed up services to the home market. Architecture is
the essence of the current paradigmatic shift: “the becoming a part of the post-spatial network
central act of coming era is to connect everything ecosystem.
to everything. All matter, big and small, will be
linked into vast webs of networks at many levels. We are temporally burrowing our way across the
Without grand meshes there is no life, intelligence, globe to form supply chain networks, trans-
and evolution; with networks there are all of these national organizations, trans-continental firms and
and more” (Kelly, 1994). Nearly five decades ago, 24-hour workdays spread across time zones. We
long before the computer or the Internet became do not necessarily connect in a spatially contiguous
popular, Teilhard de Chardin prophetically manner any longer. Our economies are dependent
proclaimed that the human evolution is heading on temporal connectivity between the globally

distributed manufacturing, sales, marketing, and used to describe the post-spatial developments.
administration networks. These different terms have different connotations
and agendas of significant variation. What the
In such a world, work or performance is broken plethora of terms indicates, despite their
down into a series of events or tasks that can be encampments, is the desire for architecture to be
assigned in real-time to any part of the globe with thought of as an active and interactive being.
no heed paid to national boundaries (geographic
space), cultural barriers (cultural space), economic Smart Architecture is the term currently popular
disparities (economic precincts), and other spatial in the sustainable design communities, where
logic. The author calls such a condition of going smartness often refers to the ability to regulate a
beyond the dualistic Newtonian and Cartesian variety of environmental parameters for an
notions of space and time Post-spatial. A more efficient energy usage (Park, 2003, Kienzl, 2002,
elaborate discussion of the theoretical dimensions Pagani, 1999, Daniels, 1997).
of the post-spatial world can be found elsewhere
(Senagala, 2003). Responsive architecture puts a performative
emphasis on increasing the interactive nature of
The Shift from Timeless Space to Timed Space building elements in response to environmental
or social or commercial needs in real-time (Ishii,
“The new produced and projected space has less 1998). Responsive environments might involve
to do with lines, surfaces and volumes than with any form of responsiveness.
the minutiae of viewpoint, the dynamite of tenths-
of-seconds. These viewpoints are simultaneously Kinetic architecture, which can be traced back to
time-points in the tele-topological continuum of the 1960s, is preferred by those who see
long-distance projection and reception.” architecture as a literally more mechanical, kinetic
Paul Virilio (1991) and moving organism (Zuk, 1970) where spatial,
mechanical movements are essential to its
Architecture of stone and brick has always been definition and existence. However, all of these
the one sphere of life that embodied a sense of terms represent directions in architecture that
timelessness. Edifices of stone resist time and involve active, instantaneous response to either
define a pattern of space for thousands of years. internal or external conditions through a central
We no longer build our world in stone for a number computational agency.
of reasons. The notion of timelessness, however,
has stayed rooted in the profession, its thinking, Architecture: Supple and Sensate and Connected
and in what people expect to see embodied in
architecture. Except for minor deviations, most of “A building is made up of other spaces within it
architectural thought has been focused on that move and change, even if its own walls remain
establishing a single inert spatial footprint that is fixed. The idea of the mobility of building and
programmed to stay immobile 24 hours a day and within building is one possible idea of Deleuzian
365 days a year. Compare this to television, which thought that might be of tremendous value in
is programmed to the second. Imagine holding a architecture. Building is not only a movement of
single pattern of pixels on the television screen sedimentation and stabilization but also a way of
24 hours a day. Despite the fact that architecture opening space and living.”
and television have different tasks and purposes, Elizabeth Grosz (2001)
both can benefit greatly from understanding the
value of timeliness and intelligent interactivity. William Mitchell has put it lucidly when he said
“we become true inhabitants of electronically
Architects and other professionals working in the mediated environments rather than mere users of
related fields are indeed giving up their millennial computational devices” (Mitchell, 1999). If we can
allegiance to the notion of timelessness. They are simply transform the role of the wall, window,
beginning to work with accommodating real-time floor and roof, we can redefine what architecture
responsiveness into an array of local forces and is all about. Such a redefinition would be an
global conditions. Such terms as smart historic transformation in and of itself. And, we
architecture, responsive architecture, adaptive are not even talking about what might emerge from
environments, and kinetic architecture are being such a transformation. This revolution is already

taking place. Traditionally, walls, roofs and floors The Tangible Media Group of MIT has
have always been delimiters and separators of demonstrated that more than simple bits of
space, not connectors. Typically, when an architect information can be communicated through
draws a wall, it usually denotes a spatially static architectural interface. The group created the so-
separation of activities. Architects, in general, have called ambientROOM, which is remotely
come to embrace spatial separation and connected to a number of devices (Ishii, 1998). In
segmentation as the primary modes of design the Active Wallpaper project, sensors in a remote
thinking. A number of inventions and design location pickup the activity level and transmit it
developments—often from outside of architectural to the ambientROOM where the information is
profession—are questioning the ontology and revealed in the form of a pattern of illuminated
epistemology of walls, floors, windows and roofs. patches. When the activity is high, the movement
What is the reality of a wall or a floor? How do of the patches is high and vice versa. In the
we know a wall or a floor and, more importantly, Pinwheel project, colorful pinwheels spin at
how do these architectural elements know us? different speeds based upon their information
How can these elements be digitally networked, input. Ishii points out that these efforts “envision
controlled and connected? In the next few pages that the architectural space we inhabit will be a
we will see select examples that manifestly new form of interface between humans and online
demonstrate different ways of answering these digital information” (Ishii, 1998). He sums up the
questions. project thus: “the ambientROOM surrounds the
user within an augmented environment – ‘putting
One more paradigmatic shift that can be noticed, the user inside the computer’ – by providing subtle,
from the projects discussed in this paper, is the cognitively background augmentations to
shift from static, frozen and fixed architecture to activities conducted within the room.”
supple, kinetic, responsive and parametric
architecture. The form, environment, qualities and Teleporting a comprehensive experience is at the
characteristics of a building become linked, heart of these experiments. Two or more spaces,
connected and dependent on certain parameters. in this scenario, could be in sync with each other
Thus, a building can become more flexible, despite their physical separation by vast distances,
interactive, connected and more sustainable over presumably even interplanetary distances! This
longer periods of time. kind of connectivity radically redraws the map of
spatially fragmented geographic locations into a
Architecture as Immersive Interface / map of temporally contiguous experiences.
Architecture as Info Pump Inventions such as these, question our outmoded
notions about context, region, and neighborhood.
It is believed that in less than a decade, more than
one billion people all over the world will be Talking to the Walls
spending half as many hours in front of the
computer screen as they will in physical space (de Acoustic Tap Tracking system, developed by the
Kerckhove, 2001). The computer screen has team led by Joseph Paradiso of MIT Media Lab,
become the 17" gateway to all the digital is capable of transforming any large flat surface
information out there. What if we are able to such as a wall or a window or a table into a tap
expand the ways by which we see, hear, touch and sensing interactive responsive surface (Paradiso,
sense information? What if we can release more 2000). In distinction, the other higher tech devices
people from the screen for more hours by developed elsewhere that use optical cameras,
distributing the interface around the architectural touch screens, lasers, pens and light curtains, have
environment? What if the walls, floors, lighting, many inherent problems. From cost to portability,
ventilation and other facets of the architectural from effectiveness to scalability, these systems
environment begin to communicate information have not had much success in transforming things
to the user? What if architecture, as a whole, at architectural scale.
becomes a gigantic immersive interface to send
and receive information? What if architecture
becomes a spatial synesthetic pump to channel,
amplify and process the temporal flows of digital
information?

interactive screen. If we stretch our imagination

around this invention, we can think of walls and
tables becoming interactive interfaces. The dumb,
blank, abstract walls that divide can become
interactive, intelligent, sensing, smart surfaces that
could truly connect in real-time. Information
becomes immersive. Walls could become portals
to spatially distant lands or culturally other worlds.
Soft, Supple and Sentient
More and more architects are beginning to explore

the form and format of sensate and supple
architecture. The Muscle, a prototype featured at
the Non-standard Architecture exhibition ii, which
was designed by the Dutch design firm
Oosterhuis.NL (ONL), dons a pneumatic structure
that behaves like a gigantic, digitally mediated
muscle. The building would flex, contract, expand
and mold itself to suit changing programmatic
conditions over time and in real-time. The
architects propose that the place be used for a
variety of activities such as a disco or a television
studio or a meeting place. The synthetic muscles
of the Muscle react as people move near the sensor
points. Another way to manipulate the structure is
by moving the sliders on a remote computer
screen. Thus, the building becomes spatially
interactive and can be plugged into the Internet.
This alien-looking blob is not necessarily how
buildings might look as a whole in the future. Some
buildings might look that way. However, the
Muscle, as ONL proposes, can literally be used as
a series of muscles in a building to control any
environmental or other parameters in real-time.
Perhaps the buildings might be able to literally
Figures 1 and 2. Acoustic Tap Tracking System by express their feelings by flexing their facial and
Responsive Environments Group, MIT Media Lab spatial muscles, thereby leading to new and supple
expressionism in architecture.
The ingenuity of Paradiso’s system is that it is Swift, Supple and Sentient

buildable for under $500 and can be scaled to be
installed on large architectural surfaces. Earlier In 1998, design firm dECOi, led by Mark
version of this device came from Paradiso’s Goulthorpe, won a competition to design an
collaborations with Media Lab’s Tangible Media interactive art piece for the foyer of the
group, which resulted in the so-called Birmingham Hippodrome. The piece functioned
PingPongPlus demonstrated at SIGGRAPH 1998. as a mediator between events happening inside
In the experimental prototype, four transducers the theatre to the human activity outside the
sense the taps on a 4’x8’ shatterproof glass pane. theatre, forming a link between the public plaza
The analog audio signals are analyzed for peak and the theatre itself. Aegis Hyposurface was
timing and fed into the computer as intelligible conceived as a responsive surface which reacts
taps such as mouse clicks. Thus, it is possible to physically in real-time to the events happening
transform any shop front window into a simple around it.

Figures 4, 5 and 6. The Muscle by Oosterhuis.NL Figures 7 and 8. Aegis Hyposurface

The responsive wall surface was made of a pliable

material or skin stretched over 896 highly
responsive and sensate pneumatic pistons (known
as actuators). These computer-controlled actuators
generate coordinated movement across the
surface, allowing it to create complex patterns that
rapidly reconfigure its appearance in response to
a variety of digital input such as user movement,
light and sound (Zellner, 1999). This structure
differs from the Muscle by ONL in significant
ways. Aegis Hyposurface translates given
information into a four dimensional matrix of
triangular surfaces. It is a 4D screen. It is as if we
have been liberated from the world of flat two-
dimensional pixels into the world of dynamic four-
dimensional experiences. Much has been said and
written about this project. However, the true
implications of this post-spatial construct go
beyond its kinetic nature. The responsive wall can
respond in real-time to any digital instructions that
would arrive either from local sensors or via the
Internet from remote sensors located anywhere
else in the world. Thus, the very ontology of wall,
its relationship to space, time, and its connectivity
to local and remote phenomena are brought into
critical focus. The sensate space-time of the
installation can potentially extend to all corners
of the globe. The post-spatial wall could become
a threshold between spatially discontinuous yet
temporally connected worlds. We are witnessing
the beginnings of a radical redefinition of
architectural reality. Figures 9 and 10. Magic Carpet System by Responsive
Environments Group, MIT Media Lab
Floors That Know

and communicate the prexel (pressure pixel)
The Responsive Environments group at MIT information to the main computer for further
Media Lab has devised a sensate floor system programmable action. Thus, with this relatively
called Magic Carpet. It cannot fly, but the magic inexpensive and robust flooring system, it is
carpet does know a lot about who is on it. In this possible to transform any floor into a sensing
system, a series of piezoelectric wires in X and Y surface capable of forming an intelligent
directions are used to sense the footstep dynamics environment. The possibilities that this invention
such as pressure and movement. The sensing opens up are immense. It is possible to control
medium used here is an inexpensive shielded wire such architectural characteristics as lighting levels,
that is capable of producing a small voltage location, ventilation, security and other
(around 15 Volts) spike when pressed. These environmental parameters. It is possible to estimate
signals are then processed using various filtering the number of people and the kind of activity
and clustering algorithms to decipher the location through analyzing the data in real-time, which can
and pressure information. increase or decrease the heating and cooling levels
in a room or control the room illumination or
The Magic Carpet system has led the team to activate the entertainment system or open up the
develop networkable floor tiles called Z-Tiles. In walls or lead to many other architectural
this system, a series of interlocking tiles, equipped possibilities. Perhaps a walk on the floor during
with embedded processors, form an adhoc network October could activate the sound of the rustle of

leaves. Perhaps a dance on the floor could adjust The system consisted of a kinetic spaceframe
the acoustic characteristics of the walls to better driven by three sets of Festo pneumatic pistons.
suit the music. Thus, the very meaning of floor The original prototype was designed to sense the
stands redefinition. A sensate interactive floor- audience movements through pressure sensitive
equipped room can literally come to life and, in mats and translate the impulses into valve
the process, transform our life-world. operations controlled via the computer. The
scenario depicted in figures 13 thru 17 envisages
some dramatic architectural possibilities. This
invention was an attempt to question the
conventional notions and limitations of
architectural programming. In conventional
programming, spatial configuration is held more
or less static by walls and floors, while people and
events are orchestrated. Topotransegrity proposes
to program the walls and floors along with people
and events in an all enthralling four-dimensional
framework. This project represents a truly
groundbreaking shift in architectural thinking.
At Home (from) Anywhere
The Internet Home Alliance

(www.internethomealliance.com), a remarkable
cross-industry collaboration between GM’s
OnStar, Invensys, ADT Security Systems, HP,
Panasonic, and many other corporate partners have
launched, in early 2002, a post-spatial initiative
to integrate OnStar’s Virtual Advisor® service
with home security control, telecommunications
control, and climatic control from any Internet
enabled appliance anywhere in the world. This
system also gives the customer visual access to
his or her house at any time. Garage doors, access
Figures 11 and 12. Z-Tiles by Responsive Environments doors, windows, all the major home appliances,
Group, MIT Media Lab HVAC system, security system, and
telecommunication systems are networked using
the HP Application Server 8.0 framework as
Programmable Buildings gateway. The participating customer would be able
to access any and all of these aspects in real-time
The Latent Utopias exhibition curated by Zaha from any computational device such as a cell
Hadid and Patrick Schumacher featured a phone or a PDA or a laptop. The customer would
groundbreaking kinetic responsive prototype be able to remotely turn on or off the kitchen stove
named Topotransegrity. It was a curious mix of or refrigerator. Panasonic’s smart doorbell would
topological manipulation though pneumatic notify the customer anywhere, through audiovisual
spaceframe structure. The project, designed by access, anytime his or her doorbell is rung.
5Subzero, a budding group of architects from
London, featured surfaces that can be manipulated One interesting fact about this whole enterprise is
through either an automated control mechanism that the alliance was not initiated by architects,
or a real-time feedback system or a pre- not overseen by architects, and certainly not
programmable system. Imagine a floor/roof controlled by architects. Once the OnStar At
system that can be manipulated to form different Home® technology matures, it will be a matter of
configurations depending on different spatial extending, scaling, and porting the technology to
needs as illustrated in the pictures. all building types.

Figure 13. Topotransegrity by 5 Subzero
Architects’s usual clerical obsession about making

static things should yield to connecting dynamic
space-times as the primary mode of crafting new
realities. Architects have the opportunity now to
take a proactive step to become a part of the
leading edge of this revolution than to wait and
continue to ride the trailing edge of technological
advancements. Architects need to shift the
discourse from fabricating shiny surfaces to
designing intelligent interfaces. If architects fail
to take notice today, the world might continue to
be plagued by the banality and spatial mediocrity
of suburban homes simply extended into the
cybernetic realm. The majority of the architectural
profession still thinks in terms of static, immobile,
Figure 14. Topotransegrity Prototype in Latent Utopias non-interactive, and unwired buildings. As de
Exhibition, Graz, 2002 Kerckhov (2001) said it rightly, “the opportunity
for the architect is not merely to improve the drafts
The notion of architects as visionaries seems to or the elevations, or even benefit from the CAD
have become passé. Non-architects are taking over system, it is to rethink the real. How do architects
that role. The question now facing architects and greet this opportunity?”
designers is to take ownership of this emerging
mode of thinking, which goes beyond making
merely well-detailed boxes or blobs, and enters
the realm of designing networkable architecture.

Figure 15. Topotransegrity System Layout
Figure 16. Pneumatic Pistons Figure 17. Topotransegrity Scenario

Acknowledgements
A longer version of this paper is going to be

published in: Senagala, M., “Post-spatial: The
Emergence of Sensate Space-time in
Contemporary Architecture,” in Rivka Oxman,
ed., On Digital Design and Design Thinking,
Design Studies journal special issue, Elsevier
Press, 2005. Page numbers and issue number not
available at the time of this manuscript preparation.
Web Sites
www.media.mit.edu
Figure 18. OnStar At Home® Architecture. Courtesy
Internet Home Alliance www.oosterhuis.nl
www.internethomealliance.com
Conclusions http://www.5subzero.org
Looking at big picture scenarios, which is what

this paper attempts to do, is like peering through a End Notes
foggy landscape. Things are blurry at best from a
distance and become clearer as we get closer. Back i. http://www.cnn.com/2004/TECH/internet/10/
in 1969, ARPANET was formed when we hooked 18/wireless.broadband/index.html October 19,
up a computer to a computer to a computer to 2004
another computer. A global network was then born
that led to the emergence of the Internet. No one ii. See www.oosterhuis.nl for this and its more
had thought, at that time, that a network of four current version as well. The exhibition was held
hosts would later give rise to novel political at Centre Georges Pompidou from December
campaigns, new commerce, and a hundred million 10th, 2003-March 1st, 2004.
connected homosapiens dwelling online. The
principle to understand here is that once a system
reaches a critical level of connectivity and
complexity, something new emerges that is beyond
anyone’s comprehension.
When we infuse computational intelligence into

our environment, when we transform our
environments to respond to the post-spatial
network society, and when we redefine the
fundamental meaning of the basic architectural
elements such as walls, floors and windows, we
transmute architecture into a connection machine.
More importantly, by digitally connecting a
building to a building to a building to another
building, something revolutionary might emerge
that goes beyond what has been envisaged in this
paper. That would be the day when your homepage
truly merges with your home.

References Pagani, F. (1999). Adaptive Buildings through

Evolutionary. Design Doctoral Thesis, Vancouver,
Born, M. (1962). Einstein’s Theory of Relativity. Canada: The University of British Columbia
New York, NY: Dover Publications, Inc.,
Park, C. (2003). Occupant Responsive Optimal
de Chardin, T. (1961). The Phenomenon of Man. Control of Smart Façade Systems. Doctoral Thesis,
New York, NY: Harper & Row Atlanta: Georgia Institute of Technology, USA
de Kerckhove, D. (2001). The Architecture of Potts, W. (1984). “The chorus line hypothesis of
Intelligence. Basel, Boston and Berlin: manoeuvre coordination in avian flocks.” Nature,
Birkhäuser. vol. 309. PP. 344-345
Grosz, E. (2001). Architecture From the Outside. Daniels, K. (1997). The Technology of Ecological
MIT Press, Cambridge, Mass., USA Building. Verlag, Basel, Switzerland: Birkhauser
Ibelings, H (ed). (2000). The Artificial Landscape Zuk, W. (1970). Kinetic Architecture. New York,
NAI Publishers, Rotterdam, Holland NY: Van Nostrand Reinhold
Ishii, H, et al. (1998). “Ambient Displays: Turning

Architectural Space into an Interface between
People and Digital Information” in the
Proceedings of the First International Workshop
on Cooperative Buildings (CoBuild ’98).
Darmstadt, Germany: Springer Press. PP. 22-32
Kelly, K. (1994). Out of Control: The New Biology

of Machines. New York, NY: Perseus Books.
Mitchell, W. (1999). E-topia: Urban Life Jim, But

Not As We Know It. Cambridge, Massachusetts:
The MIT Press.
Paradiso, J et al. (2000). “Sensor systems for

interactive surfaces” in IBM Systems Journal,
Volume 39, Numbers 3 & 4
Senagala, M. (2003). “Postspatial Architectures:

The Emergence of Time-like Parametric Worlds”
in Cultura Digital y Diferenciacion, Proceedings
of the SIGRADI International Conference,
Rosario, Argentina. PP. 375-382
Virilio, P. (1991). Lost Dimension. New York, NY:

Semiotext(e)
Zellner, P. (1999). Hybrid Space. New York, NY:

Rizzoli
Kienzl, N. (2002). Evaluating Dynamic Building

Materials: The Potential Impact of Climatically
Responsive Building Enclosures. Doctoral Thesis,
Cambridge, Mass.: Harvard University

G.. Katodrytis / Poiesis and Autopoiesis in Architecture
Poiesis(*) and Autopoiesis in Architecture

George Katodrytis1
1
American University of Sharjah
Abstract
The use of digital technology in architecture has proven to be more assertive than originally thought: it has
reconditioned the nature of the design process, and established new processes and techniques of fabrication.
Recent applications in digital technology show inquisitiveness in the contentious subject Genetic Algorithms.
This new architectural process is characterized by two main shifts: from poiesis (or poetry) to autopoiesis,
and from authenticity to mimesis. Since evolutionary simulations give rise to new forms rather than design
them, architects should now be both artists and operators of both Inventive and Systematic design.
Inventive design: The digital media should bring about poiesis (poetry). Digital spaces reveal and visualize
the unconscious desires of urban spaces, bringing forth new dreamscapes, mysterious and surreal. This implies
a Freudian spatial unconscious, which can be subjected to analysis and interpretation. The tools of digital
dreaming, meanwhile, have opened a window to the ‘urban unconscious’.
Systematic Design: Digital media should bring about an autopoiesis. This approach calls into question traditional
methods of architectural design that replace the hierarchical processes of production known as “cause and
effect,” and propose a design process where the architect becomes a constructor of formal systems. Will the
evolutionary simulation replace design? Is metric space dead? The new algorithmic evolutionary conditions
give architecture an autopoiesis, similar to biological dynamics.
Paradoxically, the new emerging process is more insightful. The emphasis of the exploration is on morphological
complexity. Architecture, through “machine” fabrication, may become more responsive, rigorous and poetic.
(*)Poiesis: poetry, creation, especially of a poetic inspiration.

Digital Uncanny
I would argue that from its inception, digital media

were considered a discipline external to
architecture. By definition the digital in
architecture does not exist. Despite this,
architecture seems to truly lend itself to digital
exploration. It creates a topology of symbolic
forms as digital constructs. More importantly, it
manifests itself in the most ambiguous element –
space – within which any projection moves freely
and without fixed boundaries. The new technology
of the digital media has managed to unravel the
repressed condition and abandoned projects of 20th
c architecture (Futurists and Surrealists), and to
challenge the only ideology that created it:
modernism and its associated technology. One may
talk about the relationship between the digital Figure 1. Project by author
uncanny and introjections.
In architectural terms, the search for modernism’s

repressed condition was concentrated in the The blurring of lines between the mental and
domain that the modernists had clearly and physical, the organic and inorganic was
polemically identified as the basis of their attack transformed by the surrealists, especially Dali, into
on tradition: the irrational, the decorative and the a formulation that stressed the intersection of the
uncanny. A good example is Tzara’s indictment biological and the constructional, building and
of modern architecture as a “complete negation psyche, architecture and hysteria, in order to
of the image of the dwelling.” “Modern produce the ultimate object of desire, or its
architecture,” Tzara argued, “as hygienic and reification at least. Characterized by its mimesis
stripped of ornaments as it wants to appear, has of the digestible, it was an architecture that, in
no chance of living…because it is the complete Dali’s words “verified that urgent “function,” so
negation of the image of the dwelling.” necessary for the amorous imagination: to be able
in the most literal way possible to eat the object
In the modernists’ tradition, the line between of desire.” Walter Benjiamin stated that the
nature and machine, between the organic and the intersection of technology and nature was
inorganic seemed clear; organicism was a represented by the displacement of symbols from
metaphor, not reality. But for the current digital Romanticism to Modernism. Here we may begin
media, the boundaries between organic and to trace the affiliations of Surrealism and
inorganic are blurred; the body itself, invaded and modernism on the level of technique - affiliates
reshaped by technology, in turn invades and that were stated by Benjamin in the aphorism: “The
permeates the space outside, even as this space reactionary attempt that seeks to detach the forms
takes on dimensions that themselves confuse the imposed by technique from their functional
inner an the outer, visually and physically. As context and to make natural constants out of them
Walter Benjiamin presciently observed, “The work – that is to say, to stylize them…” In Benjiamin’s
of Le Corbusier seems to arise when the ‘house’ terms, the structure that unified the two was
as mythological configuration approaches its end.” fetishism. For fetishism, in its multiple
Digital technology attempts to reincarnate these displacements, “suppresses the barriers which
“mythological configurations,” repressed by separate the organic from the inorganic world.” It
modernism, with the monstrous and anamorphic is as “at home in a world of the inert as in the
merging of animal and house as an oneiric world of the modern mechanization of the
machine, a machine for dreaming. After all, there dwelling in its mission of repression against the
is no architecture without dream, myth or fantasy. bric-a-brac of the nineteen century.”

Mimetic Impulse correlative behavior in which a subject actively

engages in “making oneself similar to an Other”,
dissociates it from its definition as merely
imitation.
According to Adorno, “by means of the mimetic

impulse, the living being equates himself with
objects in his surroundings.” This surely holds the
key to exploring the question of how human beings
situate themselves within their environment, and
points to an area in which the domain of
psychoanalysis may offer crucial insights into the
mechanism by which humans relate to their
habitat. It begins to suggest, for example, that the
way in which humans progressively feel ‘at home’
Figure 2. The act of mimesis in a particular building, is through a process of
symbolic identification with that building. They
What happens, then, when the fusion between the may come to identify with technological objects.
organic and the inorganic takes form? Mimesis? This symbolic attachment is something that does
A mimesis that has a multiple interpretation. not come into operation automatically; it occurs
Digital technology is mimicking architectural gradually. Mimesis in Adorno, and in Walter
space so much that it becomes believable and Benjamin, is a psychoanalytic term - taken from
“real,” that organic and inorganic matter, animate Freud - that refers to a creative engagement with
and inanimate forms become indistinguishable. an object. Mimesis is a term, as Freud himself
Form becomes malleable and changeable and predicted, of great potential significance for
interactive, as though it imitates its occupants. The aesthetics.
body fuses with its surroundings.
To understand the meaning of mimesis in Adorno
According to Walter Benjamin and Theodor we must recognise its origin in the process of
Adorno’s biologically determined model, mimesis modelling, of ‘making a copy of’. In essence it
is posited as an adaptive behavior that allows refers to an interpretative process that relates not
humans to make themselves similar to their just to the creation of a model, but also to the
surrounding environments through engagement with that model. In mimesis
assimilation. Through physical and bodily acts of imagination is at work, and serves to reconcile the
mimesis (i.e. the chameleon blending in with its subject with the object. This imagination operates
environment), the distinction between the self and at the level of fantasy, which mediates between
other becomes porous and flexible. Rather than the unconscious and the conscious, dream and
dominating nature, mimesis as mimicry opens up reality.
a tactile experience of the world in which the
Cartesian coordinates of subject and object are not It is important to recognize here the question of
firm, but rather malleable. temporality. Symbolic significance may shift,
often dramatically, over a period of time. What
Adorno’s discussion of mimesis originates in a was once shockingly alien may eventually appear
biological context in which mimicry (a mediator reassuringly familiar. The way in which we engage
between life and death) is a zoological predecessor with architecture must therefore be seen not as a
to mimesis. Animals are seen as genealogically static condition, but as a dynamic process.
perfecting mimicry (adaptation to their
surroundings with the intent to deceive or delude Mimesis for Benjamin offers a way of finding
their pursuer) as a means of survival. Survival, meaning in the world, through the discovery of
the attempt to guarantee life, is thus dependant similarities. These similarities become absorbed
upon the identification with something external. and then rearticulated in language, no less than in
The manner in which mimesis is viewed, as a dance or other art forms.

Architecture along with the other visual arts can Benjamin’s use of the term mimesis - in the flash
therefore be viewed as a potential reservoir for of the mimetic moment, the fragmentary is
the operation of mimesis. In the very design of recognized as part of the whole, and the individual
buildings the architect may articulate the relational is inserted within a harmonic totality. It is within
correspondence with the world that is embodied the new digital spaces that the act of mimesis
in the concept of mimesis. These forms may be happens.
interpreted in a similar fashion by those who
experience the building, in that the mechanism by Algorithms and the breeding of digital forms
which human beings begin to feel at home in the
built environment can also be seen as a mimetic One technique by which mimesis can be
one. constructed is by Algorithms. Algorithms may be
defined as a detailed sequence of procedures to
solve a problem. As such algorithms may be
programmed to execute a series of mimetic tasks.
Genetic algorithms constitute a class of search
algorithms especially suited to solving complex
optimization problems. In addition to parametric
optimization, genetic algorithms are used in
creative design, such as combining components
in a novel and in an inventive way and ultimately
creating a new complexity of language and form.
Genetic algorithms transpose the notions of
evolution in nature to computers by imitating
natural evolution. They find solutions to a problem
by maintaining a population of possibilities
Figure 3. Project by author according to the ‘survival of the fittest’ principle.
Because of this ability to “search” algorithmic
scripts or codes generate form, which is precise
and complex and which would be impossible to
have conceived using the basic software interface
and tools. Inevitably, a new universe of formal and
compositional possibilities opens up and
techniques for digital form-finding.
A common approach is to define a building

envelope in terms of a series of parametrically
defined elements such as the structural ribs. Some
forms are curvilinear, non-planar and irregularly
shaped, yet precise. Furthermore, some approaches
that involve rule structures seek to generate
designs via various forms of growth and/or
Figure 4. Project by author repetition algorithms. Additionally, there are
approaches that seemingly abandon any kind of
Mimesis, then, may help explain how we identify formal approach to shape generation but seek to
progressively with our surroundings. In effect, we allow designers to “discover” meaningful shapes
read ourselves in our surroundings, without being that exist within more complex geometrical
fully conscious of it. “By means of the mimetic patterns. Most of the approaches that use formal
impulse,” as Adorno comments, “the living being shape-identification algorithms require specially
equates himself with objects in his surroundings.” written computational algorithms. Architects can
The aim throughout is to forge a creative now use advanced software to breed new forms
relationship with our environment. When we see rather than specifically design them. As De Landa
our values ‘reflected’ in our surroundings, this notes, “ …only if what results shocks or at least
feeds our narcissistic urge, and breaks down the surprises, can genetic algorithms be considered
subject/object divide. It is as though - to use Walter useful visualization tools.”

Algorithms are based on non-linear wave function Intuitive vs. systematic

that through parametric differentiation organizers
vectors of density. Can we then talk about Through this highly refined algorithmic process,
fabricating of dense and large cityscapes, both intuitive and logical form-making processes
employing poly-directional structural networks? are necessary. A grid-like structure transforms
gradually into a convoluted scheme with numerous
On another level, the role of design has now been overlapping coils or folds revealing complex
transformed into that of breeding fit and beautiful patterns. As the new architectural forms become
forms. There is clearly is an aesthetic component: complex, interconnected nodes and parameters
the “sculpting” of beauty and the development of change and, in a chain reaction, affect the outcome.
a personal artistic style. As with any socio- Algorithms offer a degree of rationality therefore
technological revolution throughout history, allow the designer to become a free form maker.
architecture inevitably invents a new formal
language. One assumes the role of the artist, not of computer
programmer. The outcome requires to be visual,
Ultimately, if traditional architectural and it is about inventing form than about solving
representation has been based in resembling and problems in a systematic way. The resulting
describing the appearance of the architectural exuberance of formal composition should be
object through its use, then algorithm architectural judged as any architectural proposal, i.e. first on
notation has become operational; to design the its visual impact, and the potential to become a
choreographing of the transformation process. The building.
architectural object is transformed into event and
performance, either by understanding architecture Architects may start thinking less in terms of
as the dynamics of spatial conditions, or by the typology and more about topology and variations.
object being understood as the actualization of One way to introduce this thinking process is by
built-up potentials. contrasting the results. More variations are
Figure 5. Project by author

necessary in order to keep the “search” exciting. to new possibilities, i.e., the emergent form.
One of the design goals of the generative process
is not only to support both the “designer’s” In the task of designing rich search spaces, certain
(intuitive) view and the “programmer’s” (formal) philosophical ideas, which may be traced to the
view, but also to reconcile these two views and work of Gilles Deleuze, play a very important role.
create appropriate mechanisms as an interchange Though not invented by Deleuze, he was the one
between them. who brought them together for the first time,
making the basis for a brand new conception of
New possibilities may be the introduction of a the genesis of form. According to Manuel De
heterogeneous structure, as layers of form Landa in his essay “Deleuze and the Use of the
generated by different codes. Complex behavior Genetic Algorithm in Architecture,” the productive
comes from the interaction of simple parts. The use of Genetic Algorithmic implies the deployment
basic principle of Genetic Algorithms is searching of three forms of philosophical thinking:
for optimum balance, and the eventual survival of
the fittest. Evolutionary Algorithms can be both a 1. Populational thinking: The sequence of
structural making device and a search mechanism. operations points at spontaneous and
The structurally oriented form-finding approach multiplemutations.
becomes a useful tool for composing space 2. Intensive thinking: Without the struc-
through repetition. Structural members define both tural engineering and distributions of
the surface and the volume of forms. stress, a virtual building will not evolve
as a building.
Genetic Algorithmic and the Autopoiesis of 3. Topological thinking: The obtained with
Architeture the Genetic Algorithm must demonstrate
an incredible combinatorial productivity
like in natural forms, with thousands pos-
sibilities.
The employment of genetic design strategies

develops autonomous architectural concepts,
which replace the traditional hierarchical processes
of production known as “cause and effect”: new
organizational patterns and weavings and
performative morphologies that can modulate and
differentiate the environment. This morphogenetic
process includes pattern, repetition and
permutations.
The tendency towards architectural autonomy

Figure 6. This image was created by placing the algorithm might be understood as a moment of overall
onto a unit sized NURBS plane societal process of differentiation. Traditionally,
architecture and good design were inseparably
The new digital approach to architectural design connected with society and harmony. The new
is based on computational concepts such as algorithmic evolutionary conditions give
topological space, isomorphic surfaces and architecture an autopoiesis. The autopoietic system
parametric design. Architecture is recasting itself, as a complex, historically evolving system, always
becoming - in part - an experimental investigation uses time and involves a series of events in its
of topological geometries. Digital media is “responses,” so that simple and predictable one-
employed not as a representational tool for to-one correlations between environmental
visualization, but as a generative tool for the impacts and system responses are out of the
derivation of form and its transformation – the question. Recent developments in digital
digital morphogenesis. It explores the possibilities technology expose a degree of autonomy that
of “finding form”. Topological space opens up a architectural discourse has established by
universe where essentially curvilinear forms are differentiating itself from the immediacy of
not stable but may undergo variations, giving rise everyday talk about buildings, and thus the

complexity of the discursive detour, which layer, that of interiority, by “weaving” in one
mediates a particular impact/response, should (enclosure) into the other (interior). There are two
grow with the overall complexity of society. levels of complexity: that of weaving of the skin
and that of the volume.
In weaving, the linear fluidity of wrapped objects

creates exciting visual forms. A repeated line gives
the forms an organic and often mystical quality.
Because of the predictability of a wrapped surface,
their generation lends itself to scripting and
procedural methods using wrapping algorithm.
One can imagine a solid shape being “mummified”
or wrapped at various layers.
This dual process is what it is known as the

“exogenous” and “endogenous” control of plant
development. Diffusion-limited aggregation and
cellular automata provide models of exogenous
mechanisms of branching pattern formation. In this
case, components of the growing structure
communicate through the surrounding space, in
contrast to the endogenous control mechanisms,
which rely on information flow within the
Figure 7. Project by author developing structure. In nature, endogenous and
exogenous control mechanisms are often
Current experimental work focuses on issues of combined. For example, the development of a tree
organizational complexity (layering, is affected by the genetically controlled formation
interpenetration of domains), on the production of meristems (apices), the flow of water, nutrients,
of diversity (iteration vs. repetition), on the spatial and phytohormones through the branching
recognition of fuzzy social logics (smooth vs. structure, and the plant response to environmental
striated space), on ways of coping with uncertainty factors, such as the shading and crowding of
(virtuality vs. actuality), and on engagement with branches. Environmentally sensitive systems
new production technologies. represent one of the approaches proposed to create
comprehensive models integrating endogenous
What is important about the morphogenetic model and exogenous phenomena. Architectural design,
is the degree to which it allows for a coexistence as well as fabrication, can now employ and imitate
of various forces, engendering an autogenetic, these operations.
autopoietic and mimetic system. Autopoietic
systems produce theories of complex self- Surface Spatiality / Structural Morphology
organization that nevertheless can become
problematic. Self-organization or autopoiesis is The new apace is the outcome of the synthesis
thus impossible without the necessary random between space-oriented and structure-oriented
infux of external forces. This is precisely Felix models, developing self-regulatory patterns in
Guattari’s point when he talks about the machine which potentialities are regulated by the
and its arrangement that a machinic arrangement developing structure itself. These techniques result
of heterogeneous forces and heterogeneous in the simulation of evolutionary and environment
autopoietic processes are much more interesting based three-dimensional structures and surfaces.
than simple models of poiesis. The new research in architecture involves
structural morphology and generative modeling
Multi-layering of architectural form. The design process has now
turned from the mimetic into one of growth, based
Moving beyond the design of an envelope alone, on given data (directions or restrictions).
only spatial relationships may incorporate another Algorithmic structure represents abstract patterns

that are not necessarily associated with experience on, only when the underlying geometric
or perception. Algorithmic processes result from relationships have been defined and tested. This
events that are often neither observable nor is an ideal process in any design, starting form
predictable and seem to be highly intuitive. In this broad ideas and gestures and which then are
sense, algorithmic processes become a vehicle for developed in more detail. Beyond CAD tools and
exploration that extends beyond the limits of the interface of even advanced modeling software,
perception. this requires new methods. These new tools should
be open-ended and programmable. As parameters
change so do the variety and topological
permutations. Well-established and used scripts are
those that generate what is called weaves and
braids. Given the specific surface or volume, they
achieve an interconnectedness of elements for
bracing and strength, like continuous materials.
Structurally, both tension and compression are in
a new set of relationship; more a network of
structural matrix. Weaves are based on parameters
for surface topology description, density, and
number of strands or threads. They can adopt a
helicoidal or crosshatch pattern.
The architectural process is now evolutionary and

intuitive. As if re-enacting Surrealism in a new
Figure 8. Project by author form, the modernist repressed condition may have
been liberated, replaced by the uncanny and the
myth and the new autopoiesis of architecture.
Figure 9. Project by author
Algorithmic Architecture employs methods for

creating new architectural morphologies. By using
scripting languages and working with codes it is
possible to create forms through methods
analogous to the evolution of intelligent life:
emergent behavior and self-organizing systems.
It pursues various methods through which the role
of the designer can shift from “space
programming” to “programming space”. What the
new scripts and codes can achieve is conceptual
broad gestures at the beginning, and precision later

Figure 10. Example of a script to pipe curves

References
Vidler, Anthony (2003). Fantasy, the Uncanny and

Surrealist Theories of Architecture. Essay.
Benjamin, Walter (1986). On the Mimetic Faculty,

Reflections. New York: Schocken Books.
Deleuze, Gilles and Guattari, Felix (1987) A

Thousand Plateaus. University of Minnesota
Press, Minneapolis.
Edited by Huhn, Tom (2004). The Cambridge

Companion to Adorno. Wesleyan University,
Connecticut.
Deleuze, Gilles (1994). Difference and Repetition.

Columbia University Press, New York.
De Landa, Manuel (2001). Deleuze and the Use

of the Generic Algorithmic in Architecture. Essay.

J. Daveiga, P. Ferreira / Smart and Nano Materials in Architecture
Smart and Nano Materials in Architecture

José Daveiga1, Paulo Ferreira2
1
University of California, Los Angeles
2
University of Texas at Austin
Abstract
We describe and analyze the fields of Smart and Nano Materials and their potential impact on architectural
design and building fabrication. Distinguishing Smart and Nano materials, Smart Materials perform both
sensing and actuating operations, whereas many Nano materials are capable of self-assembly. In general,
Smart and Nano materials can perform like living systems, simulating human skin, the body’s muscles, a
leaf’s chlorophyll and self-regeneration. Recognizing that the traditional partition between Materials Science
and Architecture is obsolete, our intent is to show how these two fields are intrinsically connected, while
growing evermore symbiotic as we progress into the future.
Keeping the designer in mind, our paper begins with the question: “What Nano and Smart materials can be
used in future architectural designs?” Outlining what such materials might mean for architectural fabrication
and design, we claim that Smart and Nano Materials can imitate living organisms. Effective implementation
of these materials will therefore allow designed spaces to operate as active organs within a larger dynamic
organism, synthesizing both expressive intent and pragmatic considerations.
This paper is a collaboration between an architect and a materials scientist on the future of materials and their
influence in architecture. By giving examples of work already underway we intend to illustrate and suggest
directions ranging from the functional to the expressive, from tectonics to morphology. We conclude with a
reflection on the importance of future research between our two areas of knowledge.

Introduction of an architect and a materials scientist, we propose

that the interplay between smart materials and
“Architecture should strive to imitate the nanotechnology can be more than buzzwords in
principles of nature without imitating its forms.” the latest design review: with real knowledge, real
Frank Lloyd Wright impact is possible.
Writing in the first half of the twentieth century, When we consider smart and nano materials in
Wright couldn’t predict what potentials lie in wait their particular and peculiar characteristics, their
for current designers willing to imitate both the design and construction potentials are easily
principles and forms of nature. If contemporary understood. Yet we would be remiss to leave you
field experts in materials science are correct, the without a word of caution regarding that same
near future of science and technology will offer apparent facility. Many times, when seeking new
products and procedures which follow the ideas and approaching new concepts, one can
principles of plant and animal growth and make strides in speculative and metaphorical
regeneration in extraordinary ways. Oddly enough, thought while reality lags further and further
these large scale changes will rise from the “I.Q.” behind. Eager to embrace new, cutting edge
of smart materials and by new uses of the tiniest technology, we as designers often let ourselves
of forms, such as nano materials. Imitating the dream into the future, forgetting the work toward
principles of the abalone, which self-assemble implementation which true creativity entails. To
their own shells atom by atom, scientists and apply many of the ideas discussed here, some
technologists are in the process of creating systems technical barriers between scientific discovery and
such as minute machines (nano-assemblers) and architectural application must be overcome before
coding mechanisms, which will imitate the realizations are complete. To prevent lags between
generative processes of nature following nature’s practice and theory, investment in realization must
growth and manufacturing methods. As the field keep pace with the hours invested in written
of materials science predicts, the “I.Q.” of smart speculation. Second, interdisciplinary research,
materials will be determined by the degree of the utilizing the best minds of both architecture and
material’s response to environmental stimulus, materials science must be encouraged and
whereas advanced nano materials will be supported institutionally and professionally. When
fabricated by self-assembly. The ramifications of these objectives are most nearly achieved, we as
this work will reach out from laboratories into designers will be best able to ground our
various fields, promising large scale re-inventions discussion in possibilities, not theoretical musings.
to rival the industrial revolution in impact and We hope this paper is a primary step in that
scope. grounded direction.
Health care, building methods and materials, Smart Materials and Nano Materials
fashion, technology, transportation, food and drug
composition and distribution, and material What are Smart and Nano Materials?
processing: all of these fields will be changed by
the import of smart materials concomitant with To critically engage the world of leading science,
nanotechnology and its resultant advancements. the architectural community should understand the
Architecture will feel the impact as well, via technological innovations that smart and nano
innovation in materials and construction, but also materials are capable of producing. To best do so,
through the resultant cultural changes that will we need a working definition for both smart
emerge in the face of new technology. At this materials and nano materials.
important juncture between science, technology
and culture, the architectural community will have Beginning with “Smart:” Smart Materials are
the opportunity to embrace, question and reject materials with the ability to perform sensing and
nanotechnological advancements. Without actuating functions, similar to those in living
grounded knowledge to do so, however, the import systems, upon the application of an external
of nanotechnology to the field of architecture will stimulus [Newman 1997] . This stimulus can be
be superficial at best, and at worst, just another in the form of light, sound, temperature, electric
inundation of mediatic tricks and flashy painterly field, magnetic field and stress. The term “smart
building skins. Herein, drawing on the expertise materials” gained relevance after the 1960’s Naval

Ordonnance Laboratory discovery of Ni-Ti alloys term “nanotechnology”.

(Nitinol) which were first used for the construction
of junction components in F14 aircrafts [Purdy Types of Smart Materials
1961]. If deformed, Nitinol will recover its original
shape when subsequently heated above a specific Piezoelectric Materials
temperature. Since its discovery, this material has Piezoelectric materials produce electrical charge
been used in many different applications, such as when mechanically stressed, and undergo some
medical implants, dental fixtures and electrical mechanical deformation when subjected to an
switches. electric field. Among the natural materials with
this property are human skin and quartz. When
Over the last 35 years, smart materials have further used in computers and watches, quartz is subjected
developed and diversified to include piezoelectric, to an electric field which causes it to vibrate at a
electrostrictive, magnetostrictive, and shape particular frequency, hence sending a signal at
memory alloys, which have both sensing and precise time intervals.
actuating functions. Additional materials, such as
electrochromic, electroluminscent and Piezoelectrics can also be used as resonators in
photoluminescent, pH-sensitive and photovoltaic items like watches, and in active noise control
are capable of sensing functions. These smart systems such as headphones, to detect noise
materials also carry the potential to become through the generation of electrical signals upon
intelligent materials, i.e., besides their sensing and vibrating due to the sound waves. Piezoelectrics
actuating properties, such materials will able to are also used in mechanical vibration damping
learn and adapt with time [Newman 1997] and systems, such as high competition skis and tennis
may ultimately simulate living systems. rackets. More recently, piezoelectric paints have
been studied to detect fissures in walls.
As for “Nano:” Technically, nano materials are Conductive paint is doped with piezoelectric
materials with a nanometer scale substructure. The particles that will react to stress/strain by
term “nano” derives from the Greek word for producing an electric current. In this way,
“dwarf”. It is used as a prefix for any unit like a piezoelectric materials can also be used as sensors
second, or a meter, and it means a billionth of that to test material stress/strain in construction.
unit. Hence, a nanometer is a billionth of a meter,
which is 1000 times smaller than the thickness of Electrostrictive Materials
a human hair. The eventual objective of Electrostrictive materials, like piezoelectrics,
nanotechnology is to manipulate and control produce electrical charge when mechanically
materials at the atomic level. In other words, we stressed and generate strain when subjected to an
expect to build things the way nature does it, atom electric field. However, while the strain produced
by atom and molecule by molecule, i.e., by self- is proportional to the applied electric field in
assembly. piezoelectric materials, in electrostrictive
materials, the strain generated is proportional to
In 1959, Physics Nobel Laureate Richard Feynman the square of the applied electric field. As a result,
challenged the scientific community by asking if for the same applied voltage, electrostrictive
we could write the 24 volumes of the Encyclopedia materials are capable of generating strains that far
Brittanica on the head of a pin [Feynman 1960]. exceed those of conventional piezoelectric
Although Feynman could not predict it, this lecture materials. The disadvantage is that the deformation
was a defining moment in the field of at low applied electric fields is small.
“nanotechnology”. The phrase “nano-technology” Electrostrictive materials have been used as wiper
was first used in print in 1974 by Norio Taniguchi blades, artificial muscles, landing gear hydraulics,
[Taniguchi 1974] to refer to the increasingly and for correcting surface deviations caused by
precise machining and finishing of materials. In thermal fluctuations.
the 1980’s, K.E. Drexler, described a new “bottom-
up” approach, involving the molecular Magnetostrictive Materials
manipulation and molecular engineering in the Magnetostrictive materials change shape when
context of building molecular machines and subjected to a magnetic field. Pure elements with
molecular devices with atomic precision [Drexler high magnetostrictive properties at room
1981, 1986], which eventually popularized the temperature are Cobalt and Iron. By adding other

elements one can achieve “giant” magnetostriction Hz) than SMA materials, without a reduction in
under relatively small fields [Hathaway and Clark the output strain, and are capable of attaining
1993]; this is the case of Terfenol-D. several kHz with lower output strains.
Magnetostrictive materials were first used in
telephone receivers and hydrophones. With the Current applications for the thermally driven
discovery of “giant” magnetostrictives further SMAs include coffeepots, thermostats, vascular
applications have developed, such as ultrasonic stents, hydraulic fittings or airplanes. MSM alloys
cleaners, high force linear motors, medical and will be used in loudspeakers, joysticks, ink-jets,
industrial ultrasonics for surgical tools and robots and air foils to control drag and turbulence.
underwater sonar. Magnetostrictive devices are
rather robust in terms of wear, and thus may Biomimetic Materials
potentially replace piezoelectric materials. The term “biomimetic” has origin in the Greek
words “bios” and “mimetikos”, which mean “life”
Electrochromic Materials and “imitate”, respectively [Newman 1997].
Electrochromic materials change color upon Hence, “biomimetic” means “to imitate life”.
application of an electrical voltage. An applied Biomimetic biosensors are used to convert a
voltage initiates a chemical reaction which changes biological response into an electrical signal. In
how the material reflects and absorbs light nature there is a wide spectrum of cases where
[Hardaker and Gregory 2003]. In some smart bio systems are utilized. Fish, for example,
electrochromic materials, the change is between contain a number of fibers that vibrate as they
different colors, such as those used in swim and act as sensors for flow noise. Simulated
electrochromic windows wherein the material transducers can then be used to talk to fish. Frogs
changes between colored (reflecting light of some communicate by emitting a low-frequency and a
color) and transparent (not reflecting any light). high frequency tone through dual-tone transducer
As a result, electrochromic windows darken when as well.
a voltage is applied and become transparent when
the voltage is removed. Reflective mirrors are Types of Nanoscale Materials
electrochromic as well, reflecting light upon the
application of an electrical voltage, instead of Nanoparticles
absorbing it. Additional potential uses of this
technology include rechargeable lithium batteries, Nanoparticles, also called “zero dimensional nano
sensors, privacy/security glazing areas and high materials”, do not possess any dimension outside
contrast displays. the 1-100 nm range. Although nanoparticles can
have different shapes, crystal structures and
Shape Memory Alloys compositions, their small scale induces a change
Shape memory alloys are of two types: thermally in fundamental properties. A good illustration of
and/or stress driven (SMA), and magnetically this phenomenon can be found in the properties
driven (MSM). Thermally/stress driven shape of quantum dots, such as CdSe, which are
memory alloys, such as Nitinol, undergo a phase nanocrystalline particles containing only a few
change under stress, from a high temperature phase hundred atoms. Because electrons in a quantum
to a low temperature phase, and return to the dot are confined to widely separated energy levels,
original high temperature phase when reheated as particle size decreases, the dots emit only one
[Newman 1997, Wayman 1993, Schetky 2000] . wavelength of light when excited. Thus, CdSe
Magnetically driven shape memory alloys such nanocrystals in solution, and exposed to incident
Ni 2 MnGa, on the other hand, undergo a light, emit radiation of a particular color [Dabbousi
reorientation of the low temperature crystal et al. 1997]. Accordingly, quantum dots can be
structure upon the application of a magnetic field used in the design of road reflectors and any other
[Ullakko 1996, Ullakko et al. 1996]. When the reflective surface that will be exposed to incident
orientation of the magnetic field is reversed, the light.
structure returns to its original shape. MSM alloys
show larger strains than traditional SMAs, and Magnetic nanoparticles can be used to detect
seem to be the most promising for future particular biological entities, such as
applications. Additionally, MSM alloys can microorganisms that cause disease. In these
already obtain 50 times greater frequencies (250 applications, magnetic nanoparticles are attached

to antibodies which bind to a target. When binding create a new surface. In some cases, thin coatings
occurs, the magnetization vector of each with a nanocrystalline structure are formed in
nanoparticle is parallel, thus inducing a strong selected substrates, through a dispersion of
magnetic signal. On the other hand, antibodies nanoparticles. In other cases, the films are so thin
which do not recognize the target do not bind and that their thickness is within the 1-100nm range.
display randomly oriented magnetization vectors, The most recent nanofilms are just a few molecules
with no strong net magnetic signal [Reed and Tour thick, having been built up by self-assembly, i.e.,
2000]. atom-by-atom or molecule-by-molecule [Kotov
2001].
Nanowires and Nanotubes
Nanofilms are stable and able to cover larger areas.
Nanowires and nanotubes are materials created Once applied, they are easy to manipulate using
with one non-nano dimension. These materials standard processes and tools. Potential
tend to have one long axis (above 100 nm), and a applications of nanofilms include the development
cross-section that is within the 1-100 nm range. of scratch-resistant plastic coatings, low friction
The best examples of these structures are the coatings, and materials with a very low or negative
widely publicized carbon nanotubes [Iijima 1991, refractive index. Nanofilms are useful in the
1994]. These nanostructures are composed of microelectronic industry, data storage industry,
carbon atoms that assemble into cylinders a mere solar energy area, optical devices and medical
1.4 nm in diameter, and a few micron in length equipment.
(Figure 1). A single-wall carbon nanotube has
extremely good electrical properties (conducting
electricity better than copper) and very good
mechanical properties (its tensile strength is much
higher than steel). A multi-wall carbon nanotube,
on the other hand, has semiconductor properties,
such as nanowires of ZnO, GaN and SnO2 [Yang
2005], which are attractive to the microelectronic
industry.
In 1998, the idea of using carbon nanotubes to

fabricate simple minute gears evolved by bonding
rigid chemical ligands onto the external surfaces
of carbon nanotubes to produce ‘gear teeth’. Each
gear is made of a 1.1 nm diameter nanotube with
seven benzyne teeth. Computer simulation results
show that the gears can operate up to 70 GHz (70
billion times a second) in vacuum at room
temperature without overheating or slipping [Han
et al. 1997].
There are many applications for which nanotubes

Figure 1. Single-wall carbon nanotube with a diameter of
and nanowires may become important, in 1.4nm. These carbon nanotubes can be thought as wrapped
particular electronic and opto-electronic devices, sheets of graphite
magnetic storage, and composite materials with
very high strength-to-weight ratios.
Bulk Nanomaterials
Nano Films
Bulk nanomaterials are materials that have
Nano films are materials that have two non-nano macrodimensions in 3-D, but exhibit a nanoscale
dimensions. Typically, nanofilms are used as a substructure. Currently, there are two main
surface treatment when composition or/and processes used to achieve bulk nanomaterials : 1)
mechanical properties need to be altered, or as production of nanoparticles in powder form,
coatings when a different material is deposited to followed by high-pressure compaction and

subsequent high temperature sintering to materials can further point architecture towards
consolidate the powder, and 2) high angular the ideal of a systemic organism; wherein a
extrusion or high pressure torsion, where relatively material or construction can respond in action and
small metal work pieces (of the order of cm) are reaction to its environment. Furthermore,
highly deformed to produce a nanoscale emboldened by these new material technologies
substructure. Current research in field of architects may find new ways to question
mechanical behavior demonstrates a broad range Aristotelian notions of separation and causality,
of fascinating properties for bulk nanomaterials, expressive necessity and the demands of function.
namely a significant increase in hardness [Sanders [Kalay 1999]. These dichotomies, epitomized in
et al. 1997], yield stress (up to five times higher) the famous statement by Louis Sullivan that “form
[Chokshi et al. 1989] and, in some cases, ductility follows function” [Sullivan 1934], may dissolve
[Kock et al. 1999]. The disadvantage of these as technological innovation and expression reach
materials thus far is that they are difficult to toward new unities. Resulting new concepts of
produce in large scales. materiality can then allow designers to break from
‘function or form’ into a new cyclic concept, and
Bulk nanomaterials have potential for applications a new paradigm, that creates one body of ‘function
where high strength-to-weight ratios, high strain and form’.
rate and low temperature superplasticity and
ductility in usually brittle ceramics. These fantastic Below we discuss the impact that both Smart and
mechanical and physical properties make bulk Nano materials can procure in the realms of
nanomaterials attractive for advanced applications architecture, construction and design. It is our goal
in medical, construction, space, sporting goods, to describe how these materials will not only allow
and transportation industries. for new and better responses to the performance
needs of today’s most efficient buildings but will
Smart Materials and Nano Materials in also allow for the shift in architectural thought
Architecture regarding materiality and the qualities of materials
in architecture.
An architect seeking new and better ways of
building is often driven by a thirst to develop new Intergraded Functionalities
forms and express new ideas of space.
Architectural history is built of such developments “Making familiar products from improved
as the ‘newest’ and the ‘boldest’ of building forms materials will increase their safety, performance,
and construction advancements illustrate. We and usefulness. It will also present the simplest
believe that new smart and nano materials have engineering task. A greater change, though, will
the potential to open a new era in architectural result from unfamiliar products made possible by
design and construction, enabling architects a new manufacturing methods.”(Drexler)
higher level of intricacy which will span from the
smaller scales of a molecule to the larger concepts Drexler’s vision was strong though his urgency is
of society. only now reaching the world of architecture.
Alluding to layered structures and manufacturing
In this vein, advancing scientific areas are now, processes with multiple capabilities, Drexler’s
as they have often been, the ideal scavenge site vision was to push science and society toward
for designers eager for new solutions to their greater efficiency, with amplified results at lower
technical problems, and looking for better ways costs. In line with Drexler’s vision, material
to express emerging design concepts. As novel scientists are now drawing inspiration and data
Smart and Nano materials research is no exception, from the bat biosonar, shark-skin denticles and
architects must look ahead for the impact these whale songs to develop Smart materials. In
novel cutting edge materials will have in the parallel, a wide effort has been invested in
constructible world. New concepts of materiality emulating mitochondrias in plant cells, sea shells,
emerging from Nano and Smart materials can spider webs, protein synthesis, as well as other
allow architects to further respond and engage our living nanoscale machine-like matter to create
present – connected, fast paced, organic, plastic, nano materials. Strains of molecular manufacturers
responsive, consumable, volatile, media-centric, are working to create nanoscale machines (nano-
and tech-centric. In response Nano and Smart assemblers) and structures which will be able to

build, program, record and transmit information, weight ratio through the use of composite materials
and ultimately self-replicate to produce identical and more recently nano materials. For example,
structures to carry out the same tasks. Building at aluminum alloys with nanocrystalline grain sizes
these levels can lead to materials with greater are expected to exhibit yield strengths close to
strength-to-weight ratios, multi-tasking materials 1GPa. Yet in the face of these and other materials
that reduce material and labor costs, ultimately discoveries, the work of A.G. Evans [Evans et al.
making architecture more affordable, sustainable, 2001] confirms, by plotting the strength versus
and more accessible. density of common materials, that the strongest
building materials are, to date, also the heaviest.
In the past few years we have witnessed an In this regard, novel nano materials with high
exponential increase in building designs that can strength-to-weight ratios could, when properly
perform in an environmentally responsive manner. introduced, be in high demand.
In regards to performance of the building we have
what is many times termed green design, or While improving one existing material
sustainable architectural design. These design characteristic is a notable achievement,
approaches focus on the creation of high manufacturing composites with nanostructured
performance sustainable buildings – wherein high- components and integrating these composites as
performance is defined as an economic sound, building materials can go further still to
efficient, and conscious employment of resources conceivably surpass many existing building
in the creation of living, working and playing properties. For example, acoustic and thermal
environments [Hawkes et al 2001]. Norman insulation and strength could be incorporated in
Forster’s recent buildings strive far toward these one sole material: composite materials made from
standards as he works to seamlessly intertwine cork dust reinforced with carbon nanotubes could
design integrity with sustainability and be capable of achieving high tensile strengths and
environmental performance issues throughout his significant ductility, while acting as acoustic and
buildings’ envelopes [Foster 2003]. Thus, while thermal insulators: cork and carbon nanotubes,
‘material smartness’ is already in play, (i.e., in melded into one single material will help reducing
materials wired to sensors can be found throughout man hours by eliminating construction layers due
newly constructed buildings relaying security, to functional integration within one single
temperature data, and accomplishing many material.
disparate environmental task), truly Smart
materials are only now reaching points of possible As shown in Figure 2, material layering can often
employment. increase efficiency. For example, the above
pictured envelope could also be coated with a
Through the use of Smart materials, as defined in photosensitive color layer, such as an
the previous section, new processes of integrating electrochromic material. This material could be
multiple functions into one material will offer engineered to change color as a function of
alternative construction methods allowing for a temperature, shifting from one color to another to
decrease in material and construction costs while better reflect or absorb heat as required (figure 2).
offering state of the art technology and its secured In addition, multi-layered smart materials could
benefits. For example, novel structural combine to create a wall that can ‘breathe’ in
construction materials which are light-weight but response to interior and exterior temperatures. This
exhibit higher strength can reduce the total mass system could be engineered to react at certain
built and offer the possibility of higher, larger, and temperatures, opening or closing micropores
ultimately safer structures. The events of 9/11, controlled by smart shape memory alloys
the collapse of the Charles de Gaulle 2E terminal, embedded in the overall wall structure. In this
recurrent natural disasters and simple economic ‘breathing’ wall assembly the smart material itself
principles of material efficiency have made high possesses functional capabilities; much like human
strength-to-weight ratios a pressing concern. Yet, skin porosity is not controlled by the brain but
while architecture continues to employ the so- rather by its intrinsic functional characteristics
called modern materials of concrete, steel and (Figure 3).
glass, current materials science research has made
great strides toward improving the strength-to-

Figure 2. Simulation of a building surface with color changing paint. (8:am, 12pm, and 4pm) [DaVeiga and LaRoche 2002]
Figure 3. Smart shape memory alloy membranes at micropores for air circulation
through wall
From the examples above we can see how, when

materials contain several functional capabilities, composition and then to assemble them into larger
current construction methods will benefit from structures with unique properties and functions
increased simplicity. Though these are simply a will revolutionize materials and manufacturing.”
few relevant scenarios, they are based on real
material capabilities, waiting only for architects Using smart materials concomitant with
and materials scientists to invent the means to nanotechnology, the future of material selection
express their inherent potential. Designers, in and manufacturing will be both custom and
order to keep pace and create standards, must be dynamic. As aforementioned, the incorporation of
informed about such material processes and nano materials will prove a new and innovative
develop expert knowledge concerning their engineering, construction, and architectural tool.
applications. Materials can now be manufactured in ways that
permit the dynamic manipulation of characteristics
Advancements in Manufacturing such as texture, color, and duress. Materials which
shift their properties can be applied for aesthetic
“Nanotechnology is fundamentally changing the or functional purposes. For example, a wall
way materials and devices will be produced in the covering made of a smart material could
future. The ability to synthesize nanoscale conceivably change its surface texture depending
building blocks with precisely controlled size and on temperature, electric current and other actuating

elements. These chameleon-like properties can Conclusion

also be explored in functional areas such as the
duress of a material; in which case a material could Materials performing multiple functionalities
change its strength or flexibility in response to embedded in the ‘skin’ of the building will reduce
contact depending on the amount of force applied. man hours by eliminating construction layers and
If and when these materials can be grown on-site, complex building control systems. This will
like plants by self-assembly, architecture must be happen as novel Smart and Nano materials can
prepared for the ramifications of such innovation. act without the assistance of computational
methods of sensing and actuating and because
Scientific processes have now located the potential when the material is placed in-situ it will be
for Nano controllers as well as assemblers to self- operative in the environment. However, the
replicate and, once again imitating nature’s examples herewith outlined, if embraced and
processes. As these potentials leave the lab and employed, will bring about a new design
enter the living room, contractors must learn to paradigm. The material capabilities now offered
‘grow’ foundation structures rather than build by the forefront of materials science will require
them. In such a scenario, atomic sized controllers a new kind of understanding and cooperation
can be programmed, under an architect’s guidance between disciplines. Whilst a computationally
and specifications, to grow walls and foundations, manipulated system (i.e. HVAC) can be adjusted
in much the same way as genetic material to fit most building designs as a unit independent
programs the growth of mammals. As in the from the design; the use of smart and nano
abalone, with which we started this paper, proteins materials will require a deeper understanding in
which fill and expand like water balloons in place design and material science. As explained the new
can be stacked like bricks. A complex procedure possibilities require collaboration between
of nourishing rather than constructing these architects and material scientists to solve design
protein walls can lead to structures that grow to specific demands on materials. In a sense, the
reach pre-programmed specifications. Through process of creating new buildings will come closer
growth, rather than construction, Eric Drexler to that of advanced systems in which every part
looked to a day when the self-assembly and self- of the building has an operative nature dependent
growth of residential and office structures would on more than mere aesthetic purpose.
drastically reduce labor costs, enabling both
architect and client further creative freedoms that
are currently bound by bottom line restraints.
While Drexler wrote out dreams in line with the

above fusion of materials research and architecture
in the 1970’s, the direct usage of novel materials,
specifically designed for architecture has not yet
occurred. Historically new materials have played
an important role in the reformulating of new
ideas. For example, Corbusier would have not
been as successful in questioning the idea of
lightness and heaviness at Ronchamp if it were
not for the use of a material perceived as heavy
(reinforced concrete). With this paper, a step has
been taken in that direction, and we hope that in
the present and the near future, architects will
begin to employ the knowledge and skills of
materials development, through the help of
material engineers, in order to custom make
materials according to their needs and predicted
uses.

References Iijima, S. (1991). Nature. London. 354, 56.
Blaser, W. (1997). Ludwig Mies Van Der Rohe. Iijima, S. (1994). MRS Bulletin. Nov. 43-49.
Birkhauser Verlag AG.
Kalay, Y.E.. (1999). Performance-based design.
Chokshi, A.H., Rosen, A., Karch, J. Karch, Gleiter, Automation in Construction (8), 395–409.
H. (1989). vol. 23. pp. 1679-83.
Koch, C.C., Morris, D.G., Lu, K., A. I. (1999).
Cuito, A, Montes, C. (2003). Jean Nouvel, Ductility of Nanostructured Materials. MRS Bull.
Archipockets. Te Neues Publishing Company. vol. 24. pp. 54-58.
Dabbousi BO, Rodriguez-Viejo J, Mikulec FV, Kotov, N. (2001). MRS Bulletin. Dec. 992-997.
Heine JR, Mattoussi H, Ober R, Jensen KF,
Bawendi MG. (1997). Journal of Physical Newman, R.E. (1997). Materials Research
Chemistry. B, 101, p9463. Bulletin, May.
Daveiga, J., LaRoche, P. (2002). A Computer Otsuka, K., Kakeshita, T. (2002). MRS Bulletin.
Analysis Tool for the Design and Manufacturing
of Complex Architectural Envelopes: EvSurf. Feb, 91-100Purdy, G.R., Parr , J.G. (1961) Trans.
Caracas. Sociedad Ibero-Americana de Grafica Am. Inst. Min. Eng. 221, 631.
Digital (SIGraDi).
Reed, M.A., Tour, J.M. (2000). Scientific
Drexler, K.E. (1981). “Molecular Engineering: An American.
Approach to the Development of General
Capabilities for Molecular Manipulation”. Proc. Sanders, P.G., Youngdahl, C.J., Weertman, J.R
National Academy of Sciences. Sept. USA. p5275. (1997). Mater. Sci. Eng. vol. A234-236. pp. 77-
82.
Drexler, K.E..(1986). “Engines of Creation: The
Coming Era of Nanotechnology”. Anchor Press/ Schetky, L. M. (2000). Materials Science Forum.
Doubleday. New York. vols. 327-328. 9-16.
Evans, A.G. (2001). MRS Bulletin. Oct. Sullivan, L.H. (1934). Kindergarten chats on
architecture. In C.F. Bragdon (Eds.), Education
Feynman, R.P. (1960). “There’s Plenty of Room and Democracy. Scarab Fraternity Press.
at the Bottom”. Engineering and Science
(California Institute of Technology). Feb. p22. Taniguchi, N.(1974). Proc. International Conf.,
Prod. Eng. Tokyo. Part II, Japan Society of
Foster, N. (2003). Foster Catalogue. Prestel Precision Engineering.
Publishing.
Ullakko, K..(1996). Journal of Materials
Han, J., Globus, A., Jaffe, R., Deardorff, G. (1997). Engineering and Performance. Vol. 5, No.3. p.405.
Nanotechnology 8. 95. Ullakko K, Huang JK, Kantner C, OHandley RC,
Hardaker, S.S., Gregory, R.(2003) MRS Bulletin. Kokorin VV. (1996) Applied Physics Letters. vol.
Aug. 564-567. 69 No.13, p.1966-1968.
Hathaway, K., Clark, A. E. (1993). MRS Bulletin. Wayman, C.M. (1993). MRS Bulletin. 49.
Apr. 34-41.
Yang, P. (2005) MRS Bulletin. Feb. 85-91.
Hawkes, D., McDonald, J., Steemers, K. (2001).
The selective environment: an approach to
environmentally responsive architecture -
Sustainable Framework. Spon Press.

M. Silver / Discrete Space: Automason Ver. 1.0
Discrete Space: Automason Ver. 1.0

Michael Silver1
1
University of Michigan
Abstract
This paper will demonstrate the power of digital technology by encouraging architects to become more involved
in the creation of small scale, ad-hoc and task specific software tools. Rather than appropriating code originally
designed to solve the visualization problems faced in other fields, designers should instead develop their own
programming culture. Through the acquisition of new skills and better collaborative exchanges architects can
advance the unpredictable desires of the poetic imagination while addressing the practical challenges faced
by craftsmen, engineers and project managers.
This paper will therefore attempt to:
a. Expand the current definition of Computer-Aided Manufacturing (CNC Mills, Laser Cutters etc.)
by proposing a new kind digital construction system,
b. Transform an old, ubiquitous building craft by writing proprietary code that can alter the perfor-
mance of an advanced and ubiquitous communication technology,
c. Describe the basic features of a new architecture.
Note: Unlike classical machines (can openers, clocks and steam engines) a computer’s power is measured by
its degrees of freedom. Because different codes can run on the same hardware without significant changes in
its design a “universal machine” is able to evolve architecture beyond any fixed style or formal paradigm.
Borrowed products that often force the imagination through a filter of hidden protocols and invisible logical
processes currently limit architectural thinking. The ability to craft programs that address both the practical
challenges of building design and the human capacity to imagine new forms has not yet been fully explored
by the individual designer. As code writing becomes easier and machine languages less mysterious proprietary
computing will most likely increase in popularity. Design tools will evolve like musical instruments built to
play sounds heard first in the mind. This will be an unprecedented moment in the history of architecture.

Discrete Space: Automason Ver. 1.0 Contemporary architects are judged as much by
individual buildings as they are by their ability to
mobilize today’s most sophisticated design tools.
“With our artificial automata we are moving much
A certain fascination with technology is natural
more in the dark than nature appears to be with
to any discipline that thrives on innovation. While
its organisms. We are, and apparently…have to
new digital technologies have had an especially
be, much more scared by the occurrence of an
profound impact on the architectural imagination,
isolated error and by the malfunction which is
construction in America and around the world is
behind it.”
still very much a diverse mix of processes, often
supplemented by hand using both traditional and
John Von Neumann from “The General and
non-traditional materials. Instead of focusing
Logical Theory of Automata” (1)
exclusively on the fabrication of building
components in the factory, this essay explores how
information technologies can be deployed in the
“A Space in Architecture shows how it is made”
field to effect sweeping changes in the way
craftsmen work.
Louis I. Kahn from “Architecture is the Thoughtful
Making of Spaces” (2)
While technologies like CNC milling are far from
ubiquitous, the forms and materials associated with
masonry structures have changed very little since
the advent of computer aided design. Of course
portable masonry robots designed for both factory
and on site applications are now in the early stages
of development but there are serious reasons to
doubt that their employment will render human
workers obsolete. In fact, as we look back on the
history of information technology, especially in
the last thirty years, the opposite seems to be the
case. Instead of eliminating work, automation has
forced many to adopt new skills and become
technically specialized as both products and
production processes become increasingly more
sophisticated. Even if android masons do replace
the need for human labor in the future, we still
need to find new ways of enhancing the quality,
efficiency and performance of existing
Figure 1 construction techniques today. The narrow
definition of Computer Aided Design and
Manufacturing (CAD/CAM) must therefore be
expanded to include a much larger set of tools and
procedures. In addition to robotic pipe laying
drones, laser cutters and rapid prototyping
machines we need to include new human machine-
interfaces that operate on site as embedded
technologies that can change the way buildings
are made in the present (augmented craft). This
should be done with an eye on the body’s
connection to, rather than its replacement by,
technology. What’s more we need to honor this
relationship by developing new and powerfully
expressive building forms.
Figures 2 and 3: Block stacking robot and Asimo the Because so much architecture today is conceived
android mason through descriptive techniques like AutoCAD,

digital processes are used mostly for the purpose created in the process are entirely natural to both
of design and representation. Computation is rarely the craftsman and the mathematics. With simple
a direct process responsible for and self-evident programs building details obtain their complexity
in the work itself. To rectify this problem, we are for free, no external agent, author or extraneous
developing a building technology that is based on system is needed to design them. This kind of
the analogous operation of cellular automaton complexity is not dependent on the incessant
programs and masonry construction. differentiation of parts (complexity for a price) but
on the application of fixed rules in a discrete
A cellular automaton program (CA) consists of a system that requires only two components.
field of discreet cells divided into small groups or
neighborhoods. Defined in terms of finite states, With simple programs, designers can push
on or off, transparent or opaque, white or black, masonry to its limits while respecting constraints
etc., a CA computation evolves over time. The both functional and tectonic that define its
condition of each neighborhood in a cellular potential for complexity. The behavior and
automaton is used to determine the state of the intrinsic randomness of certain CA patterns also
next generation of cells. Repetitive and aperiodic, challenges the basic assumption that deterministic
symmetrical and asymmetrical, homogeneous and systems necessarily follow predetermined ends.
heterogeneous patterns emerge as a result of the A brick does not necessarily want to be an arch.
direct relationship between parts acting together When considering emergent phenomenon as
to form a larger system from the ground up. By global patterns produced from the bottom-up by
following local procedures based on laws of local interactions that spread through a system,
adjacency and iteration, masons following a fixed then the “existence will” of an unfolding event,
CA code build by stacking one brick at a time. process or entity becomes meaningful in a way
This process is dependent on the relationship of that an archetype does not. Most patterns generated
each masonry unit to its immediate neighbors and from the iteration of cellular automaton programs
is capable of producing a wide range of forms from generate evolving, teleonomic structures that
very simple rules. cannot be described by a definite shortcut, formula
or ideal type. The only way to know how a given
rule will behave is to set it in motion. In this regard
Henri Atlan writes:
“A teleonomic process does not…function by

virtue of final causes even though it seems as if it
were oriented toward the realization of forms, that
will appear only at the end of a process. What in
fact determines it (i.e. a teleonomic process) are
not forms as final causes but the realization of a
program, as in a programmed machine whose
function seems oriented toward a future state,
while it is in fact causally determined by the
sequence of states through which the
preestablished program makes it pass.”(3)
With a CA masonry system, details are self-

Figure 4: Typical cellular automaton pattern organized whereas the overall pattern produced
by the code forms a tight-fitting whole that is
intentionally selected by the architect. In other
In an automasonry wall, form emerges from the words, the building’s design emerges naturally
direct expression of its materials and the way they from the process of stacking bricks while the
are assembled. This follows one of the guiding overall pattern is constituted in response to specific
principles of modernism but with a difference: design requirements. Again, the code self-
structures driven by simple programs can be organizes the parts while the manipulation of the
constructed without recourse to a limited inventory initial conditions (the starting point of a given
of pared down and platonic forms. The patterns pattern) gives the designer power over the whole.

An architect’s desires, his or her personal reading masonry construction can be developed without
of the client brief, institutional protocols, first considering the body. To do this we must
functional constraints and the logic of construction address the problem of human error.
merge with the rigorous operation of simple
programs to produce an organic architecture that In a wall built with simple programs, anything that
satisfies the imperative of commodity, firmness interferes with the accurate performance of a code
and delight. In this way autopoietic systems like ends up undermining the intended function of the
cellular automata become active agents in the whole. Because an automasory structure expresses
evolution of design. Details exclusively composed function and material use, it must be computed
according to the will of the architect are jettisoned with great accuracy. An effective system of “error
in favor of a self-organizing system that embodies, correction” is therefore needed to insure the correct
in both its function and appearance, the intricate placement of parts in a network of details that are
relationship between artistic intentions and the extremely sensitive to small changes.
systems of life with which they interact. As
Katherine Hayles has pointed out, “…the prospect
of humans working in partnership with intelligent
machines is not so much a usurpation of human
rights and responsibilities as it is a further
development in the construction of distributed
cognitive environments, a construction that has
been ongoing for thousands of years…No longer
is human will seen as the source from which
emanates the mastery necessary to dominate and
control the environment…To conceptualize the
human in these terms is not to imperil human
survival but precisely to enhance it, for the more
we understand the flexible, adaptive structures that
coordinate our world…the better we can fashion
images of ourselves that accurately reflect the
complex interplays that ultimately make the entire
world one system.” (4)
Building without Drawings
“By 2008, analysts estimate, annual world wide

phone cam sales will reach 650 million (up from
150 million in 2004)” one California start-up has
created… a special face recognition software that
runs on the low cost microprocessors used in cell
phones…the company has tweaked its… analysis
algorithms to identify everything from a Coke can
to the Mona Lisa…”
Figure 5. Trio 600 PDA/Cell Phone
Wired Magazine (5)
To make a wall with simple programs masons must
More than just tools for communication, today’s first look at the initial conditions on the ground
cell phones have become powerful computers that floor. The initial conditions are divided into
serve a constantly evolving menu of functions neighborhoods (N). The number of cells in a
including: gaming, navigation, photography, data neighborhood is always constant for a given set
storage and retrieval. By comparison bricks are of rules. In an N=5 computation, the code
primitive objects used to construct stationary walls determines the state of a brick (black or white,
that change very little over time. While the clear or opaque) immediately above cell #3 in a
preceding text hints at a mathematical link between 1-5 unit group. Since masons build by shifting a
craft and technology no instrumental approach to 5-cell frame one step at time, left to right or right

to left, they are forced to check previous moves at 6. Construct accurate mosaics based on hand
least 5 times for each block laid. While this drawings or photographic images following the
approach decreases the likelihood of errors it neighborhood logic of cellular automata.
consumes a lot of energy if workers are required 7. Schedule work, access itineraries and organize
to check rules, stack bricks and evaluate patterns team members.
documented on paper. Less time and effort will
be required if the process is automated. Of course applications like these will continue to
evolve as mobile communication devises exploit,
Using a specially programmed PDA (6) cell phone, new sensing technologies, increased
nonstandard brick patterns can be constructed computational power and better, more efficient
without reference to an external image of the networking capabilities. Passive and Active Sensor
whole. Again the process is entirely local. Networks alone will effect sweeping changes in
Complex masonry work can be achieved without the way buildings are designed, managed and
the need for an equally complicated index of parts constructed. With “an internet of things” (7) where
i.e. unwieldy templates, shop drawings, every object in the supply chain can be tracked
construction documents, etc. But how does a using radio tags the prospects of developing more
bricklayer compare work that takes place in efficient and technically empowered building
physical space with information stored in a practices is more or less insured. The task of the
computer’s memory? How do masons know if an architect today is to incorporate these
error has even occurred? The answers to these developments into a holistic vision of architecture.
questions build off the recursive nature of CA
programs and the ever-increasing sophistication Ornament, Entropy and the Picturesque
of ubiquitous mobile communication technologies.
Here’s how the system works. Pattern recognition For the proposed San Jose State University
software normally designed to identify complex Museum of Art in Silicon Valley, we used a “class
objects is used to monitor simple 3D wall two” CA code (8) to produce both complex and
configurations. The PDA’s onboard camera simple patterns from binary strings of stone and
monitors what’s happing in real space while voice glass block. The location of rooms with windows
commands from both the computer and the mason and exhibition spaces requiring large, blank
determine block-stacking patterns. By relating display walls were laid out in accordance with the
pictures taken in close proximity to a craftsman’s competition brief. Once these parameters were set
body with the neighborhood logic of cellular in place, a search began for a rule that could grow
automata the system is able to prevent errors in the most appropriate form. For the museum’s
construction while supplying the workmen with exterior, internal subdivisions and fire stairs we
accurate instructions for building. An error free used a 5-cell outer totalistic cellular automaton (9)
structure makes CA patterns and their associated that damped out the complexity of the lower floors
functions (fenestration, wall space, stairs, etc.) to create a partly windowless tower with intricate
integral and precise. All of this is done “hands- openings at the base. (The top of the building is
free” and without paper drawings. Using the terminated by a skylight that draws the sun into a
system, builders can also integrate standard PDA narrow atrium facing east.) From the complexity
functions into the workflow making it easy to: of the lower levels the project culminates in a
quiescent and illuminated void.
1. Communicate progress reports to job captains Vertical supports for both the building’s walls and
off-site using mobile email accounts. floors were aligned with the initial conditions of
2. Organize teams to work simultaneously on the CA code. A non-regular grid of columns
different parts of the same wall by forming a produced different spans with beams of varying
wireless network. depth setting up an exchange between light,
3. Store and display construction documents. gravity and computation. Rather than being neutral
4. Read barcodes and track materials. infill the project’s surfaces actively shape an
5. Aid in the resolution, redesign and clarification internal concrete armature that rhythmically
of building details by establishing better fluctuates as the CA patterns ascend into space.
communications between architects and The parametric relationship between enclosing
craftsmen. walls and structure is a tectonic fact experienced

as a system of real interconnections forged during While these patterns are not themselves a product
the process of design and construction. The of organizing principles governed by structural
building is a computation. In the San Jose State laws they are also not applied decoration. In the
University Museum of Art, the nature and position process of unbuilding complexity, class two
of each masonry unit affects its immediate cellular automata drive the organization of
neighbors and the order of the whole. Because the structure and space. Far from a reduced simplicity
system is extremely sensitive to small changes, achieved through the removal of intricate details,
every brick counts in a truly organic architecture blank homogenous surfaces emerge out of
created by the rigorous iteration of simple heterogeneous patterns that negate themselves.
programs. Literally ornament self-organizes its own
disappearance. This approach escapes the narrow
dialectic that opposes formal excess on the one
hand against a strict return to pure forms on the
other. While the results look almost entopic as if
matter was being subjected to erosion, this is not
an architecture of death and decay. Ruins
exemplify the destruction of ideal origins or types;
at San Jose both complexity and simplicity are the
bi-products of a single code. Unlike Site’s Best
department store where shoppers are menaced by
a picturesque simulation of falling bricks, nothing
here is intentionally broken, fragmented or worn
away.
The design and organization of San Jose express

its code’s ability to efficiently produce
asymmetrical patterns that are organically linked
to blank, homogeneous space. These relationships
Figure 6. San Jose atrium perspective
are produced using local rules that are not based
on the recursion of simple motifs, faithfully
Figures 7 and 8 . San Jose State University Museum of Art, wood model and diagrams showing internal space requirements,
selected CA patterns, lighting boundaries, concrete frame and beam to panel variations

rendered at different scales (self-similarity). No forces on a single, deformed surface.

image of the whole can be found in the details.
Neither scale invariance nor the repetition of a While Thom’s catastrophe diagrams are used
standard module can be used to guide the mason’s mostly as a formal devise to exceed the operative
work. While the code for a completed wall can be limits of collage: abrupt changes in the composi-
ascertained through direct observation, the rules tion of a system are not necessarily restricted to
on their own give little indication of the kinds of mathematical descriptions based on an infinitely
forms that will be produced during construction. divisible space. Discrete operations can be equally
Order and randomness, consistency and effective in generating networked relationships be-
inconsistency are therefore binaries that accentuate tween distinct elements. (10) The fold as a leitmotif
the morphological potential of simple programs. for contemporary practice (Eiseman, Gehry et al.)
More importantly, they follow a material logic that requires “a continuous variation of matter” that
organizes structure and space by avoiding the “articulates possible new relationships between
capriciousness of applied decoration. vertical and horizontal, figure and ground…”(11)
The architectural effects of the fold can be replaced
These computational strategies open up by the iterative extrusion of simple programs like
architecture to new ways of thinking and are useful cellular automata where binary codes drive the pro-
as an alternative to the already exhausted tropes duction of complex and simple, random and or-
of contemporary practice. The ability to integrate dered, homogeneous and heterogeneous space. The
different functions and internal space requirements following categories, linked to their current for-
without resorting to antagonistic compositional mal expression give way to a new set of proce-
strategies has traditionally been the motive behind dures:
“folding” in architecture. The work of Rene Thom
is often used in this context to connect opposing
Figure 9. Automason’s operational flow chart

Bifurcation—Folding/Deformation—Simple Programs 2.Louis I. Kahn, “Architecture is the Thoughtful

Making of Spaces” from Architecture Culture
Affiliation—Smoothness/Continuity—Discrete Space 1943-1968, ed. Joan Ockman, Rizzoli, New York,
1993, p-271.
Differentiation—Mass Customization—Complexity for
Free
Variation———Self-Similarity ————A periodicity 3. Henri Atlan cited in Mark C. Taylor’s, “ The
Moment of Complexity” P. 169.
Fabrication—Robotics (CAD/CAM)—Augmented
Craft 4. Katharine Hayles, “How We Became Post
Human: Virtual Bodies in Cybernetics, Literature,
and Informatics”, University of Chicago Press,
Conclusion: How to Make a Building with 1999, p-290.
Stone, Glass Blocks and Cell Phone
5. Jessie Scanlon, “Phones that Get in Your Face”
As a design and production tool the computer is Wired Magazine, December 2004, p-48.
typically used to create folded, curvilinear forms
made from mass-customized parts. These are 6. Personal Data Assistant (PDA)
conventions that have evolved over time as a result
of specific machines and software technologies. 7. A network formed by linking objects together
These technologies operate according to a fixed with Radio Frequency Identification tags (RFID),
logic, one that can be transformed by new codes The tags can be monitored in the field and
developed directly by architects from the ground connected to a database accessed via the
up. Automason Ver.1.0 allows craftsman to worldwide web.
increase the sophistication and efficiency of their
work while affording them new opportunities to 8. For a detailed description of the classification
compete in a culture increasingly dominated by system for Cellular Automaton programs see:
technological change. Through the process of Stephen Wolfram, “Universality and Complexity
software development architects can employ off- in Cellular Automata” from Cellular Automata and
the-shelf technology to design and construct Complexity: Collected Papers, Westview Press,
buildings that express these newly acquired Philadelphia, 1995. Pp.140-157. A Class two
capabilities. Task-specific tools can therefore pattern starts out complex and ends up simple.
affect a much closer link between different
construction processes and the design imagination. 9. An Outer Totalistic set is a shorthand format
For the San Jose State University Museum of Art for specifying Cellular Automaton rules.
a simple cellular automaton program was used to
generate the building’s form. The program drives 10. Discreet Space more explicitly implies the
the placement of columns on the ground floor. joint; a clear division between parts whereas
Changing span depths connect the Earth’s smoothness desires a seamless blending that is
gravitational pull with the operation of an abstract almost impossible to achieve in real buildings
code. The code is rendered in glass block, mortar made from assembled parts.
and stone. Not a window is out of place because
in the building each part affects its immediate 11. Peter Eisenman, “Visions Unfolding”, from
neighbor. To achieve this, Masons will have to lean Zone Incorporations P. 234.
how to build without errors. Humans can do this
only when they establish close collaborations with
intelligent machines.
Notes
1.William Aspray, “John Von Neumann and the

Origins of Modern Computing”, MIT Press,
Cambridge Massachusetts, 1972, p-192.

Automason Ver. 1.0 (Technical Description) computer which in turn records the work, checks
for accuracy, and moves to the next group of cells.
The actual structure of the software used by the Combining the initial indexing procedure and a
mason requires some explanation. A standard PDA step- by- step record of actions the computer
cell phone is hung from the mason’s neck like a knows exactly where it is in space. (This procedure
medallion with its camera facing away from the works even if all the units in a given neighborhood
chest. This position insures that each picture are the same or if a group of blocks are removed
snapped by the camera shows the wall under to make way for operable windows or doors.) As
construction from the same relative angle and work proceeds the computer compares what it sees
perspective. The focal length of the camera’s lens against a complete image of the finished wall
is set at arms length insuring that anything outside stored in its memory. When a day’s labor is
the body’s range of motion is out of focus. complete the mason deactivates the camera and a
(Pictures are automatically taken at 2-second job report is automatically emailed to a pre-
intervals.) The computer’s memory samples and assigned address.
holds only two images at a time. All prior data is
erased. When a mason begins work the phone is
activated manually or by voice command using a
head set and microphone to tell the computer
where the first two blocks have been laid. Images
of the initial conditions for a wall are analyzed by
pattern recognition algorithms that register the
location of existing blocks (Black or white clear
or opaque). The initial conditions are determined
by the position of columns on the ground floor.
The software based on CA neighborhood types

recognizes four primary stacking variations and
their associated joint patterns. With a 5-cell group
for example, the first is a flat placement of two
rows with a pair of blocks on top at position 1 and
2 with 5 units below. The top blocks are placed
when work begins by following instructions given
by the cell phone. The next configuration contains
5 units at the corner of the wall with the same 2
blocks on the second row above units 1 and 2.
The last two configurations contain 9 blocks with
one missing from the first space of the second row
and a full group of 10. The last type occurs when
a single course of blocks is complete and the
mason has finished a complete loop. (Remember
the left side of a one-dimensional cellular
automaton is continuous with its right side.)
By analyzing pictures according to these block

configurations the cell phone then determines the
state of each cell in a given neighbor hood. From
simple shaded images the computer can
distinguish between similar and dissimilar
materials. When a block configuration is
recognized and its cells states determined the
computer calculates the next move following a
specific CA rule. With a synthetic voice command
the cell phone tells the mason which block to place.
When a block is set the mason talks back to the


M. Fox, C. Hu / Starting from the Micro
Starting From The Micro: A Pedagogical Approach to

Designing Interactive Architecture
Michael Fox1, Catherine Hu2
1
California State Polytechnic University / SCI ARC
2
Hong Kong Polytechnic University
Abstract
The paper outlines a pedagogical approach whereby a number of technology-intensive skills can be quickly
learned to a level of useful practicality through a series of discrete, yet cumulative explorations with the
design goal of creating intelligently responsive architectural systems. The culmination of such explorations in
creating full-scale interactive architectural environments leads to a relatively unexplored area of negotiation
whereby individual systems must necessarily manage environmental input to mediate a behavioural output.
The emerging area of interactive architecture serves as a practical means for inventing entirely new ways of
developing spaces, and the designing and building environments that address dynamic, flexible and constantly
changing needs. Interactive architecture is defined here as spaces and objects that can physically re-configure
themselves to meet changing needs. The central issues explored are human and environmental interaction and
behaviours, embedded computational infrastructures, kinetic and mechanical systems and physical control
mechanisms. Being both multidisciplinary and technology-intensive in nature, architects need to be equipped
with at least a base foundational knowledge in a number of domains in order to be able to develop the skills
necessary to explore, conceive, and design such systems. The teaching methods were carried out with a group
of undergraduate design students who had no previous experience in mechanical engineering, electronics,
programming, or kinetic design with the goal of creating a responsive kinetic system that can demonstrate
physical interactive behaviours on an applicable architectural scale. We found the approach to be extremely
successful in terms of psychologically demystifying unfamiliar and often daunting technologies, while
simultaneously clarifying the larger architectural implications of the novel systems that had been created. The
authors summarize the processes and tools that architects and designers can utilize in creating and demonstrating
of such systems and the implications of adopting a more active role in directing the development of this new
area of design.

Introduction Environmental Design Department at Art Center

College of Design in Pasadena, California, at the
The emerging field of Interactive Architecture is Southern California Institute of Architecture,
gaining prominence in recent years. We refer to (SCI_Arc) in Los Angeles, California and at MIT
the term “Interactive Architecture” in the context in Cambridge, MA. The primary objective of the
of this paper as systems that are specific to the course outlined above has been extremely
“tangible” and “physical”, and related to spatial consistent during each offering. The course
“environments” in the broadest sense to include addresses kinetic function as a technological
architecture. The increasing presence of sensors design strategy for building types and objects that
and actuators in domestic contexts calls for the are efficient in form, and inherently flexible with
need of architects and designers to develop the respect to various contexts and a diversity of
skills necessary to explore, think about, and design purposes. The idea is to create spaces and objects
intelligent and adaptive architectural systems The that can physically reconfigure themselves to meet
majority of students of architecture or changing needs.
environmental design however have not been
exposed to this new field. The challenge in Course Design
developing a course on Interactive Architecture
lies in the highly technology-intensive nature of The course (used plurally) was predominately
the subject matter, involving knowledge and skills design-build based and supported by lectures that
that cross boundaries into engineering, computer, provide the necessary conceptual framework. The
and behavioural sciences. course begins with an introduction to theoretical
concepts and precedent in architectural
From the onset, the authors of this paper were applications of Responsive Kinetic Systems. Basic
conscious of the difficulties design students might engineering concepts in mechanical structures
experience. We contend however, that such were also introduced. Simultaneously, students
difficulties are best tackled by removing the were asked to explore various mechanical motions
psychological barrier design students tend to have and joints from found objects and structures that
towards computing and engineering. We believe intrigued them, and then select one structure to
that this is best achieved by having students work examine closely its underlying mechanics. They
on a series of small, explorative hands-on model- were then required to re-build and re-model the
making exercises that are incremental in nature, mechanical structure to replicate and expand its
and which gradually incorporate engineering and basic kinetic capabilities.
computing components. Based on this approach,
the authors have developed a course to enable Next, students were introduced to basic concepts
students to have hands-on experience in designing in electronics, as well as BasicStamp, a
Interactive Architectural environments. This paper programmable IC chip with an integrated circuit
reports on the design, delivery, and project for the incorporation of sensors and motors. While
outcomes of this course taught at several different learning to work with BasicStamp, students were
Universities of both architecture and design to both asked to think about applying the motors to their
undergraduate and graduate students. The authors mechanical joint explorations, as well as using
also argue that when the design tools evolve sensors to trigger the motion. At this point, the
together with the developing design concept, concept of behaviours was introduced, and
crossing of boundaries becomes easier, facilitating students were asked to both design and rationalize
the acquisition of new knowledge in other the intended behaviours of their mechanical
domains. structure, and apply these interactive behaviors
towards an architectural application. It is in this
Developing Skills for Thinking manner that the students’ initial model explorations
gradually grow in complexity, integrating first
Course Objectives automatic functions, and later, more complex
autonomous behaviors, and lastly architectural
This course was conceived to develop the skills applicability and conceptual insight.
necessary to explore, think about, and design and
has been taught at the School of Design of the
Hong Kong Polytechnic University, in the

Strategies for Delivery system, then develop the application of this system
in the larger context of use. The objective behind
Most of the topics on electronics, computation and this methodology was to allow students to focus
mechanics were covered only in a very basic, on the core of the responsive system itself, that is,
introductory manner. The approach used in this the fundamental mechanical structures and how
course was different from typical design courses to embed this with interactive and intelligent
in that rather than starting from macro: finding a behaviors. This approach also minimizes the
problem, then research, then design, we started daunting psychological effects students might
micro: designing the mechanical structures first, experience if from the onset, they were told that
then ‘grow’ the system by adding sensors and they have to develop a structure or building that
motors, then designing the behaviour of this is intelligent and that can physically demonstrate
Figure 1. Two-D joint exploration Figure 2. Three-D joint exploration
Figure 3. Assembled structure Figure 4. Final kinetic model
Figure 5. Chair module (half-risen) Figure 6. Table module (fully raised)

interactive behaviors. By going micro, students control and realistic operating conditions. The
did not feel the pressure of confrontation by model incorporated sensors and motorized control
venturing into unknown territories. This hands- in order to prototype the behaviours of the systems
on demystification process is very important for as opposed to simulating them. The project is a
crossing interdisciplinary boundaries. The specific application scenario that actually affects
challenge here lies in having the students learn a the nature of the architectural construct (the room).
minimum, in order to be able to achieve a greater Specifically, the design project is a networked
sum than the additive parts, in the sense of being system of individually responsive chairs that
able to embed kinetic structures with behaviors function together, transform from an invisible floor
that demonstrate, as opposed to simulate, one’s module to a congregation of six chairs when half-
design ideas and intents in intelligent and risen, or a table when fully raised. Primary design
responsive systems. Through discretely gaining considerations were to affect the physical space
elementary interdisciplinary confidence, the and to create a module of furniture that is adaptable
ultimate goal is for students to be able to overcome both to the number of users as well as to be
psychological barriers and have the confidence to physically present when it is necessary and
communicate with engineers and programmers invisible and out of the way when it is not in use.
their design intentions in an effective manner.
Such confidences can have a strong effect in Design Example Two: Caterpillar Kiosk
architects taking a more active role in directing
the development of interdisciplinary design The Caterpillar Kiosk project has employed a
directions. To do this, designers need to have at mixture of prototyping tools and techniques from
least a superficial knowledge base of both the early mechanical joints explorations throughout
engineering in terms of mechanics and fabrication to final design implementation. It demonstrates
and also the computational substructures in order how tools and design co-evolve, with the tools
to develop the necessary skills and the conceptual facilitating the design process as opposed to
and intellectual framework for designing. directing it, an argument that will be further
discussed in the paper. The final design is a hybrid
Application Examples: Scaled Architectural structure: the core mechanism, responsible for the
Mechanics and Robotics piston-like motion, is built with fabricated
materials; this is further integrated with, and also
Several student projects are outlined below that affecting, the navigation mechanism built with
illustrate the methods and processes used in LEGO gears and motors. Primary design
designing interactive systems at a scalar level for consideration was to provide a temporary shade
understanding mechanical motions. These and shelter in areas where it is most needed, and
particular projects are all when it is needed. The idea is to embed this mobile
from a course taught at Hong Kong Polytechnic caterpillar structure with both light sensors and
University. Similar scaled exercises are also motion sensors, such that it will navigate
carried out as for initial skill-building in the autonomously towards the direction of strongest
context of the full-scale projects outlined later in UV index, but will stop once it senses human
this paper. motion within it. When it is not moving, people
can then gather inside the structure and enjoy being
Design Example One: Kinetic Chair/Table Systems shaded from strong sunlight or even heavy rain.
Application environments are perceived to be vast
This interactive design project for kinetic chair/ open areas like parks or beaches. While not in
table systems demonstrates the simplified motion, the caterpillar structure could be used for
prototypical kinetic attributes that first grew from other temporary activities like public
a simple exercise in mechanical design. The performances, or as a temporary catering facility.
simple mechanical model of cardboard that The final model here also incorporated sensors and
demonstrated the motion grew to a precise motorized control to in order to demonstrate
mechanism with gears and motors and sensors. through prototyping the behaviours of the systems
The design in its final stage retained the base as opposed to simulating them.
mechanical principles yet grew discretely to
demonstrate a full range of attributes relative to
kinetic function, human interaction, adaptive

Figure 7. Early mechanical joint explorations with LEGO Figure 8. Early mechanical joints explorations using LEGO
parts parts.
Figure 9. LEGO parts replaced with fabricated steel and Figure 10. Final model incorporating sensors and motors
acrylic modules
Additional Scaled Design Examples and that architectural expression may change as
more designers explore, design and familiarize
Several additional projects are shown below that themselves with such systems.
demonstrate a wide range of application examples
in architecture. The applications grew out of the Application Examples: Full-Scale Architectural
simple mechanical explorations with varying Environments
degrees of complexity. A diverse range of
prototyping materials were also employed in these Several student projects are outlined below that
final models in order to clarify architectural issues illustrate the methods and processes used in
of structure, transparency, lighting, etc. Typically designing full-scale prototype systems that can
it was desired that the computation, (control boards demonstrate interactive behaviours.
and sensors) be seamlessly embedded into the
structures rather than being openly expressed as Design Example One: Interactive Restaurant
an architectural quality, and that the mechanics Design Process
and motorization be openly expressed. We find
it important to note these differences in attitudes This interactive design project for an interactive
towards mechanics and computation and although restaurant was a collaborative project of an entire
such differences are not quantified here, we note studio. After students have a practical grasp on
that perhaps they are a result of the familiarity and the basic mechanics and robotics they begin to
acceptance that students have with mechanics in collaboratively design a full-scale interactive
the real world as opposed to things computational environment. (in this case an interactive restaurant.

Figure 11. Wall system combining intelligence with Figure 12. Transparent materials employed for intelligent
traditional skyli
Figure 13. Exposed mechanics with hidden computational Figure 14. Complex behaviour hidden behind simple
con architect
Figure 15. Intelligent solar reflectors with hidden mechanics Figure 16. Full scale prototype of intelligent solar reflectors
and computation
Specifically, the design project is a networked cam fabrication, to standard construction

system of individually responsive systems that techniques such as welding and timber framing.
function together as an architectural space. In almost every instance, design modifications are
Students in the classroom learn to go between made as a result of misunderstanding the physics
various aspects of realization from robotics to cad- of full-scale mechanics.

Figure 17. The use of computers for programming initial Figure 18. The use of CAD-CAM for fabrication
behaviours
Figure 19. Typical classroom under construction Figure 20. Wiring in sensors to traditional timber framing
for an acoustic wall
Figure 21. Construction hardware and computer software Figure 22. Typical hand-made circuitry
Design Example One: Interactive Restaurant as process did the realization come that the input for
a Functional Architectural Environment the individual systems should not be a direct
relational input t of human behaviours but rather
As the level of completion of the architectural the translated and cumulative effect of behaviours
environment as a whole neared completion we on other systems in the space. The façade reacted
found that a great deal of negotiation occurred in to people outside of the space, the interactive bar
terms of sensing and response. The space had a reacted to people preparing to enter the space, the
number of individual systems that were designed floor reacted to the people in the space, the ceiling
as independent systems and only later in the reacted to the floors and the vanity reacted to the

Figure 23. Aesthetics of the robotics are expressed as design Figure 24. Aesthetics of the spatial interactivity are expressed
elements as design elements
Figure 25. A chair inflating where and when a user decides Figure 26. Chair and ceiling system working together
to be seated
Figure 27. Pneumatic control manifold expressed as a design Figure 28. Interactive waiting lounge dealing with
element psychological proximities
people temporarily leaving the floor and translates Design Example Two: Interactive Spa/Urban
this to the façade and to people outside of the Retreat (iSpa) Design Process
space. Thus completing a cyclical and synergistic
effect of behaviours wherein each project is a key This interactive design project for an interactive
component in the larger system. spa/urban retreat was another collaborative project
of an entire studio. After students have a practical
grasp on the basic mechanics and robotics they
begin to collaboratively design a full-scale
interactive environment. (in this case an interactive

spa/urban retreat. Students once again learn to go wiring circuits, programming, lighting, and
between various aspects of realization from acoustics serve as a tremendous exercise in
robotics to cad-cam fabrication, to standard bringing to light a great number of unforseen
construction techniques such as welding and issues both in terms of mechanics and electronics
timber framing. In almost every instance again that build the skills and confidence to
we found that design modifications are made as a communicate as designers to expert fabricators and
result of misunderstanding the physics of full-scale technicians.
mechanics. We believe that the hand-on
experience of making mechanical connections,
Figure 29. Wiring in the classroom Figure 30. CAD-CAM assembly in the classroom
Figure 31. Welding Figure 32. Making of a hay-bale wall
Figure 33. Fabrication of metal interactive wall Figure 34. Mechanical motorized counterweight made from
found objects

Design Example Three: Interactive Spa/Urban For instance, as the floor system began to become
Retreat (iSpa) as a Functional Environment very difficult to walk upon, a path had to be carved
within the space for circulation. As another system
As the level of completion of the architectural required low lighting levels, it required the
environment as a whole neared completion we function of the lighting system to become a
found once again that a great deal of negotiation navigational system for circulating throughout the
occurred in terms of sensing and response. The space. Once again, as realistic conditions were
space had a number of individual systems that developed in a design-build fashion, new problems
were designed as independent systems and only evolved that required the modification of existing
later in the process did the realization come that concepts to complete the interactive architectural
the input for the individual systems should not be environment. We believe that such issues of
a direct relational input of human behaviours but interactivity in terms of real human behaviours and
rather the translated and cumulative effect of response are very difficult to design and are more
behaviours on other systems in the space. This clearly understood only as they evolve in a holistic
project evolved to a degree that both acoustics and fashion. The group designing of the system as a
lighting became major factors in the overall whole then makes explicit the negotiation that
design. Most of the projects were dealing with must occur in designing interactive spaces and
psychological and sociological issues that could environments with inclusive architectural
only be understood and translated through a considerations as opposed to objects or stand-
combination of systems as opposed to any singular alone systems.
system.
Figure 35. Full-Scale environment dealing with typical Figure 37. Cleansing interactive entry
spatial division
Figure 36. Full-Scale environment dealing with 3-d spatial Figure 38. Biorhythmic light room
division

Figure 39. Acoustic cylinders Figure 40. Entire environment negotiating light and sound
Design Example Three: Interactive Zoo (iZoo) exercise in bringing to light a great number of
Design Process unforseen issues both in terms of mechanics and
electronics that build the skills and confidence to
The iZoo is an interactive architectural space that communicate as designers to expert fabricators and
fosters human connections, floods the senses, is technicians. We found an increasing number of
educational and manipulative. The design project students that found ways to make things work that
is the collaborative effort of 22 SCI_ARC students were not the “correct” way but that did in fact
for a course in “Interactive Environments”. This work. For instance rather than learn to program
course was once again a combination of full-scale- a voltage regulator to dim a lighting device, the
hard-core-construction, CAD/CAM manufacture, student used traditional lighting dimmer switches
robotics, and the translation of robotics to full- that were controlled by servo-motors that he had
scale real works to create a space with behaviours a good grasp on and the project worked perfectly.
that you can interact with. After students have a Another student could not get an external amplifier
practical grasp on the basic mechanics and robotics to work correctly with the power that was supplied
they begin to collaboratively design a full-scale with the BasicStamp yet wanted a much so rather
interactive environment. (in this case, based upon than solve the technical computational problem,
the conceptual ideas inherent in a zoo. Students he simply built an acoustic box around the output
once again learn to go between various aspects of that he had and placed a microphone on the
realization from robotics to cad-cam fabrication, speaker output that he had and ran that through
to standard construction techniques such as the amplifier and once again was able to get a very
welding and timber framing. In almost every nice output. These means of problem-solving are
instance again we found that design modifications quite unconventional for traditional students of
are made as a result of misunderstanding the architecture but are invaluable in learning to think
physics of full-scale mechanics. We believe that of ways to get a desired result with what is
the hand-on experience of making mechanical available both physically and technically and are
connections, wiring circuits, programming, often more valuable than learning a direct technical
lighting, and acoustics serve as a tremendous skill and applying it to get a known result.

Figure 41. Combination of wiring, programming and full- Figure 42. Laptops for programming are ubiquitous in the
scale construction classroom
Figure 43. Key aspects of negotiation and circulation Figure 44. Dealing with 3-d spatial negotiation
Design Example Three: Interactive Zoo (iZoo) as mechanics. There was a great deal of negotiation
a Functional Architectural Environment for space for individual projects that included
issues of lighting and acoustics. A number of the
This project stands out from the previous two projects had specific needs in terms of these issues
examples in that the intent of the interactivity was that differed from adjacent projects and
not focused upon practical or functional problem- transitional areas had to be collaboratively worked
solving but was aimed at articulating the out between them. Also worthy of note is the fact
conceptual ideas inherent in a “zoo”. Conceptual that the design was initially an open-plan whereby
ideas such as Manipulation, Learning, Roll- visitors could wander as in a typical art gallery
Playing, Privacy, and Human Connections were from exhibit to exhibit. This concept was changed
the stimuli here for developing the individual rather late in the process out of the need to clarify
projects. In almost every instance again we found each piece with a physical dedicated space that
that design modifications are made as a result of was off of a linear circulation route. This change
misunderstanding the physics of full-scale allowed the individual projects have increased

Figure 45. Large scale lighting requires collaboration of Figure 46. Spatial construction at a human scale
many individuals
Figure 47. Circulation was defined out of necessity near the Figure 48. Shared lighting and acoustic needs were
end of the project vigorously negotiated
Figure 49. Although visitors exceeded design expectations, individual projects developed means to clarify spatial needs

clarity for dedicated understanding and interaction Evolving the Design Tools with the Design
yet increased the difficulty for making a cohesive
space. The issues involved in the making of a This paper also argues that the design tools (most
cohesive environment in terms of lighting of which were borrowed from other disciplines)
colouring, textures, and acoustics were of great used in the design and prototyping processes, plays
concern to everyone involved. They were not an important role in the evolution of the design
completely defined until quite late in the process itself. In particular, how the tools used in
and after the circulation had been changed. We prototyping behaviours with real-time feedback
feel however that within the context of a more influences the processes utilized in designing, and
generous design time frame that they would have consequently, have a profound effect upon that
been worked out at an earlier stage in the process which is actually designed. The integration of such
as a guiding principal. tools for simulation (rather than representation)
into the process of designing lies in an
understanding of anything as malleable to
idiosyncratic designing needs. In computing, this
is comparable to chunks of codes easily becomes
reusable ‘tools’, facilitating individuals in
developing other codes. The tools as used in
teaching this course in Responsive Kinetic
Systems, we argue, develop new heuristics if such
methodologies can in turn be tracked; they can
more aptly facilitate novel tools. The tools and the
design, on the generalized level as well as the
specific, ought to evolve together. When we are
Figure 50. Behaviours had to be learned with intuitive designing with a tool, the heuristics of the process
instruction in order to negotiate the space are thereby directed through the affordances and
limitations of the tool. However, when the tools
used evolve with the design, the heuristics are
facilitated by the tools, and not necessarily limited
by their parameters. Process then is directed or to
motion capabilities for their intended structures
along the way. This co-evolution of the design with
the design tools has in part, been made possible
by the lack of preconception to limitations of the
tools themselves on behalf of the design students.
In this sense, designers very naturally ‘break the
rules’ and consider the tools as pliable as the
developing design concept. This in turn has
Figure 51. A remotely controlled project through the internet actually facilitated the crossing of boundaries and
subsequent acquisition of hands-on knowledge in
the said domains.
Architects Directing the Development of

Interactive Architecture
Building is a hugely complex endeavour and it is

not possible to design a building without
consulting many specialists (architects, engineers,
construction managers, lighting consultants,
mechanical engineers, acoustical experts, financial
advisors, and legal experts, etc.) [Cuff 1991] But
Figure 52. Small-enclosed spaces allowed for clarity of collaboration is difficult as each specialist comes
individual projects within the overall context of the exhibit from a different educational foundation [Kalay
space 1999], and has goals and criteria and methods that

are different from others. Intelligent responsive dynamic, flexible and constantly changing needs.
architecture, being a more complex building type, It is therefore, timely for architects and designers
will require the collaboration of even more to equip themselves with foundational
specialists. However, the heterogeneous understandings in domains like engineering and
backgrounds of the participating professionals in computation, in order to assume a leading role in
the building industry are often a source for helping to shape this future. In fact, the evolving
misunderstandings and misinterpretations of the design thinking is one of holistic and experience-
communicated information, leading to errors and based, with the user’s needs and experience taking
conflicts. [Kalay 1999] central stage. That is, the success of intelligent
responsive architecture not only lies in designers
To overcome this, the paper proposes that becoming more fluent with technology, but also
architects should take courses on simple having a paradigm shift from ‘space-and-flow-
mechanics and computation in order to accumulate conscious’ to ‘human-need-and-experience-
a superficial knowledge base in these domains. conscious’; from the mindset of designing a library
This will enable the architect to share the building, to a mindset of designing the enabling
perspectives and general concerns of other factors to support the experience of acquiring
specialists, and to better communicate his design knowledge. The goal for the designer is to have
intentions, ultimately facilitating better enough skills in the areas of interactive design to
collaboration amongst the team. Further, we are communicate design intentions through first-hand
really at a point in the profession where intelligent knowledge. Such communication skills can have
responsive systems are possible and even feasible a strong effect in architects taking a more active
from an economic standpoint. It is both timely role in directing the development of
and important that architects should take on a more interdisciplinary design directions. To do this,
active role in directing the development of this designers need to have at least a superficial
area of design. The idea is not for architects to do knowledge base of both the engineering in terms
structural calculations on a building, nor to of mechanics and fabrication and also the
develop the computation that controls the computational substructures in order to develop
behaviour of a responsive system. The traditional the necessary skills and the conceptual and
role of the architect will not change, but he will intellectual framework for designing. Changes in
have new roles of engineering and consultancy, conventional thinking about what an environment
defining and designing the next generation of can do as opposed to what an environment is will
responsive buildings. ultimately revolutionize the way we use space.
Conclusion
This paper has described a method for teaching

Interactive Architecture as an exemplar of a
pedagogical approach to nourish the future
generation of architects and interactive systems
designers. Design in the new millennium, whether
it is a building or an object, will inevitably be
increasingly technology-based. Architectural
environments will increasingly be smart and
responsive and capable of complex behaviours.
The motivation lies in learning to create dynamic
environments that can physically re-configure
themselves to meet changing needs. The goal is
to make spaces that behave, respond, interact, and
adapt like human beings. The central issues
explored are human and environmental
interaction, embedded computational
infrastructures and kinetic engineering. At the
intersection of these areas exists a widely
unexplored area of design tuned to address today’s

References
Chironis, Nicholas P. (1996) Mechanisms and

Mechanical Devices Sourcebook, Mcgraw Hill.
Clarke, Arthur, C (1964) Profiles of the Future,

Harper and Row Publishers, Inc., New York, NY.
Coen, M. (1999) The Future Of Human-Computer

Interaction or How I learned to stop worrying and
love My Intelligent Room. IEEE Intelligent
Systems.
Coen, M. (1997) Building Brains For Rooms:

Designing Distributed Software Agents. American
Association for Artificial Intelligence.
Cuff, D. (1991) Architecture, the story of practice,

MIT Press, Cambridge, MA.
Davidson, Cynthia (1995) Guy Nordenson, Chuck

Hoberman, Mahadev Raman: Interview: Three
Engineers (Sitting around Talking) Any:
Architecture New York, v10, Anyone Corp, New
York, NY
Fox, M. A. (1996) Novel Affordances of

Computation to the Design Process of Kinetic
Structures, Thesis M.S., Massachusetts Institute
of Technology, Cambridge, MA.

M. Clayton / How I Stopped Worrying and Learned to Love AutoCAD
How I Stopped Worrying and Learned to Love

AutoCAD
Mark J. Clayton1
1
Texas A&M University
Abstract
The history of computing is expressed through AutoCAD as an accretion of ideas and inventions, each of
which was a breakthrough in its time. Learning to use AutoCAD, or any CAD system, is augmented by an
understanding of the historical context of its development. In contrast to a “deconstructivist” criticism of
AutoCAD that avoids all historical context, this paper discusses the user interface of AutoCAD placed in its
historical context by combining facts of history with personal reminiscences. The paper answers mysteries
about AutoCAD such as “Why a black screen?”, “Why LISP?”, “Why a command line?”, “Why layers, pens
and line types?”, and “Why 2D?” An understanding of context and history is a starting point for understanding,
mastering, and improving software.

Mortality and the Recognition of History than subversive. There are reasons, some
historical and some practical, behind all of the
It is a peculiar sensation to realize that one’s own idiosyncrasies of AutoCAD. It is more profitable
experiences and past have become history, to use the idiosyncrasies to explore the world of
particularly from the vantage point of merely ideas clustered under the rubric of architectural
middle age. In these times of rapid change, the computing.
old and familiar rapidly become the strange and
forgotten. Computing has grown from non- I suggest that the deconstructionist critique suffers
existence to pervasive impact upon every aspect from an impoverished concept map due to its
of human life in a time period of less than the insistence on disregarding evidence from outside
lifetime of an individual. PC CAD (an acronym the artifact in question. At the risk of
that may already be slipping into the obscuring oversimplification, I present a concept map in
mists of history), spans little more than a human Figure 1 to summarize Senegala’s argument. The
generation, and in my case, neatly encompasses self-referential approach is limiting and
my professional life. When I was entering college, inadequate. Context is important. I will
the personal computer was being invented. By progressively develop a richer concept map that
the time I finished college, AutoCAD had reached will lead the reader toward ever expanding circles
the market. In my first jobs as a professional of exploration. In the conclusion, I hope I can
designer I became a CAD manager. I started some present convincingly that the “subversion” in
of the earliest courses in architectural computing. AutoCAD is actually in the mainstream of
I pursued graduate studies to increase my architecture as a discipline, practice and
knowledge and hone my skills at the cutting edge profession. The view of architecture expressed in
of architectural computing. Along the way, I AutoCAD clashes not with the mainstream of the
learned to use AutoCAD, Microstation, profession but with a limited and limiting view of
VersaCAD, Dynaperspective, DataCAD, architecture that might be called “the cult of the
ProDraft, PowerDraw, MiniCAD, ArchiCAD, sketch.”
FastCAD, Arris, Sonata, Reflex, EasyCAD,
ComputerVision, CATIA, and many other Perhaps my approach is post-modernist rather than
forgotten CAD systems. AutoCAD is still with deconstructionist, as I will employ an attention to
us, and in fact dominates the contemporary history. I attempt to ground the discussion in
architectural tools market. It has competed context and integrate an extensive range of ideas.
successfully in the race to innovation and adoption I prefer to think of this paper as a documented
during that entire time. Under the hood and in the oral history, personal memoir, or even a testimonial
undercarriage is an entire history of computing. of a personal journey to enlightenment. Perhaps
Examination of that history can give the reflective such an approach is inevitably more gentle,
student, those who ask not only what and how, reflective, and compassionate than the pretended
but also why and when, a way to reach dispassion yet superciliousness that seems
understanding that can be the basis for progress. characteristic of deconstructivism.
This paper is presented as a counterpoint to an

article prepared by Mahesh Sengala that
“deconstructs” the user interface of AutoCAD
(Senegala 2004). That article accuses the makers
of AutoCAD as either consumed by ideological
subversiveness or being “dim-witted” (305). He
supports this idea by reference to characteristics
of the AutoCAD user interface. Senegala suggests
that subversive characteristics include the
command line, the “black void” screen, the
inadequate treatment of line weights, the
preference for 2D, and the disorderly collection
of icons, menus and keyboard commands.
However, AutoCAD employs people who are Figure 1. Concept map of the view that AutoCAD subverts
certainly not dim-witted and are more pragmatic architectural traditions

A chronological personal memoir would be a self- Our work was merely a moment in a continuous
centered indulgence. Alternatively, I have dialogue about computing that was initiated at
organized this paper into five themes that combine Cambridge in the late 1950s (Keller 2005). Before
Senegala’s points of criticism with historical the invention of computer graphics, architectural
context and personal reminiscences. These themes computing explored optimization, space layout,
are 1) computational design; 2) evolution; 3) proportion, pattern, and other topics. By the
overlay drafting; 4) 2 or 3D; and 5) disorder, 1970’s, computational design was producing
information and process. The final section large, complex buildings such as hospitals that
summarizes these themes to expound an empirical were demonstrably superior to those designed
theory of architecture generated from the forty- with conventional methods (although aesthetically
year discourse on computing and architecture that unacceptable to the general public) (Hoskins
has generated AutoCAD. 1977). In truth, the dialogue is much older than
computing and includes many of the preeminent
The Command Line and Computational Design architects in history, as thoroughly expounded by
Mitchell (1990).
Any computer system presents strange,
idiosyncratic, or cryptic ideas and languages to The AutoCAD command line is a reminder and
the new user. I have watched students struggle with acknowledgement of the contributions of that era.
MKDIR, grep, or dragging a floppy disk to a By expanding the command line one can scroll
trashcan. With experience, confidence and through an entire AutoCAD session to uncover
familiarity increase, and the operations become the sequence and precise steps in creating a
obvious and comfortable. Although the command drawing. One can use that record to gain insight
line in AutoCAD is much maligned as antithetical into design process or generalize a session into a
to a graphic sensibility, for an experienced repeatable automated sequence.
AutoCAD user it is a component of a highly
optimized, two handed user interface. The use of Through the command line, Auto LISP is always
the space bar as equivalent to the “Enter” or ready to spring into action to automate a repetitive
“Return” key is a brilliant feature that eliminates task or apply reasoning to a design problem.
handedness; AutoCAD is equally usable (or Invented in 1958 as one of the first high-level
unusable) by left-handed people as right-handed programming languages, LISP is still the preferred
people. language for much artificial intelligence
investigation (Spencer 1997). For many years,
It is easy to mistake the command line as Auto LISP was a common tool for architectural
counterintuitive or anti-graphical. It is also easy investigation and remains a quick and dirty way
to hide. However, it is also a persistent and to manipulate data and graphics in AutoCAD.
beautiful reminder of an era when architecture was Auto LISP may be a decisive factor in the
not graphical, and CAD was not a branch of dominance of AutoCAD as it enabled the sales
computer graphics. My introduction to force to tailor AutoCAD for the needs of specific
architectural computing as a formal subject of clients. It is puzzling why other entries onto the
study comes from spending 1986 and 1987 at CAD market have neglected the inclusion of
UCLA. Bill Mitchell, Lionel March, George Stiny, macro languages. A command line is an asset,
Murray Milne, Robin Liggett, and Jeff Hamer not a liability.
espoused a commitment to “computational
design.” Fellow students and I learned to program The Black Void and Evolution
as a means to generate design from functional and
aesthetic criteria. We programmed windows and When AutoCAD starts, the default screen is a
facades that could be controlled in pattern and black field with no indication of size, or direction.
proportion by use of parameters (Mitchell, Liggett Senegala describes this “black void” as another
and Kvan 1987). We wrote analysis programs for subversion of an architectural sensibility. Of
space layout and technical analysis, including course, one can easily change preferences and
Solar 5, Opaque, and Climate Consultant (Moore, defaults so that AutoCAD opens with a white
Milne and Geier-Wilson 1993). We devised shape background, appropriately sized grid defining the
grammars for Palladian Villas, Prairie Houses and ground plane, and perspective view, as illustrated
numerous other building types (Mitchell 1990). in figure 2. However, the black void has historical

associations that help us remember where we were, The question of how to interact with a computer,
where we are now, and where we may be going. particularly for graphics, was wide open. In 1984,
I was an intern-architect assigned largely to CAD
AutoCAD was announced as a software product drafting. The firm leased one workstation of the
in 1982, one year after the introduction of the IBM ProDraft system by Bausch and Lomb for upwards
PC and two years before the Apple Macintosh of $80,000 per year. The computer was dedicated
(Spencer 1997). The innovation of a hard disk in to CAD and in fact could run no other software.
a personal computer was still one year away, It used a proprietary operating system. Input was
awaiting the release of the IBM PC-XT. It was a through a keyboard and a digitizer tablet; the
startling software “hack” to fit a full-featured CAD mouse had not yet been introduced to the popular
system onto a 360kb floppy disk to run in less market. There were no pull-down menus or on-
than the 640kb limit of RAM enforced by MS- screen icons.
DOS. In that era, most computing was done in
text mode; it took a special display and a special In 1987 when I joined the faculty at Cal Poly, there
graphics circuit board to display graphics. A CGA had been steady progress in graphics software, but
card displayed 4 colors at a resolution lower than no earth-shaking changes. Our computer labs
that offered by my cell phone. Word processors contained PCs and PC-XT’s with a few PC-AT’s
operated in text mode sans any font display or page for experimental purposes. The state of the art
layout capabilities. Screen space was very was AutoCAD, VersaCAD, and DataCAD, each
precious. I suspect that the prototypes of AutoCAD of which employed screen menus for user
were written to allow the user to interact in text interaction and a mouse as graphical input device.
mode and then switch to graphics mode to view Figure 3 approximates the look and feel of
the results. AutoCAD from that era, as obtained from
adjusting preferences in AutoCAD 2005.
Figure 2. A 3D modeling setup in AutoCAD

The black void screen, to one who knows the lines is symbolic, not aesthetic or expressive.
history, is an expression of economy and AutoCAD derives from a long graphic tradition
minimalism. It is a reminder of the cleverness that prizes clarity and meaning rather than
and discipline that was required to achieve PC appearance. Once again, the historical perspective
CAD. AutoCAD out of the box is a neutral tool helps to clarify one’s understanding and perhaps
that provides a tabula rasa for the exploration of increase one’s ability to use the tool.
design ideas, but if you don’t like it, then change
it. An appreciation of the immense computing In 1978, drafting was an overwhelming part of
power of current computers in comparison to those architectural practice and education. As a student
of ten or twenty years ago may temper the at Tulane, I had an enormous advantage in that I
profligate wastage of such power due to shoddy could not only draw but I could draft. My lettering
modeling. There is value in educators focusing was terrible, but so was that of everyone else. I
upon how to build compact, efficient computer understood perspective, something that set me
models that can fit into modest computer apart from 95% of my classmates. Very few
resources. students ever attempted a perspective, and the
instructors generally encouraged us to work in
Line Weights and Overlay Drafting plan and section. The long nights of work in
studio were spent in producing the final drawings
The ascription of subversiveness to the for presentations, which were mostly drafted
implementation of line weights in AutoCAD can using ink on vellum and consisted of 1/8" = 1’ or
only be justified by ignoring the purpose and ¼” = 1’ plans, sections and elevations.
origins of the software. The graphic language
employed in AutoCAD is not that of fine art The faculty had modest drafting skills; they could
drawing and sketching, but that of mechanical draw, but they knew nothing about emerging
drawing. In mechanical drawing, the quality of technology for documentation that was poised to
Figure 3. Simulation of AutoCAD circa 1987 using settings in AutoCAD 2005

revolutionize practice. Using the new tools, the their day. Overlay drafting became the model for
time needed to draft a project could be reduced CAD systems. AutoCAD employs a typical
by an order of magnitude. This technology was overlay drafting metaphor, as does nearly every
not the computer; it was overlay drafting. CAD system on the market. The concept of layers
Widespread in the manufacturing industry, the became the dominant and ubiquitous way of
technology was adopted slowly by the organizing graphic information. The plastic
architectural professional. The accidents of my tracing templates also received a digital
personal experience introduced me to the counterpart, variously called symbols, blocks,
technology that became the foundation of the CAD templates, or cells in different CAD systems.
metaphor that we know today.
The overlay metaphor for digital representation
While working in Washington, DC in 1980, of drafting has proven effective in comparison to
draftsmen introduced me to the tricks of their craft, manual drafting. It is nevertheless a clumsy way
such as lead holders, micro-leads, rapidographs, to represent a building. Buildings are not made
parallel rules, plastic templates, drafting machines, of blocks, layers, line styles, and line weights, nor
triangles, and French curves. I learned to draw Xrefs, paper space, or sheets. The traditional
on both sides of a sheet to make alterations easier. drafting representation is itself a metaphor for a
I became proficient at using a Leroy set for building. Nearly all CAD systems, AutoCAD
lettering. Through friendships with the blue-collar included, enforce a metaphor of a metaphor onto
workers in the printing companies, I received tours the hapless architect.
and demonstrations of the equipment that gave
them pride. They showed me vacuum press Like all metaphors, the idea of overlay drafting
blueprint machines and full-size photography breaks down when translated for the computer.
rooms. They showed me how to make sepia Line types and line thickness are tough problems
“backgrounds”, photographic negatives, and in a world of infinite zoom, and the solution in
composites out of multiple sheets. A photocopy AutoCAD does not seem ideal. Pen settings in
machine (a relatively new piece of office AutoCAD are inevitably annoying, but are
equipment that many architects believed was an understandable in the context of pen plotters. In
unneeded luxury) could be used to copy repetitive the 1980’s, the line weight depended not on a
drawings, such as standard details, entourage, and software setting but upon which pen was in which
even plumbing symbols, onto transparent adhesive location in the pen carriage of the plotter. The
film (sticky back paper). These could be adhered distinction between paper space and model space
to a drawing, eliminating hours of work. Dry helps to correct the metaphor by applying more
transfer lettering increased the legibility of text, tangible aspects of the metaphor, such as line
substituting for a skill that distinguished an weight, color, and visibility, in paper space. Model
experienced draftsman from a newcomer. Patterns space can then be focused upon representing the
pre-printed onto adhesive film enabled large areas building itself, where line weight, line style, or
of hatching to be created quickly and precisely. color have no real meaning beyond their symbolic
Even colors and tones could be easily added as importance.
transfers.
Autodesk has partially replaced the overlay
I bought a pin bar for registering multiple sheets metaphor in Architectural Desktop 2005. In the
of mylar to support overlay drafting, which I used new metaphor, an element is a conceptual part of
at home on competitions that I entered as a student. a building that may be repeated. A construct is a
I read books on advanced drawing production, conceptual part of a building that is not repeated
such as those by Fred Stitt (1984). In retrospect, I and is made largely of elements. Various views
realize that I acquired an advanced knowledge of of constructs can be arranged into sheets, sheet
drawing production that was far beyond what I sets, and a project. The metaphor is based on the
could have learned from professors and schools. latest thinking regarding how to put together a
Overlay drafting was the cutting edge of large set of drawings to describe a large building.
architectural production of the day. By implementing the metaphor with external
references (Xrefs), Architectural Desktop can
When software developers invented CAD systems, manage hundreds of sheets and thousands of
they mimicked the state of the art of drafting in drawings while maintaining adequate performance

on commodity hardware. Building Information Ruled surfaces, surfaces of revolution, surface

Modelers promise to substitute a new metaphor meshes, and Euclidean primitives are all provided.
that is focused upon the physical components of a Auto LISP is available for automating the creation
building. of meshes. It is relatively simple to generate a
table of vertices in Excel as a table and execute it
Although the overlay drafting metaphor is in AutoCAD. The Advanced Modeling Extension
ubiquitous, it is neither inevitable nor unique. For (AME) was introduced in the early 1990s as a
example, a cognitive metaphor has been suggested pioneering Constructive Solid Geometry system
as more supportive of the design process (Clayton that was far more advanced than the surface
et al 1994). Rather than grouping objects by modelers available in competitor’s products. With
layers, the graphic objects are identified within Release 13, solid modeling was fully integrated
interpretations as features that have significance into AutoCAD.
in various tools for evaluating the performance.
More recent versions have provided ever more
2 or 3D powerful tools for editing and viewing 3D models,
as illustrated in figure 2. A right click on the
Senegala suggests that AutoCAD imprisons AutoCAD screen pops up a menu that includes
designers in a two-dimensional flatland. I hear the 3D Orbit command, from which one can
and read this criticism frequently, but it is rapidly and intuitively set up perspective views.
incongruent with the actual software. Few people For even more advanced 3D modeling, Autodesk
seem to realize it, but AutoCAD incorporates provides tight integration with 3D Studio Viz.
powerful surface modelers and solid modelers AutoCAD is clearly an extensive environment for
based on geometry libraries that are among the 3D modeling. It is particularly valuable for
best available in the software industry. Its object- teaching 3D modeling by virtue of its tools for so
oriented solid modeler provides for Boolean many different 3D paradigms. A student can
operations on solids as well as extrusions along compare 2 ½ D to mesh modelers and constructive
arbitrary paths, shell commands, and face-editing. solid geometry.
The built in renderer performs very well, providing
both ray tracing and shadow mapping, and Disorder, Information and Process
extensive texture mapping, background
compositing, and entourage. The techniques are One final criticism of AutoCAD deserves
well-explained in a widely used textbook for attention: the user interface is disorderly and lacks
architectural computing (Kolarevic 1998). a coherent model for a design process. This
criticism betrays a profound lack of understanding
A historical perspective can explain the regarding the stated market for Autodesk
widespread ignorance about the 3D capabilities products. In the 1980s when AutoCAD s
in AutoCAD. Old impressions gained before the hegemony was surprising, the conventional
addition of more powerful functions seem to wisdom was that the company had hit on a
persist in spite of additions to the feature set. One’s fortuitous formula for selling CAD software to
image of AutoCAD seems colored by the version architects. AutoCAD was easy for dealers to
that was current at the time of first impressions. customize. Thus they could sell block libraries,
The early focus of the software on 2D drafting is Auto Lisp routines, menu customizations and
easy to understand in the context of the demand training to their clients. Architects wanted and
in industry for overlay drafting and the limitations needed the hand-holding provided by a registered
of the IBM PC hardware. AutoCAD dealer and the dealer could make
profits far beyond the margin on the retail price
AutoCAD evolved at a pace with the rest of the by selling services. The smart dealers sold
industry with respect to 3D. For example, in the AutoCAD and the smart architects found a smart
late 1980s, VersaCAD on the Macintosh, provided dealer.
only surfaces extruded parallel to the Z-axis.
AutoCAD still provides this version of so-called AutoCAD has always been a CAD engine rather
2 1/2 D modeler through manipulation of the than a finished shrink-wrapped, turn-key product.
thickness property. Surface modeling is also The purchaser is supposed to customize it for the
available in AutoCAD using the 3Dface command. needs of a particular firm. It is now and in the

past also was irrelevant if the user interface was GIS system. No competitor vendor, with the
unappealing to a particular user. If you like the possible exception of Bentley, even attempts to
menu structure, keep it. It you don’t like it, change address the design process of architecture and
it. The software can adapt to any process. other disciplines that are responsible for the built
environment.
AutoLISP is not the end of AutoCAD
customization tool development. In the early Miscellaneous Rebuttals
1990s, the AutoCAD Runtime Extension (ARX)
libraries were devised to allow a developer to write Several other criticisms are directed at AutoCAD
programs in C, compile them with any typical C in Senegala’s article. He suggests that a sketching
compiler, and link them into a running copy of tool should have a sense of ambiguity and should
AutoCAD. This was and is an amazing trick that support multiple layers of undo and redo.
enabled developers to conceive and implement AutoCAD is not intended to be a sketching tool,
extensions to AutoCAD virtually without limits but is quite serviceable for exploratory 2D and
in terms of scope and speed of execution. An 3D modeling. A user can be sloppy with drawing
example of an ARX tool created in a research lab and then “grip edit” back to a grid or other snaps.
in the mid 1990s is the Semantic Modeling One can edit the vertices of a polyline one by one
Extension, which tied together expert systems, to correct a sloppy line. If ambiguity is important
automated consultants, and AutoCAD on multiple, to the user, then the latest version can scribble on
heterogeneous machines across the Internet one’s drawings automatically to make them look
(Clayton et al 1999) soft and ambiguous. Photoshop can be used to
smudge one’s renderings to convey lack of
Once Autodesk focused exclusively on the commitment. With regards to undo and redo
Microsoft platform, the company moved quickly commands, AutoCAD has long supported
to incorporate the Microsoft Component Object powerful tools for settings points in a session,
Model (COM), exposing its entire class hierarchy backtracking through the session, and then moving
to developers for manipulation from within any forward again.
COM compliant application. A conversant
programmer can run AutoCAD commands from Conclusions
within Excel or call Excel routines from within
AutoCAD. By virtue of its 23 year history and dedication to
preserve old features, AutoCAD is a rich document
COM programming is not for the novice, but other of the history of CAD. For one who can read
features in AutoCAD provide many levels of AutoCAD, the software offers reminders of the
integration with data and software. Block attributes best ideas from the past and explanations for the
and extraction enable quick and accurate quantity current state of architectural computing. From a
take-offs and furniture tracking. Data linking historical perspective, the user interface of
provides dynamic, real time communication with AutoCAD reveals a history of computational
database systems. The hyperlink property is design and computer graphics. From the
available to link any entity with a URL. Blocks perspective of architectural practice, the argument
can be provided with behavior and drawings is compelling that AutoCAD is an excellent tool
located on the Web dropped into a drawing. that satisfies at a high level more needs than most
AutoCAD continues to provide capabilities for competing products. A concept map of this
process modeling and data integration that are easy discussion is provided in figure 4.
to use and accessible, although rarely used by the
average user. Deconstruction of Senegala’s article reveals a
hidden assumption and bias in favor of a view that
Of course, with the Desktop series, Autodesk has drawing is indistinguishable from design, a view
provided CAD software tailored to specific that one might term “the cult of sketching.” By
markets. Architectural Desktop 2005 provides a focusing on the dialectic between the AutoCAD
carefully designed process for building design. user interface and hand sketching, he privileges a
Civil Desktop and Land Desktop tailor AutoCAD myopic focus on hand sketching. While this
for civil engineering and landscape architecture. “stance” is very common among those in the
Autodesk Map 3D turns AutoCAD into a powerful academic community, from an objective

standpoint, sketching is not what architects do. drawing. AutoCAD puts few constraints upon
Architects collaborate, produce construction the study and practice of architecture and is thus
documents, manage business, design details, an impressive computer-aided design tool. In all
research materials, and act in many other ways. its elderly crankiness, it is a lovable software
In a typical firm, a few architects out of many package for those who have “been there, done
sketch and visualize grandiose schemes. Viewed that.”
from within the cult of sketching, the AutoCAD
user interface may appear ludicrous. However,
from a more complete view of architectural
practice the AutoCAD user interface is highly
optimized for production, collaboration, and
information management.
Figure 4. Redrawn concept map critiquing the user interface

of AutoCAD
The tradition of computational design research and

the critical importance of drafting production are
two truths of architectural computing that belie
the “cult of the sketch,” Because it is aligned with
the mainstream of architecture discipline and
practice, AutoCAD is not subversive. However,
it does throw down a gauntlet to those who would
privilege the graphic image as the only expression
of architecture.
My experiences have guided me to recognize that

architecture is a diverse and challenging profession
that engages one’s intellect in mathematics,
engineering, writing, management, and personal
relationships as well as drawing. CAD has for
me always meant computer-aided design in all of
the myriad aspects of design. It has never meant
computer-aided drafting or computer-aided

References
Clayton, M. J., Kunz, J. C., Fischer, M. A., and

Teicholz, P. (1994). First drawings, then semantics.
In Reconnecting: ACADIA 94. Association for
Computer Aided Design in Architecture.
Clayton, M. J., Teicholz, P., Fischer, M., and Kunz,

J.. (1999). Virtual components consisting of form,
function and behavior. Automation in
Construction (8) 351-367.
Hoskins, E. M. (1977) The OXSYS system. In J.

S. Gero (Ed.) Computer Applications in
Architecture, (pp. 343 – 391). Essex, England:
Applied Science Publishers, Ltd.
Keller, S. (2005). 10283EFE0F02 or the Seagram

Building. R. Hejduk and H. van Oudenallen,
(Eds.), The Art of Architecture/The Science of
Architecture. Washington, DC: Association of
Collegiate Schools of Architecture Press.
Kolarevic, B. (1998). Architectural modeling and

rendering with AutoCAD R13 and R14. New York:
Wiley.
Mitchell, W. J. (1990). The Logic of Architecture:

Design, Computation and Cognition. Cambridge,
MA: The MIT Press.
Mitchell, W. J., Liggett, R. S., and Kvan, T. (1987).

The Art of Computer Graphics Programming: A
Structured Introduction for Architects and
Designers. New York, NY: Van Nostrand
Reinhold.
Moore, G. T., Milne, M., and Geier-Wilson, R.

(1993). Architectural research. Progressive
Architecture (74[8]), 83-85.
Senegala, M. (2004). Deconstructing the software

interface: a critical close reading of AutoCAD,
International Journal of Architectural Computing
(3), 299-314.
Spencer, D. D. (1997). The timetable of

computers. Ormond Beach, FL: Camelot
Publishing Company.
Stitt, F. (1984). Systems Graphics. New York:

McGraw-Hill Book Company

P. Anders, W. Lonsing / AmbiVIewer: A Tool for Creating Architectural Mixed Reality
AmbiViewer: A Tool for Creating Architectural

Mixed Reality
Peter Anders1, Werner Lonsing2

1
MindSpace.Net
2
Independent Researcher
Abstract
This paper presents a new mixed reality system for architecture, AmbiViewer. The system employs digital
video, onboard modeler and global positioning to merge physical and simulated entities on the screen. The
system can be used to model projects on-site, and in view of the project environs. The paper also discusses
the use of AmbiViewer in creating cybrids, compositions of virtual and material reality. The paper concludes
with a description of a small project undertaken with AmbiViewer and its implications for cybrid architecture.

Introduction external environment, the display, and the user’s

senses is crucial to the effect. For instance, a user
Merging digital and physical environments into a of a HMD (head-mounted display) may be
blended reality is a focus of inquiry among simultaneously aware of 1) his physical
technologists, artists and designers. This surroundings, 2) virtual objects that map onto
technology, known as augmented, or mixed, reality physical objects, 3) virtual objects that float
(AR and MR respectively) has appeared in venues independent of the physical space, 4) and virtual
ranging from computer labs to gallery installations objects that are fixed parts of the visual display –
since the early 1990’s. In this paper we will such as a menu or icon. Sophisticated tracking of
describe the creation of a tool for architects to the viewer involves sensors, whether they be
generate three-dimensional compositions – full- motion detection systems, microwave detectors,
scale objects – on actual sites, and to view the Global Positioning Systems (GPS), and other
results as navigable mixed realities. This tool, we means or combinations thereof. Importantly, the
argue, presents an important opportunity for user’s bodily and cognitive use of space keeps the
conventional and computer aided design experience from being chaotic. The experience of
processes. Its implications reach beyond imaging a coherent, comprehensive space is maintained.
and fabrication to suggest alternative architectural (Anders 2003)
products, ones that fully engage the mixed realities
of contemporary culture. Mixed reality refers to a spectrum of synthetic
physical/virtual experience. Paul Milgram and
This paper will describe the potential use of Herman Colquhoun Jr. of the University of
augmented and mixed reality in architecture, and Toronto have developed a helpful taxonomy to
specifies an apparatus to support architectural distinguish mixed reality’s varied effects.
design. We follow this with a description of an (Milgram and Colquhoun 1999) It is situated
MR system called AmbiViewer along with a within a larger scale of experience, one that
discussion of its hardware and software extends in their argument between “real” and
subsystems. At the paper’s end we describe a small “virtual” environments. Milgram and Colquhoun’s
project undertaken to evaluate the system’s description of mixed reality offers a variety of
performance in the design/creation of mixed hybrid effects ranging from Augmented Reality
realities. Concluding remarks discuss projected on the “real” side of their real/virtual scale, to
developments for AmbiViewer as a tool for Augmented Virtuality (AV) on the “virtual” side.
designers and creators of mixed realities. Before Augmented Reality, we noted, overlays virtual
proceeding, however, we should first define the elements onto physical environments. Conversely,
terms we will be using. Augmented Virtuality, overlays “real” elements
onto virtual environments. The effects of AV
Terminology would resemble special effects in contemporary
film, where images of real actors are collaged into
Our topic, the merger of material and simulated animations or computer-rendered sets. The degree
elements, is known in the literature as augmented to which the viewer interacts with the result –
or, alternatively, mixed reality. Augmented reality whether AR or AV – determines the result’s
often employs transparent displays that overlay effectiveness as a mixed reality. Whereas virtual
computer graphics onto the viewer’s visual field. reality closes the user off from physical
It has been used in the development of heads-up surroundings, most mixed reality opens the virtual
displays for equipment repair, in surgery that world to the immediate environment. The resulting
allows physicians to see into their patients’ bodies, montage hybridizes the user’s experience, and at
and as a means for discerning damaged pipes in its extreme blurs distinctions between simulation
murky water. While visual technology and actuality.
predominates in AR, some forms of augmented
reality overlap the sense of sound and touch as Architecture and Augmented Reality
well. The visual focus of AR is a challenge to its
engineers. However, if the overlaid graphics are In the eyes of AR’s developers the construction
linked to external objects, a moving viewer’s head industry offers a promising market for the
and gaze must be tracked to maintain a consistent technology. Their reasoning is straightforward:
illusion. The spatial relationship between the AR deals with spatial and symbolic phenomena

in ways potentially useful to architects, builders Seeing the Invisible

and facility managers. These industries’ increasing
familiarity with – and reliance upon – computers Among other virtues of mixed reality is the ability
makes them well suited for mixed reality of its users to perceive invisible aspects of their
technology. However, as with virtual reality before surroundings. Grant Foster and his colleagues at
it, the success of such forecasts hinges on a variety the University of Reading have described the uses
of issues. Some of these lie outside the domain of of a simulation overlaid onto a building. They
technologists: compatibility, reliability, flexibility, have developed DAMOCLES, a system that
and purchase costs, training and upkeep. Not least enables vision or hearing impaired users to
important is the degree to which AR systems serve navigate a building by aural and visual cues.
the needs, and values of architects and their clients. (Foster, Wenn, Harwin, 1998) Beyond helping the
blind, however, Foster believes that AR systems
The overlay of spatial computer models onto can let users see the invisible. Equipment that
buildings has uses in nearly all stages of the generates heat, for instance, can be visually
building’s life. A project’s database may facilitate augmented to keep operators from harm. Such
a building’s design, its construction and thinking lies behind similar proposals for
maintenance. Computer scientist Gudrun Klinker visualizing the invisible. (Kieferle and Wössner,
suggests that a project’s mixed reality would attend 2001) Professor Anthony Webster at Columbia
all stages of development, from the earliest siting University has also explored the use of AR in
of a building to its subsequent occupation. In a architecture – particularly in the field of
paper written on uses of augmented reality models construction. With colleagues Stephen Feiner and
on construction sites, Klinker and her colleagues Blair McIntyre, he was able to render to view the
speculate on AR’s use in the building’s life cycle. hidden reinforcing rods in a concrete column using
a head-mounted display. (Feiner, Webster,
“With AR, ... virtual geometric objects can be Krueger, MacIntyre, Keller, 1995) The ability to
integrated into the real environment during all see through obstacles, such as concrete, murky
phases of the life cycle of the building. Before water, or human flesh is a constant theme in AR’s
the construction project is started, AR can support literature.
marketing and design activities to help the
customer visualize the new object in the Just as Klinker and others project a site’s future
environment...During construction, AR can help with augmented reality, its past may be similarly
evaluate whether the building is constructed revealed. Feiner and his colleagues at Columbia
according to its design...After construction is have modeled the campus of their university as
completed, maintenance and repair tasks benefit an armature for historical research. Using a heavy,
from seeing hidden structures in or behind walls.” but mobile, AR system users navigate the campus
(Klinker, Stricker, Reiners 1998) and “see” buildings torn down years ago, or they
can explore the school’s underlying tunnels that
This bears directly on the life-cycle of an proved vital in the 1968 student uprising. (Feiner,
architectural project, stressing the digital model’s 2002)
role in constructing and serving buildings. The use
of virtual models as annotational guides for
Mixed reality has also found uses in
construction is prevalent in AR literature. Indeed,
telecommunications. A number of researchers
among the earliest uses for AR was in overlaying
including Jaron Lanier have developed a desktop
instruction manuals onto the viewing field of
augmented reality, called teleimmersion, that lets
factory workers at Boeing. (Curtis, Mizell,
users converse with remote colleagues by using
Gruenbaum, Janin, 1999) (Mizell, 1997)
an elaborate screen display. The user is able not
Subsequent work by others illustrate the use of
only to see a partner, but, by moving his head, can
AR for the repair of copy machines, door
view the remote space as though through a
assemblies and other equipment. (Reiner, Stricker,
connecting window. The effect is apparently quite
Klinker, Müller, 1999) (MacIntyre and Feiner,
convincing since the virtual images are taken from
1996) (Molineros, Raghavan, Sharma, 1999)
a number of corroborating video cameras. (Lanier,
2001)

opportunity, then, for a mixed reality sketching

system that lets users compose designs spatially,
spontaneously, and on site. Solutions so created
could then apply to future stages of the project.
Developing such a system requires a performance
specification based on the perceived needs of
designers at the initial stages of a project. These
needs include portability, ubiquity, visualization,
and indeterminacy and are summarized below.
Portability
Conventional sketching is done with materials that

can easily be carried from place to place: paper,
Figure 1. Image of a cybrid showing construction and pencil, sketchpads, etc. Many current AR systems
cyberspace (on left) used for telepresence. The remote space depend on a fixed infrastructure (lab space,
is replicated as a virtual annex to the physical space tracking systems, desktop computers) for their
performance. An AR sketching system, while not
as concise as its sketchpad counterpart, must at
This, plus the extremely high speed of the least be portable by one person and allow
advanced Internet connection makes the presence flexibility in setup and use. Any dependency on
of the distant space almost palpable. This fixed infrastructure on site – tables, outlets, and
combination of social space and telematics cable connections – should be minimal.
suggests architectural mixed realities that have
been termed cybrids. (Anders, 1999) (Figure 1) Ubiquity
Architectural Needs / Opportunities for AR As mentioned, many current AR systems require

a stable sensor environment to track the user’s
It is apparent from the foregoing discussion that camera or head mounted display. This often entails
AR has much to offer architects. AR may indeed an indoor environment that – in turn – limits the
affect all stages of a building’s construction and apparent size of an effective virtual object. A full-
use – presumably, a mixed reality could inform size simulation of a sited building does not
the architect’s design from the project’s onset. But conduce to a lab setting, for instance. The AR
this is not so easy as it might seem. Although AR system must be independent of fixed transmitters
technology allows three-dimensional models to be that would limit its use to specific locations.
sited in actual space these models require prior Ideally such a system could be deployed anywhere.
design articulation before they can be merged into
the scene. As a result, AR cannot be employed Visualization
until fairly late in the design process. This is
unfortunate because the earliest stages of design Visualization refers to the inscription of designed
are highly information sensitive and would appear images for assessment and application to later
to benefit most from digital assistance. phases of the design. Pencil and paper provides
an ideal model for a visualization system, however,
There are several reasons that AR use is limited at the resulting images are necessarily shown
this stage of design. Research on Mixed Reality statically from a particular viewpoint – and solely
is not devoted to design exploration so much as as inscriptions on a page. An AR system should
developing effective illusions: end products such provide such visualization capacity, yet let its user
as animations, installations, or other special explore the design – effectively sculpting it from
effects. The virtual components of these mixed multiple vantages – rather than fixing it with
realities are pre-designed. In architecture, unrelated sketches. Three-dimensional CAD
however, the virtual, imagined objects of design models are a useful precursors to such digital
are not known at the beginning of a project. To visualization.
the contrary; spontaneity, inspiration and
indeterminacy characterize this phase. We see an

In our present application the hardware devices

are integrated by use of several layers of software.
These include software for 1) determining
positions of cameras and fiduciary features, 2)
image capturing, 3) image processing to identify
and evaluate features, 4) interactive modeling of
three-dimensional objects, 5) rendering
perspective views of the model depending on
generated values, and 6) composing the captured
and rendered model into composite images.
Portability and Ubiquity
The system in its current form is fairly portable,

Figure 2. Digital video camera with GPS receiver attached using batteries rather than power outlets with no
at top determines viewer location, direction of view and focal
length. Smaller laptop-mounted cameras have since been used
need for physical connection to a positioning
to minimize weight and inconvenience system. It is possible to obtain wireless connection
to the Global Positioning System (GPS) using
Bluetooth. Excluding the weight of a laptop
computer, the system weighs only ounces: a small
Indeterminacy GPS tracker, digital camera, and a visual marker
(a balloon or ball). We can transport it easily on
We have already discussed the need for the foot in a knapsack, or in a car. (Figure. 2)
spontaneous use of AR in sketching designs. This
entails creating a system in which virtual elements Because of the need for outdoor use and the
of the mixed reality are created on the fly and not architectural scale of the subject matter we
simply ported in from other applications. While required a positioning system that was external
such an exchange would be useful in later design and flexible. AR commonly employs controlled
stages, a sketching system must have on-board environments with customized technologies fitted
modeling capacity. A system that meets these to the setup. For our uses, however, we required
needs has been under development since the mid- that our outdoor tracking system be both
1990’s. Over the past year we have collaborated affordable and robust. The Global Positioning
in developing and testing the results of this effort, System effectively makes the surface of the Earth
AmbiViewer, a tool for designing and creating a controlled environment with situated satellites.
mixed reality. GPS’s use in AR has precedents in the work of S.
Feinman, B. McIntyre, B. Jang, R. Azuma among
The AmbiViewer System others. It enhances the mobility of its users and
their equipment, freeing them to visit the sites of
The configuration of AmbiViewer resembles those their choice. The present system is well suited to
found in other augmented reality set-ups. Like the scales of architecture, civil engineering, and
them it requires computer processors, user tracking regional planning, and we anticipate its use in
systems, display and rendering software. For siting buildings, bridges and infrastructural
economy our display system is presently limited elements.
to computer monitors, although a worn display is
preferred for the final configuration. Another Changing locations of viewers and features is
feature that distinguishes Ambiviewer is an efficient and affordable with the system. Under
interactive modeler that allows real-time creation normal circumstances GPS results in signal delay
and manipulation of virtual objects. Necessary intervals of about one second and accuracy within
hardware includes 1) tracking devices, like GPS three meters. In more developed areas where
or similar, to concurrently determine positions of Differential GPS (DGPS) is available – the United
objects and users, 2) digital cameras as real-time States and Europe for instance – submeter
image capturing devices, 3) fiduciary features that accuracy is standard. DGPS systems are bulky and
act as reference objects and 4) at least one mitigate against our need for portability. Also,
computer with display. since DGPS systems are presently expensive we

have, for now, limited our focus to the Visualization and Indeterminacy
conventional accuracy of GPS for camera and
fiduciary location. (Figure 3) While conventional sketches do not distinguish
between real and virtual elements, analogous AR
There are other ways to get this accuracy without compositions present a number of special
the expense and difficulty. R. Azuma observes challenges. Composite images in AR require an
that greater degrees of accuracy outdoors require input source, digital camera, modeling source, and
a hybrid approach to technology. (Azuma et al, a three-dimensional rendering engine. In addition
1999) For this reason our system employs a visual the source and model images must be convincingly
fiduciary tracking subsystem to compensate for combined into one image. Because live-streamed
GPS uncertainty. This results in a surprisingly video is the only adequate image input for a real-
effective, if rough, measuring tool. At present GPS time application, the captured stream has to be
precludes indoor tracking because receivers assessed frame-by-frame in order to detect,
require line-of-sight access to the satellites. That identify, and evaluate the features using methods
said, we are still in the product’s development and of image processing, including Hough
further hybridization remains an option for indoor Transformations.
use.
Figure 3. Interfaces for two concurrent GPS locations, one for the fiduciary feature, the other for the
viewer camera. AmbiViewer allows simultaneous tracking and location. Each GPS unit can display the
satellites used in setting the location and direction, below

Such image processing is demanding and creates and generated images. Since calculating the
a significant bottleneck in many AR applications. perspective views from the site is the major task
It puts the entire computational load on the of the software, the quality of the result depends
processor and requires a selective use of data. largely on the accuracy of calculated spatial and
Only focal length and exact viewing positions are temporal values. This accuracy is helped by using
generated using the fiduciary feature while other, GPS to assess user position and the fiduciary
possibly valuable, input must remain unused. feature to determine the focal length and exact
direction of view. (Figures. 4, 5, 6, 7) To enhance
Happily, generating a real-time 3D model is no the realism of the virtual model we can employ
longer a problem for AmbiViewer. By creating GPS to determine sun position, or add atmospheric
an object-oriented modeler completely based on values via the fiduciary system. These GPS values
OpenGL, the former bottleneck of 3D rendering – refined by those obtained from the fiduciary
is overcome. The final steps of image processing feature – are applied in both the modeler and the
are executed on the graphic card, or GPU, using rendering. However this added task, while
its processor and memory. Composing the captured improving realism, burdens a processor already
and generated images into a composite scene is encumbered by dataflow management and
then a comparatively simple matter of properly geometric control of the model. Presently this
combining memory buffers and displaying the option decreases the overall performance of the
results onscreen. The process is uninterrupted and system, although it remains an option for us as
automatic. The quality of the final composite is computational power increases.
directly related to the quality of the video stream
Figures 4, 5, 6, and 7. These images show a simulation set into an actual landscape. The study was done to assess environmental
impact of the proposed windmill. It was possible to view the mill from all angles and distances using the system.

The current configuration of AmbiViewer opens

a number of variables to the user including site
and viewing locations, number, size, shape and
color of virtual components, and their situation
within the spatial setting. Most importantly, the
user has live control of the virtual elements beyond
viewing angles and camera movement. Creation
of new elements, modifying or moving existing
ones, or deleting unwanted objects is done in real-
time within the application and always in view of
the site. This freedom provides the user with the
creative indeterminacy required at the earliest
stages of design. In addition users can, in the
Figure 8. An initial sketch using Ambiviewer. The bottom course of modeling, archive the design process as
ball on the right is the fiduciary marker for the mixed a video for later reference, a feature that could be
reality. The rest of the balls, cylinders and hovering bar are useful for educators as well as practitioners. Users
simulations
could also easily access and re-do prior states of
the virtual model through use of this “time-stamp”
record.
Figure 9. The design of the playhouse shown under

development in the Ambiviewer composition screen. The panel
on the right allows creation and control of new objects in the
scene. Note fiduciary ball on lower left
Figures 10 and 11.Shown above is the playhouse as built.

On the left is the playhouse with situated 3D elements using
AmbiViewer. Such sketches will inform the cyberspace of the
cybrid playhouse

A Design Experiment References
In the summer of 2004 we undertook a small

project – a playhouse – to test the system’s use in Anders, Peter (2004). A Procedural Model for
the design and creation of an architectural mixed Integrating Physical and Cyberspaces in
reality, or cybrid. At this time the system had Architecture. Doctoral dissertation, University of
support from GPS for its camera system. Although Plymouth, Plymouth, U.K.
it was hard to calibrate precisely, the modeler did
allow the rough massing of elements on the site. Anders, Peter (1999). Envisioning Cyberspace,
After preliminary efforts at placing arbitrary New York: McGraw-Hill.
objects in the yard (Figure. 8), a suitable site and
preliminary scheme emerged. (Figure. 9) Since the Azuma, R., B. Hoff, H. Neely III, R. Sarfaty, M.
AmbiViewer version employed had no ability to Daily, G. Bishop, L. Vicci, G. Welsch, U. Neumann,
export the model to other applications, the S. You, R. Nichols, J. Cannon (1999). Making
structure was built in ensuing weeks from sketches augmented reality work outdoors requires hybrid
based on the AR model. (Figure. 10) Successful tracking. R. Behringer, G. Klinker, D. Mizell
enough as a playhouse, we intend to use the (Eds.), Augmented reality: Placing artificial
physical shelter as a “site” for virtual additions. objects in real scenes (pp. 219-224), Natick, Mass.:
(Figure. 11) We anticipate that AmbiViewer would A. K. Peters.
make such cybrid compositions possible and “real”
to observers, especially if other systems are
employed to realize sound and tactility. Curtis, D., D. Mizell, P. Gruenbaum, A. Janin
(1999). Several devils in the details: Making an
Conclusions AR application work in the airplane factory. R.
Behringer, G. Klinker, D. Mizell (Eds.),
We have here presented an architectural authoring Augmented Reality: Placing artificial objects in
tool that employs mixed reality concepts to merge real scenes (pp. 47-60), Natick, Mass.: A. K.
simulations into actual settings. Using GPS and Peters.
visual fiduciary features we have attained Feiner, Stephen (2002). Augmented reality: A new
sufficient accuracy to model and locate designs way of seeing. Scientific American, April, 48-55.
for construction, autonomous simulations or
hybrids of physical and virtual elements. An on- Feiner, S., B. MacIntyre, T. Höllerer (1999).
board modeler not only lets us generate these Wearing it out: First steps toward mobile
products, but allow interactive manipulation of augmented reality systems. Y. Ohta, H. Tamura
virtual elements even after a project is completed. (Eds.), Mixed reality: Merging real and virtual
As the system is further integrated and its worlds (pp. 363-375). New York: Springer-Verlag.
portability improved, we look forward to
developing means for direct fabrication of physical Feiner, S., A. Webster, T. Krueger, B. MacIntyre,
and virtual elements of cybrids and their user and E. Keller (1995). Architectural anatomy.
interfaces. Presence 4(3), 318-325.
Foster, G. T., D. E. N. Wenn, W. S. Harwin (1998).

Generating virtual environments to allow
increased access to the built environment.
International Journal of Virtual Reality 3(4), 12-
19.
Jang, B., J. Kim, H. Kim, D. Kim (1999). An

outdoor augmented reality system for GIS
applications, Y. Ohta, H. Tamura (Eds.), Mixed
reality: Merging real and virtual worlds (pp. 391-
398). Springer-Verlag: New York.

Kieferle, Joachim and Uwe Wössner (2001). Thomas, B., B. Close, J. Donoghue, J. Squires, P.
Showing the invisible: Seven rules for a new De Bondi, M. Morris, W. Piekarski (2000).
approach of using immersive virtual reality in ARQuake: An outdoor/indoor augmented reality
architecture. Hannu Penttila (Ed.), Architectural first person application. Anne Rawlinson (Ed.),
Information Management. Proceedings of the 19th Proceedings of the Fourth International
eCAADe Conference, Helsinki, Finland, 29-31 Symposium on Wearable Computers, (pp. 139-
August, (pp. 376-381). 146). IEEE Computer Society: Los Alamitos,
California.
Klinker, Gudrun, Didier Stricker, and Dirk Reiners
(1998). The use of reality models in augmented
reality applications. Reinhard Koch and Luc Van
Gool (Eds.), (pp. 275-289) Springer: Berlin.
Lanier, J. (2001). Virtually there. Scientific

American, April, 66-75.
Lonsing, Werner, Digitale Bildverarbeitung. Part

1 in: Bauinformatik No. 5, 1992, 188 - 194.
Part 2 in: Bauinformatik No. 6, 1992, 246 – 255,
Werner – Verlag: .
MacIntyre, B., S. Feiner (1996).Future multimedia

user interfaces. Multimedia Systems 4(5), 250-268.
Milgram, Paul, and Herman Colquhoun Jr. (1999).
A taxonomy of real and virtual world display
integration. Y. Ohta, H. Tamura (Eds.), Mixed
reality: Merging real and virtual worlds (pp. 5-
27). Springer-Verlag: New York.
Mizell, David W. (1997). Virtual reality and

augmented reality in aircraft design and
manufacturing. Proceedings of the Second Annual
Symposium on Frontiers of Engineering (pp. 29-
31). National Academy Press: Washington, D.C.
Molineros, J., V. Raghavan, R. Sharma (1999).

AREAS: Augmented reality for evaluating
assembly sequences. R. Behringer, G. Klinker, D.
Mizell (Eds.), Augmented Reality: Placing
artificial objects in real scenes (pp. 91-98). A. K.
Peters: Natick, Mass.
Reiners, D., D. Stricker, G. Klinker, S. Müller

(1999). Augmented reality for construction tasks:
Doorlock assembly. R. Behringer, G. Klinker, D.
Mizell (Eds.), Augmented Reality: Placing
artificial objects in real scenes (pp. 31-44). A. K.
Peters: Natick, Mass.
Shreiner D., M. Woo, J. Neider, T. Davis (2003).

OpenGL Programming Guide: The Official Guide
to Learning OpenGL, Version 1.4, Fourth Edition
by OpenGL Architecture Review Board. Addison-
Wesley: Reading, Mass.

A. Schmidt, P. Kirkegaard / Navigating Towards Digital Tectonic Tools
Navigating Towards Digital Tectonic Tools

Anne Marie Due Schmidt1, Poul Henning Kirkegaard2
1
Utzon Center, Aalborg University, Denmark
1
Department of Civil Engineering, Aalborg University, Denmark
Abstract
The computer holds a great potential to break down the barriers between architecture and the technical
aspects relating to architecture, thus supporting innovative architecture with an inner correspondence
between form and technique. While the differing values in architecture and technique can seem like
opposites, the term tectonics deals with creating a meaningful relationship between the two.
The aim of this paper is to investigate what a digital tectonic tool could be and what relationship
with technology it should represent. An understanding of this relationship can help us not only to
understand the conflicts in architecture and the building industry but also bring us further into a
discussion of how architecture can use digital tools.
The investigation is carried out firstly by approaching the subject theoretically through the term
tectonics and by setting up a model of the values a tectonic tool should encompass. Secondly the
ability and validity of the model are shown by applying it to a case study of Jørn Utzon’s work on
Minor Hall in Sydney Opera House - for the sake of exemplification the technical field focused on
in this paper is room acoustics. Thirdly the relationship between the model of tectonics and the case
will be compared and lastly a discussion about the characteristics of a tectonic tool and its implications
on digital tectonic tools will be carried out.

Introduction the computer’s tectonic ability. Especially Leach

is concerned with an inner correspondence
Sketches, physical models and words are integral between form and technique which is only possible
tools in the architect’s repertoire; that can support to achieve by a focus on the very first sketches.
the architectural process. While these tools have He sees the computer’s ability to test designs in
been present for centuries, the palette of tools is terms of technical ability as its primary tectonic
today - since the introduction of drawing programs capacity and describes the computer as "…an
in the early 1980ties - increasingly dominated by efficient search-engine that is premised on the
computers. The application of computer programs notion of efficiency". Leach envisions that the
in architecture comes in various shapes – from the designer would supply the edge-conditions for the
two-dimensional drafting that can be said to mimic design and then with little interference let the
the traditional techniques of the pencil drawing to computer decide the configuration with regards
the three-dimensional modelling and parametric to structure, acoustics, environmental concerns,
designs. These tools are to some extent already constructional or programmatic issues. While
integrated into the architect’s work while the full Leach is on the right track with a focus on
potential of the computer’s ability to make interaction with the computer tools, his excitement
technical parameters accessible is still to be about the new media lures him into accepting tools
explored. The potential of breaking down the with a focus only on technical aspects and
barrier between the technical fields and efficiency as tectonic, thus minimizing the
architecture is that it becomes possible to develop architectural impact to a choice between a few
innovative tectonic solutions by reinstating the options.
architect as a master builder of the digital age.
The aim of this paper is to investigate what a
A natural backdrop for a discussion of the tectonic tool could be and what relationship with
breakdown of the barrier around technique in technology it should represent. For the sake of
architecture is the term tectonics that deals with exemplification the technical field focused on in
the connection between the scientific and artistic this paper is room acoustics but the findings are
aspects to create architecture. To understand what supposedly applicable to all technical fields. This
we should expect from a tectonic tool a closer look is done firstly by a theoretical approach to the
into the term tectonic is taken. The term is problem in order to set up a model of the values a
described by Kenneth Frampton (1995), Eduard tectonic tool should encompass, secondly to
Sekler (1965) as well as many other theorists. compare this model to a case study of Jørn Utzon’s
Sekler writes: work on Minor Hall in Sydney Opera House which
is seen as a valuable resource in understanding
‘When a structural concept has found its the role of technology and its influence on tectonic
implementation through construction, the visual architecture, thirdly the model and the case will
result will affect us through certain expressive be compared and implications to for a digital
qualities which clearly have something to do with tectonic tool is discussed.
the play of forces and corresponding arrangement
of parts in the building, yet cannot be described in Theoretical model – a navigational tool
terms of construction and structure alone. For these
qualities, which are expressive of a relation of form One way of setting a frame to discuss tectonic
to force, the term tectonic should be reserved.’ computer programs, is to begin with tectonics and
to set up a model of the value system behind
Sekler thus emphasizes two distinct set of values architecture. This is set up in a theoretical model
in tectonics; the expressiveness and the ‘play of – see figure 1. This model can – which will be
forces’, meaning the technical dimension. It seems presented in the case study – be used to identify
evident, then, that in order to obtain tectonic the various values behind architecture much like
architecture, a double focus is needed in order to we would use a navigation tool to tell us where
see the work of architecture as a synthesis of we are positioned between different areas.
technique and aesthetics. Secondly it can be used to point out a direction
towards tectonic digital tools – as will be done in
Recently many theorists, such as Leach (2004), the discussion - much like we use a navigational
Stacey (2004) and Abel (2004), have emphasized tool to set out a direction for our next move even

though this should not be seen as one route but The positivistic approach is evident through the
rather a general direction. language used: ‘defined goal’, ‘maximum
efficiency’, ‘exact and concrete result’ and
‘maximum efficiency’; only in the last sentence;
where Nervi describes efficiency as ‘not entirely
limited by economic factors’, other values are
hinted, but not elaborated upon.
The other point of view, phenomenology, validates

the bodily experience of architecture and, in the
case of acoustics, enables a focus on the subjective
feel for the sound. This focus does not
acknowledge definite statements such as an
‘optimal’ acoustical room; rather it understands
the human experience as subjective and closely
connected to dreams and imagination. Gaston
Bachelard describes in ‘The Poetics of Space’ how
we perceive our homes:
Figure 1. Strategy for tectonic tools
"…all really inhabited space bears the essence of
The structure of future and current tectonic tools the notion of home. In the course of this work, we
should make it possible to maintain the double shall see the imagination build ‘walls’ of
focus which, as described in the introduction, is impalpable shadows, comfort itself with the
the main quality of tectonic architecture. In order illusion of protection – or, just the contrary, tremble
to understand construction and structure as well behind thick walls, mistrust the staunchest
as other technical parameters in building, a ramparts. In short, in the most interminable of
positivistic approach is needed, but to maintain dialectics, the sheltered being gives perceptible
that it is not only a technical solution that is sought limits to his shelter. He experiences the house in
but a tool capable of encompassing the bodily its reality and in its virtuality, by means of thought
experience in the architectural field, the and dreams." [Bachelard, 1964, p. 5]
phenomenological approach is needed too. The
model is thus an attempt to navigate in the reality The emphasis is here on ‘essence’, ‘imagination’,
as it is seen: tectonic architecture as consisting of ‘being’ and ‘thought and dreams’.
both technical solutions and bodily experiences.
These different values can be identified in the From these two quotes it is seen how different the
following quotes by Pier Luigi Nervi and Gaston actors behind architecture and engineering can
Bachelard. A positivistic statement is expressed think. The underlying assumption behind the
by Nervi, an engineer known for creating tectonic model is that it is possible to unite these very
architecture: different value-systems and that it is necessary to
create tectonic architecture. This was also argued
"The design process is fundamental for the by Karl Bötticher, a nineteenth century German
creation of buildings and determines their form writer on tectonics (Schwarzer, 1993).
from the first preliminary studies when the
architectural idea is born to the final construction Another theoretical support is the field of social
phase where every structural element is studied constructivism in which many writers work with
in detail. It can be defined broadly as the invention dissolving the barriers between technology and
and study of the necessary methods to achieve a society. While the ideal for scientific knowledge
defined goal with maximum efficiency. (…)It in the positivistic tradition is that it is objective
would be senseless, in fact, to make a study which and free of value from the researcher and the
did not aim for an exact and concrete result, and surrounding society, the work of social
which did not try to reach it with maximum constructivists deals with understanding how this
efficiency, that is, a broadly defined efficiency not actually is not the case and that how and what
entirely limited by economic factors." [Nervi 1965, scientific knowledge is produced, is indeed closely
p. 105]. connected to society. A social constructivist, Bruno

Latour (1988), argues that the access to and ability Minor Hall for the Opera House in Sydney. This
to apply technology, is to have power and that the is chosen because it is seen as a valuable resource
ones in charge (the Princes in Latour’s in understanding the role of technology and its
terminology) actually aim at achieving this power; influence on tectonic architecture due to Utzon’s
"The two most common clichés about technology, insistence on a tectonic solution. To zoom in on
its inertia that would be too strong for anyone to the interplay between the agents and their values
resist, and its inner complexity that would be too in the architectural process, Henriksen’s theory of
much for any one to fathom, are real enough, not reality – describing four aspect of reality: facts,
as the cause of the Prince’s moves, but as the logic, values and communication (Henriksen et al,
effects that the Prince strives to achieve." (Latour, 2004) – is applied.
1988, page 38). Let us for a while imagine that
this is true – that technology is not characterized First we need some facts about the case in order
by inertia and complexity but rather should be seen to be able to analyze the development and
as active and simple, then it becomes much easier recognize different values. In his competition entry
to imagine a softening of the edges between from 1956 it is obvious from the drawings that
technology and architecture and to imagine Utzon had not thought much about the interiors
tectonic tools that can close in on the middle of of the halls or of the realization of the structure as
the model in figure 1 and create tectonic a whole. The competition brief asked for a Major
architecture. Hall which should primarily cater for concert
recitals for an audience of 3000 to 3500. A
This middle-part of the model where the values secondary use for the hall was to be opera with no
overlap is the position of the values that is specified seating requirement (Nobis 1994, p. 13).
necessary if we should be able to develop The Minor hall was to be used for theatre
computer tools to support tectonic architecture. performances and seat around 1100.
The technology should in tectonic tools not exist
on its own terms but as a support to develop After having won the competition in 1957, Utzon
tectonic ideas. This means that these programs are immediately began work by gathering a team of
significantly different in interface and advisers and collecting their reports in a document
functionality than the programs that exist within known as the Red Book (due the colour of the
the field of architectural and technical expertise cover). In the field of acoustics the Danish
today. acoustician Vilhelm Lassen Jordan was hired. The
halls showed in the Red Book (March 1958) were
Social constructivism concludes that technology primarily done by Jordan which Jordan himself
is not value free but should be understood from hints at in the quote: ‘Without analysing the
the relationship with society. In the context of cooperation between architect and acoustician too
computer programs it means that the computer minutely, it is fair to say that regarding this first
tool, as a technological device, is subject to design, the acoustician was responsible for many
interests. The social constructivist frame of suggestions.’ (Jordan, 1980, p. 95). This is further
thinking thus enables us to see the current state of supported by the fact that Utzon himself does not
computer programs merely as a symptom of the mention this proposal at all in his Acoustical
values behind them. Report from 1965, that recaptures the design
development of the halls.
Narrative - Minor Hall in Sydney Opera House
The drawings for this first proposal shows a hall
Having argued for the theoretical possibility of a which seems to be inspired from traditional
break-down of the barriers between technology classical halls called shoe box hall for their
and architecture, the existence of such a break- rectangular floor and ceiling and straight walls.
down in reality is sought for in the study of a real- Jordan describes the halls in the Red Book with
life case. In the case study the model will be used the remark "The proportions of the area are 2:6:7
to identify the different values held by the actors, (mean height: mean width: depth), which is
thereby achieving a closer understanding of the appropriate for a typical theatre hall, where the
role of technology in the architectural process. stress is laid more upon definition than
reverberation." (Nobis, 1994, p.33). The proposal
The case study chosen is the development of the did not have any familiarity with the organic shape

of the sails and it turned out that the proposal did a long time. It was very dynamic and
not fit geometrically under the curve of the sails. breathtakingly beautiful. Every one of the waves
showed a different character in its movement. Jørn
For the next years Utzon and Ove Arup, the said,"Yuzo, can’t we design the ceiling of the
construction engineer, were fully occupied with Minor Hall something like that?" whilst looking
the construction of the podium and the sails and it at the breaking crest of the waves." (Mikami, 2001,
was not until after the Red Book (from March p. 118)
1958), that Utzon began to be more involved in
the design of the halls. In a meeting in mid-1958 The ceiling of the Minor Hall was conceived as
it seems that Jordan has been asked to explain segments radiating from the centre of the stage,
some of the fundamentals of acoustics to the each segment having a curved concave shape. The
architects. He explains how the volume is radial configuration would focus the attention of
determined by the shape and functionality of the the listener towards the stage, to where the sound
room and comments on the shape of the surfaces was coming from, and would enable entrance to
"In Minor Hall the curved surface can be accepted the hall to be hid between the segments. Utzon
if the surface is highly absorbent" (Author’s describes the effect "therefore, for the eye of the
translation) (Minutes of meeting 30.6.1958). seated spectator the theatre appears absolutely
closed, even when the doors are actually open"
(Utzon, 1965). The first drawing of this scheme
occurs in August 1959 (Nobis, 1994, p. 31) and
the following three years the scheme is developed
extensively. One of the most important aspects of
the design of the halls, when addressed by Utzon,
was that the experience of the halls should support
the whole opera house experience, every aspect
in the journey from the city towards the harbour -
the rising up the stairs, the entrance into the
southern foyer and following the shape of the halls
under the sails to the entrance of the halls - were
all part of building up to the climax; the
performance in the halls. To him the relationship
between the sails and the halls was therefore of
uttermost importance. In a later interview with
Figure 2. Model of the ‘stepped-cloud’ scheme [Source: Peter Luck he described how he wanted the
Weston, 2001] auditoria to fit under the sails like walnuts. "A
walnut…when you see a walnut from the outside,
you get a feeling there’s something inside, with a
Utzon and his employees developed a scheme he slightly wobbled form. And then you open up,
called the ‘stepped-cloud’ – see figure 2. This is you’re not surprised but it’s quite different inside,
described as the first scheme by Utzon in his but it’s still in harmony with the outside. And I
Acoustical Report from 1965. One of his former left the shells open so you could see up under the
employees Yuzo Mikami, who drew the proposal, shells, feel the structure." (Nobis, 1994, p. 4)
recalls the beginning of the development From an acoustical point of view, however, the
‘stepped-cloud’ scheme was not advantageous
"Jørn asked me to work on the new design of the because of the concave curvature of the ceiling
Minor Hall auditorium. He took me to a nearby that would result in echoes and focusing of sound.
seashore by the Sound in Hellebaek, where the However, Utzon actually did consider sound when
gentle movement of water formed the ridges of the proposal was drawn. He saw the hall as a
waves one after another. The continuous ridge musical instrument in wood and tried to represent
went up higher and higher as it came nearer to the the sound as it radiated from the stage.
shore, and finally the crest of the wave began to
break, overriding the ridge and coming down with Utzon wrote in the Report on Acoustics "The first
a drumming sound onto the wet sand on the beach. acoustical engineer to participate in the project was
We watched the movement of the waves for quite Dr. Jordan. He gave up at a certain stage when we

came up to a dead corner with Minor Hall…" page 9). Both of the new consultants were send
(Utzon, 1965, p. 9), thus suggesting that Jordan drawings of the ‘stepped-cloud’ scheme in August
was not able to give satisfactory advice regarding 1962. Cremer responds:
the new shape of the hall.
"The large radius concave curvature of the ceiling
In 1960 Utzon asks Jordan for "a list of literature is rejected. Should such curvatures eventuate, their
dealing with acoustical problems" (Minutes of radius must be small. Particularly, the curvature
meetings, 17.11.1960). Finally in 1961 Jordan is of the rear ceiling area will lead to sound focusing
quoted for that he "agreed on the shape of MI-hall onto the last rows of the stalls. Also of great
subject to minor modifications" (Minutes of disadvantage is the dome-like raising of the mid
meetings 16.03.1961). part of the ceiling with its large step toward the
stage, in which the lighting is housed…This
In retrospect Jordan, however, describes the produces very strong delayed reflections in the
development as such middle of the stalls" (Nobis, 1994, p. 38).
"In his approach to the design, the winner of the Gabler drew a number of sections with arrows
competition (the Danish architect Jørn Utzon) was representing the reflections of the sound to explain
obviously influenced, primarily by the simplicity the problems of the current scheme. Furthermore
of the Greek amphitheatre but also by some he included a number of suggestions as to how to
profoundly personal ideas, indeed sculptural substitute the large concave curvature with a
conceptions, which were original to the point of number of smaller convex curves – see figures 3a/
being revolutionary in architecture, but which, b.
unfortunately, had absolutely no association with
the classical design of concert halls." In his report from 1965 Utzon remarks on his
(Jordan 1980, p. 92) acoustical development:
The ‘stepped-cloud’ scheme was presented in the "A number of laws clearly define when
Yellow Book, given to the Australian Government disadvantages such as echoes occur, and so on,
in February 1962 and was, despite its acoustical therefore it is possible to draw on paper the shape
difficulties, accepted. which will, in an original way, give direct sound
to all the seats. This is absolutely creative work.
On Utzon’s return from Sydney, where he had The shape which gives a certain amphitheatre
presented the Yellow Book, he went to Berlin and feeling, as in the Minor Hall, provides the most
Vienna. In Berlin he met the German acousticians brilliant and clearest sound from the orchestra pit
Professor Lothar Cremer - one of the most or stage – it is creative work because such a thing
important acousticians of this century, Professor as "This is right" or "This is wrong" does not exist.
Emeritus at the Technical University of Berlin - The solution can only be found by experiment and
and Professor Werner Gabler – an architect my experiment became actually realistic only after
specialized in acoustic spaces. The meeting is I had learned how sound behaves. It takes a long
recorded in the minutes "Jørn Utzon explained the time, and good co-operation with acoustical
drawings for the Sydney Opera House. It was engineers, to understand the properties of sound.
decided that Mr. Gabler and Professor Krämer On the other hand, it also took a long time for the
should, with the help of a model, test the acoustics acoustical engineers to understand the freedom
in the Sydney Opera House together with Dr. with which my concepts allow us to work.
Jordan or as a second opinion." (Minutes of Normally, acoustical engineers are repairing a
meetings 02.06.1962) fixed project." (Utzon, 1965, p. 2)
In the Report on Acoustic Research Utzon Furthermore he describes how he in Berlin was
continues the quote above "… and I contacted introduced to two rooms – one with only
Professor Cremer and Professor Gabler in absorptive surfaces and one with only reflective
Germany, whose halls I had studied and whose surfaces and how this experience was important
Berlin Opera House I admired very much. The in his understanding of how sound is perceived
Berlin Opera House had dimensions very close to and how it behaves (Utzon, 1965).
my dimensions for Minor Hall." (Utzon, 1965,

The scheme was being reworked completely

within a few months to encompass the convex
curves – see figure 4b. A great deal of the
intentions from the ‘stepped-cloud’ scheme, such
as the radiation and the entrance to the hall, was
retained. Nobis writes
"When one compares the new scheme with the

one shown in the Red Book it becomes clear that
Utzon has regained control of the acoustic
requirements on his own terms. The Red Book
scheme was an engineer’s solution to a set of
requirements. The SOH796 scheme was an
architectural solution which incorporated the
acoustic requirements." (Nobis, 1994, p. 43)
The convex proposal was developed further with

regards to the construction of the ceiling profiles.
With the ceiling profile almost set, the profiles
were further developed to use the full size of the
giant plywood sheets manufactured in Australia
at that time. The halls were to be constructed as
box plywood beams hung from the concrete beams
of the sails.
Fromonet remarks that by portraying the ideal

acoustic profile with the aid of a single geometry,
Utzon was able to integrate his conception of
sound with that of the constructive dimension. "In
the margin of one of his sketches he mischievously
dubbed this synthesis of structure and acoustics:
‘Strucacustithese’ " (Fromonot, 1998, p. 166)
In his acoustical report from 1965 Cremer

describes the results of the testing of the Minor
Hall. Two problems are found – the rear wall and
the walls around the stage. Suggestions are made
to improve these two problems and Cremer writes
in a letter to his assistant Joachim Nutsch, who
was working in Sydney with Utzon, that he
considers these tests finished. The well-known
Figures 3a and b. Letter from Gabler to Utzon - diagrams
explaining focusing effect of concave ceiling and to the right ending to this story is that despite the acousticians’
suggestion to change the ceiling from concave to convex and architect’s satisfaction with the project the
[Source: Nobis, 1994] architect left the project in 1965 and therefore the
interiors are different from what had been planned
Immediately after Gabler’s letter, the drawings by Utzon and the rest of the team.
being produced in Utzon’s office to develop the
Minor Hall began to show a concern for sound Understanding the case
paths – see figure 4. Nobis remarks that this was
the first time such considerations were shown in In order to understand the case, the model in figure
the Minor Hall drawings and that it was 1 helps us understand the values and realities of
investigated how to use the sound path differences the actors. The first concept, which was drawn by
as the determinant of the ceiling profiles (Nobis, Jordan, should be able to tell us something about
1994, p. 40). Jordan’s values and reality. Firstly, as Nobis has

problem was solved with the proposal in the Red

Book: the required number of people was seated
in an appropriate room with the desired room
acoustic quality.
However, viewed from a tectonic point of view,

the Red Book proposal is ambiguous. Drawing a
curved wall where a curve works against the
purpose of the room and therefore has to be
neutralized by rendering it highly absorbent, is a
contradiction. If the curve creates a condition that
needs to be solved, why draw it in the first place?
By creating that solution, new problems arise that
has to be solved by adding another layer to the
solution. What is interesting here is Jordan’s
relationship to technology; it is clear that he is not
aiming at an inner correspondence between form
and technique, but rather at a goal described by
positivism as Nervi described. This value can thus
in the model be positioned to the outer left, in the
part of positivism that does not overlap with
phenomenological values.
The second concept was clearly drawn from a

different stand point. The phenomenological
values guiding this scheme are evident in both
Utzon’s and Mikami’s descriptions of the scheme
where they emphasize the expressiveness and the
Figures 4a and b. Sketches by Utzon which shows reflection
lines [Source: Nobis, 1994]. Model of the convex proposal
sensation the human being would experience in
[Source: Nobis, 1994] the hall. Mikami stresses the expressiveness of the
waves as "very dynamic and breathtakingly
remarked upon, the proposal tells us about his beautiful." (Mikami, 2001). And Utzon, in one of
inspiration from classical concert halls. Secondly his many analogies of how one would experience
the conflict between the geometries of the sail and the Opera House "A walnut … you’re not
the hall itself reveals that geometrical studies were surprised, it’s quite different inside, but it’s still in
not the part of Jordan’s everyday life. Thirdly it harmony with the outside. And I left the shells
tells us something about his understanding of open so you could see up under the shells, feel the
acoustics. structure." (Nobis, 1994). The phenomenological
values are evident from their language that comes
As mentioned it seems that Utzon left it completely close to Bachelard’s language and emphasizes the
up to Jordan to supply a proposal for the Red human experience. Because the values guiding this
Book. With this carte blanche Jordan drew a proposal were not linked to any technical concerns,
concave back wall – probably to follow the curve it can be found in the outer right of the model; in
of the seating rows – that could produce focusing the area of phenomenology that does not overlap
and thereby strong echoes in the front of the room. with positivism.
When Jordan was asked about the curve he
remarks that "In Minor Hall the curved surface One of the very odd events in the case of the
can be accepted if the surface is highly absorbent" development of the Minor Hall to Sydney Opera
(Minutes of meetings, 30.6.1958). This reveals to House is that Jordan did not comment on the shape
us Jordan’s sense of logic and thereby his of the ceiling in the ‘stepped-cloud’ scheme; there
underlying set of values. From a positivistic point are no documents showing any protests from him.
of view the statement is absolutely true, the desired One possibility is that Jordan did not know the
result is obtained, there are no echoes. scheme; this is highly unlikely since he during the
Furthermore, from a positivistic point of view, the three years of developing the scheme was present

at several of the regular meetings held at Utzon’s begin by determining the problems of the room,
office in Hellebæk. Neither is there any reason to which could for instance be an echo. Secondly the
believe that Jordan was not competent in his field. acoustician would then carry out a number of tests
Another explanation could be that Utzon was not eliminating elements of the room as cause of the
interested in the acoustics but wanted to focus on problem. If the room shows an echo and the room
the architectural expression. This probably could with the ceiling covered does not, there is naturally
be the case for the very first sketches of the idea a good probability that it is the ceiling causing the
but this scheme was developed for three years. To echo and that the ceiling needs some adjustments.
support a rejection of this view is also the fact
that the ‘stepped-cloud’ scheme actually shows a This method, however, has its limitations in a
concern with sound in a conceptual way – the hall tectonic process because it, as mentioned, requires
is perceived as a funnel-shape that should transport that the architecture is static. Jordan repeated the
the sound to the audience and Utzon displays a procedure with a number of proposals for the
great interest in acoustics by asking for further Major Hall but the expenses and time consumption
readings on acoustics. Utzon’s interest in acoustics of the testing equipment – a 1:10 scale model of
can, with regards to the model in figure 1 be the hall – prevented extensive use. The technology
interpreted as an attempt to move from the outer available thus only gave room to a one-way
side of phenomenology into the overlapping area dialogue, where Jordan would test a proposal that
of the model in the middle. was believed to be finished aesthetically.
The most likely explanation is that what we see What Utzon wanted to do with acoustics was
here is a representation of Jordan’s understanding different. He was searching for the essence in the
of technology. Jordan saw technology as acoustics and he was interested in the human
instrumental – a way to realize an architectural experience and wanted to work with the audible
idea. He shows this clear distinction between as an intertwined part of the visible. His approach
architecture and acoustics in the Red Book was both phenomenological and moving towards
proposal where he uses the acoustics to mend the the tectonic part of the model in figure 1. From
shape of the room. This is supported by Utzon’s this point of view, technology is only a means to
comments in his Report on Acoustics "Normally, reach the greater goal, the human experience, but
acoustical engineers are repairing a fixed project" is never the less important in the search for the
(Utzon, 1965). Jordan did not comment on the best shape of the hall – the one that will support
proposal because he did not want to interfere with the visual, the audible and any other aspect of the
the architectural development of the proposal. This experience of the hall. From this value followed
way of applying technology is in good agreement the logic that architecture could not be carried out
with classic positivistic values, where it is very in different tempi, all the aspects had to be
important that the objects studied are static - or investigated simultaneous in order to be solved in
fixed as Utzon calls it – because only that way the one solution in one instance – to achieve a tectonic
scientist can be certain that the only variable is solution. In order to do this the barrier between
one he introduces. If, for instance, the scientist architecture and technology had to be broken
was interested in knowing the effect of a certain down.
medicine he could ensure the tests’ validity by
monitoring the health of two batches of mice – When Gabler and Cremer entered the project they
one that was given the medicine and one that was brought with them a different understanding of
not. The test would, however, not be reliable if reality. Their basic values were still from within
the conditions of the mice were not alike - if one the sphere of positivism but their logic and
forgot to feed one batch of mice that might be a understanding of technology were different than
more probable cause of death than the medicine. Jordan’s. This is evident in the communication
In order for the test to be reliable, the between them and Utzon. They did not only
circumstances around the test thus need to be kept comment on the ‘stepped-cloud’ scheme but drew
static. explanations of the principles behind their
comments. Also they tried to explain some
In terms of architectural projects it is possible to fundamental concepts of sound and acoustics to
apply the same logic if one is repairing a fixed Utzon by letting him experience the anechoic and
project. A reliable way of doing it would be to the reflective room. Thereby they brought the

acoustics from the field of positivism into Utzon’s through him. In order to be able to embed the
field of phenomenology – the figures and data principles of acoustics into his architecture, Utzon
became connected to the human experience of the mediated between the two fields by reading about
room. On the model in figure 1 the collaboration acoustics and ultimately consulting other
thereby moved from the outer edges of acousticians. One of the most important aspects
phenomenology and positivism to the middle part in this mediation is the communication. Some of
of the model, where there is an overlap of the the main points of the case study – and something
values and therefore a possibility to create tectonic that can be brought into the sphere of the
architecture. development of future tectonic tools – are the small
tricks that Gabler, Cremer and Nutsch employed
The reason for this different understanding of in order to ‘translate’ the positivistic values of their
technology might be that Gabler himself was field into phenomenological terms. It is seen that
trained as an architect and therefore used to the actors in the real-life situation diversify their
translate acoustical requirements into architectural expected predetermined way of acting of their own
matter. He must also have known that the concept field. Gabler and Cremer began drawing in order
Utzon presented was not necessarily finished, but to communicate their ideas and Utzon gained
only to be considered as a step in an iterative access to the logic behind the positivistic aspects
process. Also Cremer’s assistant Joachim Nutsch, of acoustics and he on his turn began working with
who was working with Utzon in Sydney, was reflection paths on his drawings. Likewise Gabler
known to be a talented sketcher able to and Cremer successfully broke down the barrier
demonstrate in drawing what he suggested to be between the positivistic and phenomenological
altered. When Gabler, Cremer and Nutsch were fields by introducing Utzon to the anechoic and
introduced to the project it is thus evident that it reflective chamber. This experience can be
was not primarily technical talent that was added characterized as communicative in the same sense
to the project, rather it was another way of as a dialogue is communicative but is at the same
communicating technical considerations. time a completely different kind of dialogue than
Especially Gabler and Nutsch seem to have had a what would be expected within positivism – it goes
clear understanding of the intimate relationship beyond facts and deals with the human experience.
between the architecture and the acoustics that
Utzon wanted to achieve. One thing worth noting is that Jordan, Gabler,
Cremer and Nutsch shared the positivistic values
Discussion – their aim was the best achievable acoustics.
Despite of this, they worked in different ways in
Values, communication and logic the collaboration with Utzon. Their understanding
of technology and thereby their logics are one way
The differing values held by Utzon and Jordan of explaining this discrepancy. Henriksen et al
made it difficult to create tectonic architecture. (2004) describe logics as a way of transforming
Utzon wanted advice on what rules to play by facts into contours of the future change, a way of
when working with acoustics and Jordan wanted envisaging change. While the acousticians agreed
a scheme that he could optimize. Jordan’s way of on the aim of their involvement in the project it
reasoning is very logical when seen within the thus seems that their logics led them to different
context of positivism. It was also evident from the paths to achieve this goal. Timing is one of the
case that the creation of tectonic architecture is most obvious differences between Jordan and the
based on mediation between the two values which German team. While Jordan was applying
was what Gabler, Cremer and Nutsch enabled. As technology on a ‘finished’ concept, Utzon, Gabler
such, the future tectonic tool could be seen as a and Cremer wanted to use the technology as a
mediating device able to transcend the barrier creative force in the design.
between differing values.
Digital tectonic tools
While Jordan mastered his field and was
recognized as one of the best Danish acousticians Translating these findings into consequences for
at his time he was not able to transfer his digital tectonic tools, it seems that the tectonic
knowledge into a creative force. It was difficult tools should communicate technology in a manner
for Utzon to obtain access to the technology that firstly translates the positivistic values into a

phenomenological sphere by letting the human Despite the obvious difficulties with the analytical
experience be the focus point. In other words, the programs, it is surprisingly enough these tools that
relevance of the numbers and figures of the today are closest to being tectonic, not the
positivistic field should be understood by its phenomenological ones. For instance the
impact on the human experience of sound. development within the acoustic field is
Secondly the program should be able to maintain interesting. This field has taken a full loop from
a dialogue with the user. From this we should also being a subjective field with only rules of thumb
realize that if the aim is tectonic architecture, applied to – around 1900 - becoming an ‘objective’
existing technical computer tools can not simply science applied to architecture. Today the
be applied to architecture because they are based introduction of auralization – the process of
on other values. Furthermore the logics of the rendering audible (by physical or mathematical
participants in the process indicate that the tool modelling) the sound field of a source in a space
should be based on a timing that allows the – makes it possible for instance to hear how a
architecture to use the technology as a creative concert will sound in an un-built room. The ability
force. is still mainly being used as a gimmick to impress
clients and is not yet systematically used to base
The reason why the findings from a 45 year old design decisions on, but recent studies have
case are relevant today, is that to some extent we showed the reliability of the programs (e.g.
are still working with these divisions in Pancharatnam, 2003). The significance of this, in
architecture even though the computer has a tectonic terms, is firstly that through the ability to
potential in breaking down the barrier between hear architecture the digital media is coming one
architecture and technology. In our current digital step closer to the materiality of build architecture.
tools we can recognize the positivistic and By adding another dimension – sense – to the
phenomenological values in our distinction expressiveness of the architecture, the focus on
between the architectural computer tools and the the image is reduced, which has been one of the
engineer ’s analytical computer tools. The discrepancies between the digital and the tectonic.
architectural tools are primarily drawing programs By hearing the materials as well as seeing them
that vary from a digital version of the drawing there is, for instance, suddenly a representational
board in two-dimensional drawings to difference between materials looking like stone
sophisticated generative three-dimensional and actual stone because they will not sound alike.
programs. The phenomenological concern of these While the acoustical simulation programs as
programs is evident in that they aim to shed light mentioned are far from being straight forward for
on the visual aspects of the space created. The architects at the moment, this is only a matter of
analytical tools are concerned with analysis, refining the interface and the structure of the
evaluation and optimization of given parameters program. The possibilities are evident. The hearing
in construction, acoustics, interior climate etc. and added to the vision will give a closer understanding
can as such be said to work within the positivistic of the connection between the space and the
side of the model. materiality.
One of the analytical tools within acoustics is the Secondly such a development is significant
acoustical simulation program CATT. A test of the because a digital tectonic tool allows a continuous
potential of this to assist the creation of tectonics dialogue between the positivistic and
(Schmidt and Kirkegaard, 2005) made it clear that phenomenological aspects of architecture.
the tool is highly deterministic and this positivistic Changes in the technical part of the building – for
value guiding the program renders it un-suitable instance by changing materials or the logics behind
to support an architectural design process. Some the constructional system – will become evident
of the problems caused by this positivistic value in the phenomenological part of the program. In
was very alike the approach Jordan showed in the the case of the acoustical program it will be
case study. For instance the program focuses on a possible to hear and see the difference the changes
late stage of the architectural process thereby make as well as detect the changes in a number of
making it difficult to test preliminary concept ‘objective’ parameters such as reverberation time,
proposals. Furthermore it is difficult for an sound pressure etc. In the model in figure 1 this
architect to ‘communicate’ with the program due development can be seen as a move towards the
to its highly specialized target-group. middle of the model. Thereby the technical choice

will move from being only technical to being able combined with a logic to use the technical
to encompass the bodily experience in the parameters as a ‘creative force’, the values move
architectural field. One of the most important towards the middle of the model in figure one and
aspects of this is to recognize that there is never it becomes possible to create tectonic architecture.
only one solution to the technical dimension in The case showed that communication, logic and
architecture – we are always confronted with a values are all significant components in the break-
choice where the answer is subjective. down of the barrier between technology and
architecture.
Conclusion
These components were taken on to a discussion
This paper deals with the relationship between of digital tectonic tools. It was argued that the
architecture and technology. The term tectonics is current digital tools can be understood from the
taken as a starting point for this discussion because model as supporting phenomenology and
the term deals with creating a meaningful positivism as separate aspects while it – as in the
relationship between the two. Likewise a great case study – is necessary to mediate between the
potential to break down the barrier around values in order to work tectonically. Some
technology was argued to exist in the digital media. potential to do this is identified in an acoustical
Therefore the paper links tectonics and the digital simulation tool that enables both a positivistic
and investigates what a digital tectonic tool could approach by calculation and a phenomenological
be and what relationship with technology it should approach by the ability to hear the sound in a un-
represent. An understanding of this relationship, build room.
it is argued, can help us not only to understand the
conflicts in architecture and the building industry
but also bring us further into a discussion of how
architecture can use digital tools.
The investigation was carried out firstly by

approaching the subject theoretically through the
term tectonics and by a setting up a model of the
values a tectonic tool should encompass. Tectonics
as a meaningful relationship between art and
science is seen as mediating between the human
experience expressed in phenomenology and a
technical view upon the world as described in
positivism. As such, these values can be seen as
the meeting of the architectural and engineering
values in an architectural process. The model thus
describes the meeting between these values and
is meant as a navigational tool which can be used
to identify the various values behind architecture
much like we would use a navigational tool to tell
us where we are positioned between different areas
that are more or less defined. Secondly it can be
used to point out a direction towards digital
tectonic tools much like we use a navigational tool
to set out a direction for our next.
Secondly the ability and validity of the model are

shown by applying it to a case study of Jørn
Utzon’s work on Minor Hall in Sydney Opera
House. It was shown that when positivism and
phenomenology existed separately, it was difficult
to obtain a tectonic result. When, however, both
the positivistic and phenomenological values are

References Pancharatnam, S. and Ramachandraiah, A.: 2003,

A Study on the Computer Aided Acoustic Analysis
Abel, Chris (2004) Architecture, Technology and of an Auditorium. The Institution of Engineers
Process, UK: Architectural Press/Elsevier (India), Journal of Architectural Engineering, 84.
Bachelard, Gaston (1969) The Poetics of Space. Sekler, Eduard (1965) Structure, Construction,
Translated by Maria Jolas. New York: Orion Press, Tectonics. In: Structure in Art and in Science,
1964; Boston: Beacon Press. Gyorgy Kepes, ed., Studio Vista, London, 89-95.
Frampton, Kenneth (1995) Studies in Tectonic Schmidt, A. M. D. and Kirkegaard, P. H.: 2005,
Culture - The Poetics of Construction in From Architectural Acoustics to Acoustical
Nineteenth and Twentieth Century Architecture, Architecture Using Computer Simulation, Journal
Cambridge Massachusetts: MIT Press of Building Acoustics, 12 (2), 85-98.
Fromonot, Francoise (1998) Jørn Utzon. The Stacey, Michael (ed.) (2004) Digital Fabricators,
Sydney Opera House, Milan: Electa/Gingko University of Waterloo School of Architecture
Press, Toronto
Henriksen, Lars Bo et al (2004) Dimensions of
Change – Conceptualising Reality in
Organisational Research, Copenhagen: Schwarzer, M.: 1993, Ontology and
Copenhagen Business School Press Representation in Karl Botticher's Theory of
Tectonics, Journal of the Society of Architectural
Jordan, Vilhelm Lassen (1980), Acoustical Design Historians, 52, 267-80.
of Concert Halls and Theatres – a personal
account, London: Applied Science Publishers Utzon, Jørn (1965), Report on Acoustics,
Unpublished, Mitchell Library Collection, ML
Leach, Neil; Turnbull, David; Williams, Chris MSS 2362/21, Sydney.
(eds) (2004) Digital tectonics, 1st edition, Wiley-
Academy, UK. Weston, Richard (2001) Utzon, 1 st edition,
Germany: Edition Bløndal
Latour, Bruno (1988) ‘The Prince’ for machines
as well as for machinations, Technology and social
process, Edinburgh, Edinburgh University Press,
20 – 43
Mikami, Yuzo (2001) Utzon’s Sphere. Sydney

Opera House – how it was designed and built,
Tokyo, Japan: Shokokusha
Minutes of meetings 1958-1966, Unpublished,

Mitchell Library Collection, ML MSS 2362/4,
Sydney.
Nervi, Pier Luigi (1965) On the design process.

In: Structure in Art and in Science, Gyorgy Kepes,
ed., Studio Vista, London, 105-110.
Nobis, Philip (1994) Utzon’s interiors for The

Sydney Opera House - The design development
of the Major and Minor Hall 1958-1966,
Dissertation, University of Technology Sydney.


C. Sanchez-del-Valle / Adaptive Kinetic Architecture
Adaptive Kinetic Architecture: A Portal To Digital

Prototyping
Carmina Sanchez-del-Valle1
1
Hampton University
Abstract
This paper presents a definition for adaptive kinetic structures in architecture, generated from an examination
of research in engineering and architecture. This characterization introduces the challenges presented both by
modeling form and environment, and simulating their interaction. Adaptive kinetic structures react to a changing
environment, as well as generate their own. These conditions make them appropriate subjects through which
the design and implementation of tools for ‘digital prototyping’ may be explored.
Digital prototyping serves performance and simulation-based design. In general terms, it is an interdisciplinary
integrated approach for modeling, predicting, and analyzing the behavior of a system. It is at the core of
virtual engineering. In the aerospace, automobile, and manufacturing industries, it is practiced extensively
through discrete-event and continuous simulations, as well as simulation environments. This paper provides
an overview of digital prototyping commercial software for engineering applications that can be transferred
to architecture, and identifies some of the unresolved issues. It thereby extends the vision of the comprehensive
building information modeling initiative.

Introduction respond to changing conditions, such as the

weather, the time of the day, and the location of
The assertion that architecture needs adaptive the sun. Thus, the justification for adaptive kinetic
kinetic structures responds to the profession’s structures is founded on three reasons: economy
continuous search for adaptable design that of means, responsibility towards the natural
optimally serves changing human needs. This environment, and the satisfaction of human needs
search is coupled here with an examination of and desires. These are no different from those
approaches that allow modeling and simulating given for most architectural projects; however,
complex behaviors within complex environments. adaptive kinetic structures can better modulate
When an adaptable kinetic architecture responds efficiencies, broaden the contemporary aesthetic,
by adjusting itself to a changing environment, it and give it more relevant meaning. The single
forms an ecological system. It is ecological attribute that distinguishes the adaptive kinetic
because the system components have shifting from other forms is that embodied energy becomes
interdependencies. Currently in architecture there fully visible when it produces work, when it moves
is no single modeling and simulation environment in a perceptible way. Applied energy is work, and
that allows designing these structures in an work as motion is intrinsic to the system as is the
integrated manner. There is also limited data about transformation of energy.
the actual behavior of the systems and its
consequences. The lessons afforded by the
excellent examples built by many over time await
full compilation and ongoing analysis for future
use. This means that at the heart of any model or
simulation is empirical research.
A host of tools to model form and material in detail,

assigning attributes to components and spaces, and
to execute mainly discrete simulations to evaluate
performance have been developed for architectural
design. Through digital fabrication (rapid
prototyping and manufacturing) digital models are
translated into physical prototypes and
construction components. Although novel in its
actual implementation, it follows an ancient
tradition: to evaluate a structure to predict its
performance in the future: it is necessary to build,
test, and document its behavior. Fabrication is but
one phase of virtual engineering. Virtual
engineering is “a simulation-based [design] Figure 1. Conceptual diagram of the adaptive kinetic system
method requiring a virtual environment, where the
geometric and physical properties of systems are
accurately simulated,” covering the life of the
design from concept development to Uniquely an adaptive kinetic structure moves in a
manufacturing processes, to prototype. [Lee, predictable and visible manner responding to
p.433] This paper explores the application of expected excitations. It has a built-in capability to
digital prototyping or digital mock-up, which change state through motion, while maintaining
requires an integrated interdisciplinary approach its structural integrity and stability. This movement
to modeling, simulating, and analyzing the can be produced automatically through
behavior of a system. programmed responses to changes, or manually.
Motion is fundamental to the life of the system.
The term ‘kinetic’ in this context stands for having We use ‘system’ instead of ‘building’ to better
the capacity to be affected by reversible reflect all that entails architecture. An adaptive
geometrical changes in whole or in part without kinetic system is characterized by energy/control,
losing the integrity of the system. When a kinetic purpose/function, condition/state, geometry,
structure is also adaptive, it gains the ability to structure, and material. The system is immersed

in an ever-changing environment. At the same time Buildings that house machines:

the system, in architecture, also creates its Thames Water Tower, England, Brookes Stacey
environments. Randall
Buildings with machine-like components:
The Kinetic Factor High Sierra Cabins, U.S.A.; Jones Partners
Architecture.
Far from whimsical, kinetic structures have proven Building components as machines:
to be essential in space, in extreme or hazardous Sapporo Dome, Japan; Hiroshi Hara and Atelier
environments, and in emergencies caused by BNK
natural disasters and human will. Examples of
adaptive kinetic buildings are usually found What is kinetic architecture? Perhaps influenced
among those that have been referred to as by the Pop Art movement, Zuc and Clark coined
intelligent, smart, responsive, dynamic, and active the term “kinetic architecture” offering the first
[Fox, 2003; Kroner, 1997]. These characteristics formal definition: “an architecture free to adapt
are mainly exhibited by structures that are foldable to changes taking place within the set of pressures
and unfurlable, collapsible or demountable, acting upon it and the technology that provides
deployable, transportable and portable, mobile and the tool for interpretation and implementation of
nomadic, mass-produced, ephemeral, and these pressures.” [1970, p.11] Their vision was
transformable. far-reaching, incorporating mechanisms, sensors
and embedded processors allowing the structure
One may argue a building that moves in whole, or to react automatically to change, as in the work of
in parts, is like a machine. A machine applies Fox and Oosterhuis, among others. Their proposal
energy to do, or produce work. It uses energy. fits between mechanistic and organic models for
Kronenburg references Vitruvius’ siege towers as architecture. Yet, the ‘biologic’ argument takes
one of the earliest precedents to mobile precedent in their thinking, because they surmise
environments, [Kronenburg, 2002, p.140] with “every principle of kinetic structure may be found
kinetic components. A siege tower was a tall frame in the physiology and morphology of the human
several stories high with an interior staircase set body.” [1970,p.20] Current research conducted on
on a platform with wheels pushed by soldiers, and the biomechanics of organisms further supports
in some cases special mechanisms facilitated their position. Of the eight groups or types of
movement. The exterior was plastered with lime kineticism advanced by Zuc and Clark, only five
or padded rawhide, depending on whose account are truly kinetic within the definition established
one consults. In Vitruvius’ time, a siege tower was at the beginning of this section. These five can be
a building constructed under the purview of an further organized into three groups: mechanisms
architect. Today we categorize objects functionally (kinetic components), reversible and non-
into succinct groups, where siege tower is war reversible self-erecting structures, and deformable
machine. Yet, Paul Shepheard writes “a Boeing or transformable structures. To date these continue
[airplane] is not a building, but as a refabrication to be the main types.
of the world, it is architecture.” [2003, p.vii] He
proposes architecture has three scales: landscape, Zuc and Clark’ self-erecting structures find a new
building, and machines. The difference between form in Scott Howe’s kit of parts composed by
the three is time, where landscapes are “long term geometric primitives or “motion primitives,”
strategies,” and on the other side of the spectrum intended for pre-fabrication and automated
machines “have short lives and are closely fitted construction. The kit defines a form of “kinematic
to their purpose.” [Shepheard, p.33] Therefore, architecture that includes mechanisms to construct
four conditions exists where buildings can still be itself, or to change the configuration or form of
conceived of as machines exhibiting kinetic the structure over its lifetime.“ [Howe, 2001] It
capacity, only excluding full mobility to avoid consists of: joint-based, panel-based, module-
infringing into the territory of vehicles. based, and deployable. Howe’s proposal aims for
efficiency and accuracy. [2001] Howe has defined
Buildings as machines: three levels for achieving construction automation
Automated Car Tower, Volkswagen Autostadt, based on the combination of kinematic and robotic
Germany mechanisms:

Kinematic Mechanisms: “a structure containing interaction between environment and building is

two or more elements that have the capacity to clearly stated in Kroner’s definition for intelligent
alter their configuration in relationship to each architecture: “built forms whose integrated
other based on a known or given transformation.” systems are capable of anticipating and responding
to phenomena, whether internal or external, that
Robotic Mechanisms (RM): “a structure containing affect the performance of the building and its
one or more kinematic mechanisms, one or more occupants.” [Kroner, 1997] Zerefos et al proposed
actuators, one or more sensors, and a controller.” the concept of responsive architecture: “Dynamic
A RM produces work. modifications of the external envelope of the
building and of the interior spaces, according to
System-Wide Work Cells: “Complex systems the needs of the occupants, as well as the extended
consisting of multiple kinematic and robotic use of new ‘intelligent’ materials that extend the
mechanisms require coordinated behavior and effectiveness of existing control systems.”
work areas.” [Howe, 2001]. [Zerefos et al, 2000; based on Atkin, 1988] For
Kroner the occupants are passive, they are affected
Responding to a question on the nature of his by the intelligent system, while for Zerefos et al
design for the 2000 Olympics arch, Hoberman the kinetic function explicitly responds to human
explained the arch was a hybrid between structure needs. The Zerefos’ team also includes materials
and mechanism, where the structure reacts by as active participants in the process of
changing size or shape, and the mechanisms do modification.
the work “yet [the structure] is stable and self-
supporting.” [2004] Hoberman’s definition for Of all versions, Fox’s proposal for “intelligent
kinetic architecture is one offering “the possibility kinetic systems” is the most encompassing:
of movement,” to create “transforming “architectural spaces and objects that can
environments, responsive building elements, or physically reconfigure themselves to meet
interactive public spaces.” [2004] Fox defines changing needs,” in the interaction between human
kinetic architecture as ‘buildings and/or building activity and the conditions of the physical
components with variable mobility, location and/ environment. Fox presents kinetic architecture as
or geometry.” [2003] Terzidis defines kinetic an adaptable interface between people and their
architecture as “the integration of motion into the environments. [2003, p.115] He qualifies change
built environment, and the impact such results as “responsive and adaptive to optimize.” For Fox,
have upon the aesthetics, design and performance a kinetic system is justified by the need for
of buildings.” [2004] Hoberman and Terzidis adaptability or accommodation in any of these
expand the definition beyond space or building, areas: spatial efficiency, shelter, security and
to environment. Both consider motion and its transportability. Fox proposes three categories of
consequences, essentially kinetics. Through the kinetic systems: embedded, deployable, and
use of the term aesthetics, Terzidis implies there dynamic. The embedded exist within a “whole in
is a human imperative, while Hoberman’s use of a fixed location,“ and has the ability to control the
“responsive” and “interactive” environment is less whole. In the deployable, the kinetic is an “inherent
specific. capability to be constructed and deconstructed.”
This second category would be more specifically
In the discussion of various authors kinetic defined as reversible. The dynamic “exist within
architecture intersects intelligent and responsive a larger whole“, but “acts independently“. [Fox,
architecture. The occupants and the idea of 2003] Fox’s intelligent kinetic systems types fit
almost exactly into Zuc and Clark’s scheme.
Table 1. Comparing the types of kinetic structures; Zuc and Clark

[1971] and Fox [2003]

Table 2. Comparing the characterization of kinetic structures according to Zuc and Clark
[1971] and Fox [2003]
Evidence of Zuc and Clark‘s encompassing vision Adaptability and Transformability

for kinetic architecture is also demonstrated by
the fact that almost thirty years later the The idea that adaptive kinetic structures are forms
characteristics of kinetic architecture are basically of intelligent and responsive architecture has been
presented using similar terms: presented thus far. In this section, a connection is
established with transformable structures.
An up-to-date full characterization of a kinetic Definitions for transformable structures have
structure needs to distinguish between: energy varied depending on the knowledge domain, and
source or force required to move, types of the historical period in which they have appeared.
movement, degrees of freedom, level of Fifty years ago, Giedion referred to furniture that
transformation (morphing capabilities) and speed, changed form and purpose as ‘transmutable,’ and
reversibility and permanence, controllability, ‘convertible.’ [1947, p. 434] Eileen Gray in
precision of movement, dynamic elements and collaboration with Jean Badovici had patented the
their function, level of articulation, type of ‘paravent’ window, an operable folding window
articulations, structure, strength, stability, assembled between an inner and outer layer of
geometry, material, and environments manifested. shutters. She had written “a window without a
Any description of a dynamic system must also shutter is like an eye without an eyelid.” [St. John
include the means for energy production or Wilson, p.117] Norbert Wiener, had published the
transformation, energy transmission to produce seminal work where he stated ”the many automata
work, and the storing of energy. These last three of the present age […] coupled to the outside world
elements have been largely ignored in the both for the reception of impressions and the
discussion of adaptive or kinetic structures. performance of actions […] contain sense organs,
effectors, and the equivalent of a nervous system
to integrate the transfer of information from the
one to the other.” [1961, p.43] Giedion, his
contemporary, was not looking at furniture, or at
architecture, as automata. Although aware of
motion as a particularly difficult issue when
designing folding furniture, he was not ready to
consider possible implications on architecture.
Therefore, the kinetic, adaptive or transformable
was mostly absent from the discourse of the time.
It was not until the 1970s, when Jencks borrows

from cybernetics to put forth his predictions for
the future that it resurfaces: “Substitutes for fur,
skin and other anthropomorphic effects will
proliferate until it will be quite possible for
everyone to have a responsive environment.”
[1973, p.112] Certainly, Archigram’s early work
directly references cybernetics, among other
things. Today, Saggio writes about the ‘effective
Figure 2. Conceptual diagram of the adaptive kinetic
interaction process
mutation’ of architecture, or ‘physical interactivity’
when referring to the ability of architecture to

change through controlling lights, moving used gobots transformer robots as an analogy to
partitions, and the increasing responsiveness of the early work of Jones and Pfau. Morphing or
materials and mechanical systems. [2001, p.27] adaptive-shape systems are somewhat akin to
Dorsthorst and Durmicevic define transformable Gumby, where the structure acts as if it was
building as unfinished products, dynamic continuous.
structures “made of pre-made components, which
are assembled in a systematic order suitable for A transformable structure through folding is
maintenance and replacement,” [2003] adjusting created using moves similar to origami, the art of
to users’ requirements and facilitating re-use and paper folding. Fox would categorize these under
recycling. They establish the domains where “deployable kinetic.” A folding or foldable
transformation takes place as: spatial, structural, structure acts as if it was continuous, rather than
and material. This is an area of agreement among an articulated assembly, where folds act as hinges.
many authors. They contribute the notion that the It can consist of rigid planes with contiguous
capacity to transform impacts a “sustainable elastic edges. The edges or folds are elastic because
development in the future.” they permit deformation without breaking. Planes
can also have some tolerance for surface
Three metaphors are used here to discuss the levels deformation, thus accepting a slight curvature.
of transformation of adaptive kinetic structures in Interestingly, elastic edges and planes maintain
relation to general types of geometrical their length and area. They are referred to as
configurations: origami, transformer robot toys, “inextensible” and “developable.” A developable
and the cartoon character Gumby. Liapi [2002] surface is one where “the sum of angles around
and Sánchez [1996] have used origami as a an arbitrary point on the surface is 2 ð.” [Miura,
metaphor for folding structures. Betsky [1990] 1993, p.3] Tents, space structures, such as solar
Figure 3. The three types of transformation of an adaptive kinetic structure

sails, and dECOi’s piston wall project where type is a kinetic structure shaped as a continuous
transformation occurs through “inextensional yet composite body.
deformation” of triangular plates. In this case, the
surface or structure is fragmented and articulated. Adaptive and Active in Engineering
Transformer robot toy figures are made of In structural engineering an example of active
specialized distinct parts and flexible joints. structure are “statically determinate trusses where
Through a sequence of operations, mainly rotation some of the members are linear actuators, enabling
and translation, the toy is transformed, for example the truss to articulate.” [Williams et al 1997, 77]
from robot to airplane. In architecture examples They can generally be defined as controlled
of “transformer” structures are the Schröder/ structures that contain sensors and/or actuators
Reitveld House, the GucklHupf pavilion by Hans highly integrated into the structure, and have
Peter Wörnl, and Michael Jantzen’s M-house. Also structural and control functionality. [Wada et al,
representative of this type is Morphosis’ 1983 1990] Similarly, an adaptive structure has actuators
proposal for the Venice III house, where “weights that allow the alteration of “system states and
and pulleys control sun-sails that respond to the characteristics in a controlled manner.” [Wada et
wind and sun, changing the physical aspect of the al, 1990] It is also defined as a structure that can
house from an unfinished ruin to a temporary tent.” purposefully vary its geometric configuration, as
[Betsky, 1990, p.189] Howe’s ‘motion primitives’ well as its physical properties. [Miura and Furuya]
belongs to this group. A transformer-like structure And in more detail: “structural systems whose
is submitted to reversible change, and therefore it geometric and inherent structural characteristics
is kinetic. It is fragmented and articulated. can be changed either through remote commands
and/or automatically in response to external
The third metaphor, the 1950’s cartoon character stimulations.” [Miura, 1993] The main difference
‘Gumby’, helps visualize the behavior of an between active and adaptive is that the latter
adaptive-shape structure. Gumby was made out responds to external conditions, in addition to
of clay with a soft wire inserted into the limbs and commands.
head to maintain the pose while the film frame
was produced. Although uncommon in Research into active and adaptive structures in
architecture, this metaphor has some relevance to engineering considers kinematics and dynamics,
aerospace engineering, particularly in the design mechatronics, adaptronics, robotics, and bionics.
of adaptive-shape airplane wings. For example, Mechatronics comprises “products and processes
in the adaptable or morphing wing, skin and with moving parts requiring manipulation and
structure act together to change while maintaining control of dynamic constructions to a high level
the strength needed to keep the wing functioning. of accuracy.” [Giurgiutiu et al, 2002, p.170] This
The morphing wing needs to have a memory of is further complemented by the field of
the various states adopted so that it may return to adaptronics involving “the creation of material
those states multiple times. Therefore, the change systems with intelligence and life features
needs to be reversible without degradation of integrated in the microstructure of the material
formal integrity or structural function. The system.” [Giurgiutiu et al, 2002, p.170] All of
materials necessary to build a fully functioning these must be carried on into the study of adaptive
morphing wing are not yet available. Escrig kinetic buildings.
describes a type of mobile structure as: similar to
an “animal body with bones, tendons and Simulation Environment for Adaptive Kinetic
muscles.” [1996, p.18) His descriptive analogy Structures
represents a Gumby-like structure. Kas Oosterhuis
Associates ‘Trans-ports ‘projects 2001 and the Fox has argued kinetic structures must be seen as
‘Muscle’ are close approximations to this third parts of larger systems, and “to achieve this it is
metaphor. The first project is “a spaceframe necessary to use advanced computational design
composed of pneumatic bars that are individually tools,“ among other resources. [2003] This
controlled by software so that they work together assessment is shared by colleagues both in
like the filaments in a muscular bundle.” [Zellner, architecture and engineering. [Angelov et al, 200]
1999,p.73]; The second is a pressurized volume Liapi argues that in engineering there are no
covered by “a mesh of Festo muscles.” The Gumby standards for the design and analysis of

transformable structures. [2000, p. 267] She including digital mock-up in preparation for
proposes in the case of origami-based assembly concluding with the simulation of the
transformable structures: “animated simulations manufacturing facility system required. [Murphy
of the transformation process during folding can et al, 2001] In this implementation model for a
identify problems in the initial geometric simulation environment ‘digital prototyping’ is
conception (…), and can be used for further only one of many phases. The Murphy et al
morphological explorations.” [Liapi. 2002, p.385] proposal concurs with the main goal of the
The modeling and simulation of kinetic structures performance-based design in architecture: “to
is complex. Frei Otto, a strong advocate for predict the behavior of a building from conception
physical prototypes conceded some structures can to demolition.” [Malkawi, 2004] But, it is often
only be accurately established through digital noted that not much has been offered relative to
modeling. [Dickson, 2004] Yet, as has been this type of comprehensive simulation
pointed out before, the precision of future digital environments. [Axley, 2004, p.4]
models depends on the testing of the physical
models of unbuilt proposals not yet analyzed. Design challenges facing whole system simulation
applicable to adaptive kinetic structures:
A simulation is goal-oriented, it serves a purpose.
It can assist the “design, analysis or optimization” Differences between the representation needed by
of a system. [Birta et al, 1996, p.77] To be the simulation tools and that generated in the
simulated, a system needs to be conceptualized. A geometric modeling (CAD) environment.
simulation acts on a model of reality or proposed [Mahdavi, 1999; Shepard, 2004] The ‘generative
reality. The models used in the simulation need to components’ initiative promises to facilitate
be validated to confirm they accurately represent sharing data.
the system under investigation. They also need to
be verified to determine if the concept of the No ”infrastructure for distributed simulation
system has been correctly translated into the environments, and transaction data between
simulation. [Birta et al, 1996, p.77] Expert distributed simulations.” [McLean et al, 2001]
knowledge has to be identified and coded in, and
experimental data needs to be collected. Physical “No satisfying solution for the simulation of the
prototypes after all are necessary. entire system in one software tool,” for multi-body
systems and Computer-Aided-Control
Kieran and Timberlake [2004, p.59] have recently Engineering. [Gerl, 2003; Williams, 1997]
made the case for the transfer of aircraft modeling
methods to architectural design. They argue Need the “ability to define simulation models of
simulations provide a “whole model”, where various levels of ‘fidelity’ from early in the design
information about the factors constraining the process and to have the information determined
design have been “embedded” in each part. They by those analyses influence the design process.”
refer to simulation as an enabling device to ensure [Shepard et al, 2004]
that the parts create a “unified whole.” [Kieran
and Timberlake, 2004, p.65] Yet, the “Mechanisms for multidisciplinary analyses and
implementation of such an integrated environment ‘performance based comparison procedures’ that
has not been fully realized. facilitate the collaboration of highly diverse
interdisciplinary teams.” Also shared
In the aerospace industry, a call has been made representations. [Augenbroe, 2002; Hu & Fox,
for the integration of simulations “to cover 2003]
lifecycle, from conceptual design, through
predefining detailed manufacturing processes to The purpose of developing a digital prototyping
the final assembly without any physical environment for adaptive kinetic architecture
application required.” [Murphy et al, 2001,p.829] would be to evaluate the effectiveness of its form
This entails developing a system or environment and behavior as a system to achieve the design
for ‘digital manufacturing’ where different types objectives. The behavior would not only involve
of simulation types interact. The simulations motion, but structural, environmental, and control
progress from conceptual to detailed design, to performance, including human interaction.
tolerance, mechanical, and ergonomic simulations, Commercially available packages for engineering

applications simulate structural performance,

which may include kinematics and dynamics,
interference detection, and motion volumes, and
control system, or environmental performance,
and ergonomics, but not all. Below is a list of
currently available software with some digital
prototyping capabilities that must be carefully
evaluated for possible transfer to architecture. Table 3A
A model of an integrated simulation environment,

currently in development for application in the
automobile industry, the “Simulation Environment
for Engineering Design” or SEED [Shepard et al,
2004], can be useful for digitally prototyping
adaptive kinetic structures. It is built on the
concept that a holistic simulation environment is
composed by three already existing components
that interact with SEED’s four new components.
Furthermore, it is based on the premise that such
an environment is to be shared by an
interdisciplinary team, must be accessible to
designers, and must be based on multiple
representations of increasing detail as they are
affected by the results of the simulation. Table 3B
Table 4. Commercially available software with some digital prototyping capabilities

Conclusion Acknowledgment
When placed in the architectural domain, adaptive A summer spent at NASA Langley Research
kinetic structures are extremely complex systems. Center in 2001, through an ASEE Summer Faculty
Investigating the modeling and simulation of Fellowship, provided in-depth exposure to
adaptive kinetic structures provide opportunities emerging concepts in the area of adaptive systems.
for joint ventures in design among multiple Informal conversations with Prof. Tuba Bayraktar
disciplines. Simulation-based design, performance on the complexity of integrated simulation
based-design, digital prototyping all become part environments from the perspective of mechanical
of the same. Most importantly, this effort extends engineering have allowed me to see another side
the vision of the comprehensive building of the field. Finally, I appreciate V.M. Price’s
information modeling initiative. editorial comments on the clarity of the ideas and
the manuscript.
When we consider architecture as a dynamic
system we are forced to confront issues of human
power, space control, environmental manipulation,
material economy, operational effectiveness, and
energy investment. An adaptive kinetic structure
materializes energy because it approximates a
machine, and because it can be seen to mimic a
living being. The intent here is not to present them
as new commodities, but as ideas architecture still
needs to catch up to and grapple with, if it intends
to remain present and in conversation with its
times, its needs, its machinery.
An integrated platform for evaluating design in

architecture is needed. Adaptive kinetic structures
provide the subject through which experimental
architectural ideas may be tested. It shifts the
discussion from digital design to digital
prototypes. It recognizes the need, and provides
the means for, collaborative interdisciplinary
work. It pushes for a shared language that connects
and challenges professional expertise without
reducing its complexity or its possibilities.

References Fox , Michael A. (2003). Kinetic Architectural

Systems Design. In Robert Kronenburg (Ed.)
Angelov, C., & R.V.N. Melnik, J. Buur. (2003). Transportable Environments 2. London & New
The synergistic integration of mathematics, York: Spon Press.
software engineering, and user-centred design:
exploring new trends in education. Future Gerl, Johannes. (2003, June). Overview about the
Generation Computer Systems, (19), 1299-1307 Recent Developments for Combined Control
System and Mechanics Simulation in Automotive
Augenbroe , Godfried. (2002). Trends in building and Railway Engineering. Paper presented at a
simulation. Building and Environment (37), 891– meeting on Simulation: An Essential Tool for Risk
902. From Elsevier Database. Management in Industrial Product Development,
Lyon: France.
Axley , James W. (2004). Design and Simulation
in Innovative Practice. Journal of Architectural Giedion, Siegfred. (1947). Mechanization Takes
and Planning Research, 21 (4), 273-291. From Command: a contribution to anonymous history
Wilsonweb database. (1969 edition). New York: W.W. Norton and Co.
Betsky , Aaron. (1990). Violated Perfection: Giurgiutiu , Victor, & Abdel-Moez E. Bayoum,
Architecture and the Fragmentation of the Greg Nall, (2002). Mechatronics and smart
Modern. New York: Rizzoli. structures: emerging engineering disciplines fore
the third millennium. Mechatronics, 12 (2), 169-
Birta, Louis G. (2003). The Quest for the Modelling 181.
and Simulation Body of Knowledge.
Retrieved from www.csi.uottawa.ca. Hoberman, Charles. Retrieved August 16, 2004
from www.hoberman.com.
Birta, Louis G., & F. Nur Özmizrak. (1996). A
Knowledge-Based Approach for the Validation of Howe, Scott. (2001). Kit-of-parts and Automated
Simulation Models: The Foundation. ACM Construction, Department of Architecture, The
Transactions on Modeling and Computer University of Hong Kong. Retrieved on March
Simulation, 6 (1), 76-98. From Science Direct 2005 from thegreenroom.arch.hku.hk.
database.
Hu , Catherine, & Michael A. Fox. (2003). Starting
Dickson, Michael. (2004) Frei Otto: Researcher, from the Micro: A Pedagogical Approach to
Inventor, and Inspired instigator of Architectural Designing Responsive Architecture. CAADRIA
Solutions. AA Files, (50). 2003. Retrieved from CuminCAD database.
Dorsthorst, Bart Te, & Elma Durmicevic. (2003, Jencks, Charles. (1973). Architecture 2000,
May). Building’s transformation capacity as the Predictions and Methods (2nd printing), New York
indicator of sustainability; transformation & Washington: Praeger Publishers.
capacity of flexible housing. Paper presented at
the 11 th Rinker International Conference: Kieran, Stephen, & James Timberlake. (2004).
Deconstruction and Material Reuse, Gainesville: Refabricating Architecture, New York: Mc-Graw
Florida. Retrieved from www.cce.ufl.edu/pdf/. Hill.
Escrig, F. (1996). General survey of deployability Kronenburg, Robert. Houses in Motion (2 nd

in architecture. In F. Escrig and C.A. Brebbia edition), Great Britain: Wiley-Academy.
(Eds.), Mobile and Rapidly Assembled Structures
II, 2nd International Conference on Mobile and Kroner, Walter M. (1997). An intelligent and
Rapidly Assembled Structures, Computational responsive architecture. Automation in
Mechanics Publications. Construction, (6), 381-393.

Lee, Kunwoo. (1999). Principles of CAD/CAM/ Sánchez del Valle, Carmina.(1996). Transformable
CAE Systems, Addison Wesley Longman. Folding Space. ACADIA Conference Proceedings:
Design Computation: Collaboration, Reasoning,
Liapi, Katherine A. (2002). Transformable Pedagogy. (pp. 45-54), Tucson: Arizona / USA.
Architecture Inspired by the Origami Art:
Computer Visualization as a Tool for Form Shepard, M.S., & M.W. Beall, R.. O’Bara, B.E.
Exploration. Proceedings of the 2002 Annual Webster. (2004). Toward simulation-based design.
Conference of the Association for Computer Aided Finite Elements in Analysis and Design, (40),
Design in Architecture, Pomona: California/USA. 1575-1598.
Mahdavi, Ardeshir. (1999). A comprehensive Shepheard, Paul. (2003). Artificial Love: A Story
computational environment for performance based of Machines and Architecture, Cambridge: The
reasoning in building design and evaluation. MIT Press.
Automation in Construction, (8), 427–435.
St. John Wilson, Colin. (1995). The Other
Malkawi, Ali M. (2004). Developments in Tradition of Modern Architecture, London:
environmental performance simulation. Academy Editions.
Automation in Construction, (13), 437– 445.
Retrieved from Elsevier database. Terzidis, Kostas. Course description for 2309:
Kinetic Architecture. Course Bulletin Schedule
McLean, Charles, & Swee Leong. (2001). The Fall 2004, Harvard Design School. Retrieved from
Expanding Role of Simulation in Future www.gsd.harvard.edu.
Manufacturing. In B.A. Peters, J.S. Smith, D.J.
Medeiros, & M.W. Rohrer (Eds.). Proceedings of Wada, B.K., & J.L. Fanson. (1990). Adaptive
the 33rd conference on Winter Simulation, Session: Structures. Reprinted from Journal of Intelligent
Future of simulation, (pp.1478-1486), Arlington, Materials Systems and Structures, (1), 157-174.
VA. Retrieved from www.portal.acm.org. In Selected papers on smart structures for
spacecraft, SPIE Milestone series, (MS 167).
K. Miura. (1993). Concepts of Deployable Space
Structures. International Journal of Space Wiener, Norbert. (1961) Cybernetics or control
Structures, 8 (1-2). and communication in the animal and the machine
(2nd edition). New York & London: The MIT Press
Moore, P.R., & J. Pu, H.C. Ng, C.B. Wong, S.K. & John Wiley & Sons.
Chong, X. Chen, J. Adolfsson, P. Olofsgard, J.-O.
Lundgren. (2003). Virtual engineering: an Williams, Robert L., & Rex J. Kuriger.(1997).
integrated approach to agile manufacturing Kinematics, statics, and dexterity of planar actives
machinery design and control. Mechatronics, (13), structure modules. Automation in Construction,
1105–1121. (7), 77-89.
Murphy, Craig A., & Terrence D. Perera. (2001). Zellner, Peter. (1999). Hybrid Space: New Forms
Role of Simulation in industries: The Definition in Digital Architecture. New York: Rizzoli.
and Potential Role of Simulation Within an
Aerospace Company. In B.A. Peters, J.S. Smith, Zerefos, S. C., & A. M. Kotsiopoulos, A.
D.J. Medeiros, & M.W. Rohrer (Eds.). Pombortsis. (2000). Responsive Architecture: An
Proceedings of the 33rd conference on Winter Integrated Approach for the Future. Proceedings
Simulation, Session: Manufacturing applications, ACADIA 2000: Eternity, Infinity and Virtuality,.
(pp. 829-837), Arlington, VA, 2001. Retrieved Washington D.C.
from www.portal.acm.org.
Saggio, Antonino, (2001). New Subjectivity. In

Peter Cachola Schmal (Ed.). Blobmeister,
DigitalReal, Birkhauser.

Chapter 2
Smart Systems, Techniques and Environments
Jaewook Lee, Yehuda E. Kalay
Collaborative Design Approach to Intelligent Environments
Mark Cabrinha
From Bézier to NURBS: Integrating Material and Digital Techniques through a Plywood
Shell
Daniel Barker, Andy Dong

A Representation Language for a Prototype CAD Tool for Intelligent Rooms
Matthew G. Fineout
The Tower of Babel: Bridging Diverse Languages with Information Technologies
Thomas Modeen, Christine Pasquire, Rupert Soar

Design Ground - An Iconic Tactile Surface
Tsou Jin Yeu, Chan Yi Lee, Mak Kwok Pui, Ru Xu Du, Liang Jian, Yeung Kim
Applying Scientific Simulation to Integrate Thermoelectric Conductor Module into
Architectural Design – Smart Wall for Thermal Comfort

J. Lee. Y.E. Kalay / Collaborative Design Approach to Intelligent Environments
Collaborative Design Approach to Intelligent

Environments
Jaewook Lee1, Yehuda E. Kalay2

1
University of California, Berkeley
2
University of California, Berkeley
Abstract
Intelligent environments are buildings and other settings that can recognize the changing needs of their users
and/or the changing nature of their context, and respond to them by adjusting some key environmental
parameters (temperature, light, sound, furnishings, etc.). Unlike the currently common approach, which is
based on systems theory (i.e., adjusting the parameters of the environment to match some pre-defined use
profile), the approach proposed in this paper is based on dynamic, collaborative design: it views the (built)
environment as comprised of multiple independent object-agents, each of which is responsible for one small
aspect of the environment. Each can sense the immediate changes pertaining to its domain of responsibility,
and propose corrective measures, which are negotiated with other agents to form a collective response. The
paper hypothesizes that such an approach can be made more context-sensitive and dynamic, is easily scaleable,
and can respond to the needs of multiple different users of the environment at the same time. The paper
presents the rationale for developing the multi-agent approach, its hypothetical implementation, and its
application to hypothetical case studies.

Introduction
possible conflicts may arise between the needs and
Once man-made environments, such as buildings, activities of multiple different occupants of the
automobiles, and urban places are built, they are same space at the same time; and how these
often ‘frozen’ in one or a few interchangeable conflicts can be resolved.
configurations, intended to support a single
activity or environmental condition, or a number This paper describes research that attempts to
of closely related activities/conditions. When the answer some of these questions, by providing a
activities or the conditions do not match those for framework for developing intelligent
which the environment was designed, adjustments environments using a collaborative design
are needed. They can take the form of opening or approach, rather than the systems approach. The
closing a window, turning on the light, re- approach proposed by this research is derived from
arranging the furniture, or remodeling the building. the observation that adjusting the environmental
Each adjustment requires conscious action by the and physical parameters of a building after it has
occupant(s). ‘Intelligent environments,’ on the been built can be compared to the process of
other hand, can actively support diverse human designing the building in the first place (before it
activities and different environmental conditions is built). During the design process, the individual
by automatically and dynamically adjusting their participants (architect, engineers, client, etc.) of a
configuration to meet the changing needs of their design team interact with one another in ways that
occupants without explicit human intervention. eventually lead to a joint solution, influenced by
the goals, contributions, and constraints of each
The idea of making buildings “intelligent” has participant. Modern design (and organizational)
been first suggested by Nicholas Negroponte, in theories (Benne, 2004; Kalay, 2004) claim that this
his influential book “Soft Architecture Machines” negotiated process is not based on a top-down
(1975). Such buildings will be able to adjust their systems approach, as cyberneticists would have it
environmental and configurational settings to (“the building is a machine,” according to Le
match the occupants’ needs and activities without Corbusier (1927)), but rather on methods of
explicitly being ‘told’ to do so through manual collaborative decision-making that respects the
environmental controls. As such, they will be able needs and wishes of each participant. Such
to perform optimally for a wide range of activities dynamic, negotiated, and collaborative adjustment
and needs, rather than an “average” daily activity of building parameters typically ends when the
or need set at the time of their design. Negroponte building is constructed, and it is “locked” into one
went as a far as stating that such intelligent or a few fixed states. Thereafter, a system-based,
environments will obviate the need for architects: mechanistic method is used to monitor and control
they will “redesign” themselves whenever the need adjustable building parameters (e.g., light, energy,
arises. security, elevators, etc.): the “system” is adjusted
to conform to some pre-conceived schema.
Attempts to make buildings more intelligent have
proliferated since the advent of affordable, The approach proposed in this paper, on the other
ubiquitous computing devices (Intel Research, hand, wishes to extend the dynamic adjustment
2005). These attempts, which focus primarily on of building parameters after it has been built, using
the technical aspects of buildings and information continual negotiation rather than a pre-conceived
systems, are typically guided by a systems schema. One can argue that such a mechanism is
approach, whereby the building adjusts its already in place, in the form of the on-going
parameters to conform to some pre-defined actions taken by the building occupants (turning
schema (e.g., a given temperature, lighting, or on lights, opening windows, re-arranging the
another parameter). This approach, which is based furniture, remodeling the building, etc.). This
on traditional cybernetic principles, focuses on the research proposes to replace the actions of the
physical environment itself, rather than on the human participants with intelligent software agents
dynamic interrelationship between human that will extend the on-going negotiation and
activities and the environments in which they collaborative decision making that characterizes
occur. Little attention is paid to how an the design phase of buildings, thereby off-loading
environment (e.g., a room, a building, etc.) should these decisions and actions from the human
behave as users and their activities change; which occupants.

The research discussed in this paper focuses on The Building Operation/Environmental Control
the processes that control intelligent environments, approach, on the other hand, places more emphasis
rather than the physical devices themselves that on indoor environmental quality (e.g.,
sense and respond to the required changes. The temperature, illumination, etc.), energy saving, and
design of the sensors and actuators has been building operation. Such building automation
researched by others (Coen, 1998; Holmes et al, shares the concept of ubiquitous computing with
2002; Krogh et al, 2001; Streitz et al, 1998). the approach described above, but aims at
maximizing operational efficiency and thermal
Conventional Approaches to Intelligent comfort by various embedded sensors and
Environments controllers, rather than the effectiveness of the
users themselves. The iDorm project (Holmes et
After Negroponte’s first experiments in the 1970s al, 2002) uses various networked devices that
with the adaptation of computer technologies to monitor the current state of a dorm room and adjust
the building industry, many attempts were made environmental conditions to meet user’s
to develop intelligent environments. Two major preferences. A set of distributed sensors collects
approaches emerged: equipping the environment various data about the room’s condition and sends
with devices that can help users do their job more them to the iDorm main control unit. The system
effectively (e.g., process information, etc.), and adjusts the connected effectors (e.g., heater, cooler,
approaches that adjust the environment’s own door lock, blind, window, etc.) based on the
qualities (e.g., light, temperature, etc.). The first received data. Similarly, the ACHE (Adaptive
approach is known as the Information Systems Control of Home Environment) project (Mozer,
approach, and the second as Building Operations/ 1998) is a house equipped with an intelligent
Environmental Controls approach. control system. It uses neural network-based
predictors that control basic residential comfort
According to the Information Systems approach, systems including lighting, air heating, water
an environment is considered a means (a machine) heating, and ventilation. The system can learn the
for processing information. Most such systems users’ preferences conveyed through their manual
have been based on ubiquitous computing adjustment of lights or thermostats and strives to
technology that inserts programmable microchips balance their comfort with energy savings.
into building components and appliances to
monitor and control them through networked Two fundamental drawbacks of the conventional
communication (Weiser, 1993). The primary goal approaches to developing intelligent environments
of this approach is to make computers invisible are the absence of an overall methodology for
and become part of the environment, mostly at environment-wide behavior control, and the lack
room-scale. For example, in the i-Land project of attention to the environmental impacts on
(Streitz et al, 1998), a set of roomware human behavior.
components, such as “DynaWall” (an interactive
electronic wall), “CommChairs” (mobile and In a broad sense, a setting modification by an
networked chairs), and “InteracTable” (an intelligent environment can be viewed as a design
interactive table), creates a collaborative work activity that transforms a present situation into a
environment that supports office workers and their desirable one (Kalay, 2004; Rittel, 1973; Simon,
activities. In the case of the Intelligent Room 1984): a (human) designer, in the course of solving
(Coen, 1998), computer vision and speech a design problem, identifies a design problem and
recognition systems with built-in Artificial the goals a design solution should achieve,
Intelligence are used to minimize the number of generates possible solutions by gathering relevant
embedded devices and to identify ordinary human (external) information and using her/his (internal)
activities, creating natural human-computer knowledge-base (e.g., past experiences, reasoning
interactions. Computer devices in this kind of rules, etc.), and evaluates the candidate solutions
intelligent environments are mainly used for by testing and verifying them compared to the
controlling other connected devices to access, store goals and constraints.
and display information similar to conventional
personal computers. Thus, the environment itself This design-oriented view can provide the
is treated as a container, a backdrop for users’ theoretical foundations for the development of
information-processing activities. intelligent environments. The task an intelligent

environment deals with at a given point in time A Collaborative Design Approach to Intelligent
can be considered a dynamic design problem. The Environments
intelligent environment perceives users’ activities
and the current state of the environment (problem Most design problems in the real world are too
identification), determines the goal state and complex to be solved by a single designer. Rather,
constraints that reflects users’ needs (goal design is a collaborative activity, involving
formulation), finds potential environmental multiple specialist participants. Since each
settings that can achieve the goal state and abide participant has limited knowledge and abilities,
by the constraints (solution synthesis), and the design project can only be accomplished by
evaluate the settings and select the best-fitting one the combined efforts of the participants who have
among them (evaluation). particular tasks to complete based on their
specialty. In other words, while each individual
The conventional approaches to intelligent participant maintains autonomy to deal with her/
environments that were discussed above have his individual tasks, coordination (e.g., task
more or less disregarded such design-oriented allocation, scheduling, conflict resolution, etc.) is
view. Thus, design-based models and required to accomplish shared organizational
methodologies for intelligent environments have goals. Therefore, collaboration can be defined as
not yet been proposed. As a result, a design the “agreement among specialists to share their
solution (i.e., setting modification) generated by abilities in a particular process, to achieve the
an intelligent environment, based on assumptions larger objectives of the project as a whole” (Hobbs,
made at an earlier time, may conflict with the 1996).
actual needs of the user(s), and therefore might
be overruled (by manually adjusting such We posit that an intelligent environment can be
parameters as windows, lights, and thermostats), built with compositional objects (e.g., walls, doors,
or lead to dissatisfaction of the users (Arens et al, windows, furniture, lights, etc.) as an ensemble
1997). of autonomous intelligent objects, each of which
knows how to interact with context-specific user
Hence, although intelligent environments have activities. Thus, an intelligent environment can be
been developed to improve human-environment viewed as a team-like organization of multiple
interaction in one way or another by utilizing independent agents. Moreover, a setting
computing technologies, they remain largely static modification performed by the intelligent
and passive entities. They have paid virtually no environment can be viewed as a dynamic
attention to the environmental impacts on human collaborative design activity, in which design
behavior, especially of the differing needs of problems are distributed to multiple participants
multiple simultaneous users of the same (i.e., intelligent objects) and solutions are
environment. The drawbacks of conventional synthesized through collaboration and negotiation
approaches can be summarized as follow: among them. This approach can overcome the
limits of the knowledge possessed by individual
· Lack of design-oriented view and theo- intelligent agents, while the division of labor
retical models for controlling the envi- makes the system overall simpler and more
ronment. responsive to unexpected needs of the users. Since
each intelligent object comprising the environment
· Prone to conflicts between users and en- is functionally, spatially, temporally, and
vironments. informationally bounded (i.e., has limited
knowledge and effect), a shared organizational
· Less attention to the effects of spatial goal (i.e., environment-wide modification) can
quality of the built environment on hu- only be accomplished by combining modifications
man behavior. applied to many individual objects. The division
of labor and individual autonomy, on the other
· No ability to deal with differing, even hand, can improve the efficiency of the overall
conflicting needs of multiple simulta- performance of the intelligent environment by
neous users. reducing the cost of information processing.

In general, when the uncertainty of the task Implementation with a Multi-Agent System
increases, the amount of information that needs
to be processed among decision-makers tends to Multi-Agent Systems
increase. In the case of an organization built
according to a mechanistic model, where tasks are Considering the difficulties in constructing
centrally preplanned and executed with strong intelligent environments and the drawbacks of the
hierarchical authority, the increased uncertainty conventional approaches, a method is needed that
of the task may reduce the degree of organizational can be responsive to users and user activities, and
performance due to the limits of capacity in that can efficiently handle the conflicts that arise
handling exceptions and processing information in the course of its solution generation. We have
(Galbraith, 1977) — a phenomenon that was well- found that a multi-agent approach can be a
demonstrated by the I-CAAD system (Pohl et al, promising method for developing intelligent
1994). In order for the organization to cope with environments.
growing uncertainty, the organization needs to
either increase its capacity of information An agent can be defined as any entity that can
processing or decrease amount of information perceive its environment through sensors and act
passed among the members to accomplish the task. on that environment based on its own reasoning
A team-type organization is a case in point: capability (e.g., a human agent, a robotic agent, a
individual members have more or less self- software agent, etc.) (Russell et al, 2003). Thus,
contained tasks. Hence, the amount of information autonomy, interactability, and adaptability are the
transmitted between members is reduced, essential attributes that an agent should have. This
compared to a hierarchically structured concept of autonomous agent has been a core
organization with rigid rules. research subject in Artificial Intelligence and
widely used in many industries including robotics,
Although collaboration can overcome individual process control systems, email clients, and search
limits to accomplish shared goals, it might also engines, for the purpose of developing intelligent
induce inter-personal conflicts due to differences applications.
in perspectives, goals, or knowledge among the
participants. In a collaborative design process Agent-based computing has the potential for
where each participant has her/his own tasks to conceptualizing, designing, and implementing
do, satisfying the goals of one individual may complex multi-user distributed systems (Jennings,
interfere with satisfying the goals of another 1999). In general, agents can be built in any
individual (Kalay, 2004), and this may prevent imaginable environment. Their behaviors are
satisfying the overall organizational goals. strongly dependant on the nature of a task
Similarly, when objects are constructed as environment. In theory, for any task environment,
intelligent entities that know how to behave either single or multi-agent systems (MAS) are
according to built-in reasoning capabilities, each possible. A single-agent may work well when a
one of them may perceive the same situation task environment is simple, small, and/or static,
differently from other intelligent objects, due to whereas multi-agent systems are more appropriate
their spatial and temporal boundedness. This may for complex, large, and/or dynamic environments
lead to behavior conflicts between objects. In an (Russell et al, 2003). The power of MAS lies in
intelligent environment in which multiple objects, the division of labor and the cooperation of agents
users, and activities interact with one another, like human organizations. Whether organizational,
conflicts between various entities need to be physical, or computational, the most basic
efficiently resolved to properly modify the setting technique for tackling any large and/or complex
of the environment as a whole, much like partial problem is to “divide it into smaller, more
design solutions proposed by the multiple manageable chunks” (e.g., individuals, mechanical
participants are tested and verified for consistency components, software modules, etc.) (Booch,
and ability to meet the goals and abide by the 1994). As such, multiple agents can represent the
constraints. When approaching an intelligent decentralized nature of the problem, multiple loci
environment as an organization of multiple of control, multiple perspectives, or competing
intelligent objects, coordination and interests. Within an agent organization, the agents
communication among the objects are major need to interact and negotiate with one another to
concerns.

achieve their individual as well as common goals object can be made to ‘know’ when and how to
(Jennings, 1999). invoke its behavior. Such programmed objects can
be viewed as intelligent (object) agents.
Such MAS-based approach has recently been
applied to building intelligent environments. In In a multi-user environment (e.g., an office), each
these environments, multiple computational agents user has her/his own preferences for
communicate with one another to control various environmental settings. These preferences can also
building components. A system developed by be programmed and collected in a User Profile,
Xerox PARC (Huberman et al, 1995) utilized a which stores user preference including user ID,
market-based MAS for managing the thermal object IDs, property variables, and their values
conditions of an office building. In this system, (e.g., the kind of music she likes, the lighting level,
individual temperature controllers (i.e., multiple heat/humidity levels, height of the chair/desk, etc.).
agents) of the offices bid to ‘buy’ or ‘sell’ cool or A user profile can be encoded in a card key or a
warm air by participating in an auction that is badge that can be read by objects through wireless
moderated by a central computer auctioneer, a communication (e.g., Bluetooth, RFID, etc.).
specialized agent. Similar MAS approaches are Similar approaches have been used in designing
found in other applications for the development intelligent environments (Sharples et al, 1999;
of intelligent environments (Boman et al, 1998; Colley et al, 2001). Objects will need a mechanism
Coen, 1998; Colley et al, 2001). The primary that can identify the user’s profile, and combine it
concern of these MAS-based intelligent with their ability to sense the environment to
environments has been the interaction of the agents invoke the appropriate action.
themselves and, as a result, the aspects of human
behavior and their relationship to the built
environment have largely been neglected.
In our proposed model of intelligent environment,

each building component is represented as an
intelligent or smart agent that knows how to
behave given any activity of any user, and has the
ability to perceive contextual changes of the
environment and adjust its behavior in accordance
with its immediate context to better support users’ Figure 1. User Profile + Activity Profiles + Object Profiles
context-specific activities. A layered multi-agent = Setting Modification
approach can efficiently endow an environment
with intelligence by organizing agents and
controlling their behaviors. Built-in conflict In addition to object profiles and user profiles, a
resolution mechanisms will minimize conflicts and third kind of a profile, Activity Profile that
ensure environment-wide behavioral consistency. describes the activity of the users is needed,
because the same users may perform different
Object, User, and Activity Profiles activities at different times and have different
setting preferences for their activities. With object,
The first step in designing an intelligent user, and activity profiles, the environment can
environment is to embed processors and identify users and their activities, and thus modify
mechanisms within objects to allow them to sense the settings of the environment accordingly
and respond to user activities. For example, the (Figure. 1).
door of a room can open itself when a user
approaches. This behavior must be programmed Design Activity of Object Agents
into the object (i.e., when and how to respond).
Such object behavior description can be stored in Problem Identification and Goal Formulation
the form of a profile, Object Profile. In addition
to an object profile, each object requires a control The design activity of individual object agents of
interface that actually generates actions based on intelligent environments is initiated by perceiving
external stimuli and the object’s profile. Thus, with a user or users. When an object agent detects a
an object profile and a control interface, every user in the environment, the agent retrieves a part

of the user’s profiles (i.e., user and activity profile), space of the object agent into the goal space. The
which is related to the object that the agent next step is to determine the solution state within
controls, and loads it in the memory of the agent. the goal space and select an action or a set of
Each user profile contains a set of requirements actions (means) from the object profile, which can
that describe her/his preferred configurations of transform the current state of the chair to the
the environment. User requirements may include solution state (e.g., change the height of the chair
functional (on/off, open/close, etc.), spatial from ‘x’ to ‘y,’ etc.). Before an actual modification,
(position, size, etc.), environmental (color, the chair agent needs to test and verify the selected
temperature, lighting, etc.), or informational action(s), which is the next phase of the design
(music style, etc.) properties of objects themselves activity.
as well as interrelationship(s) with other object(s)
(in-between distance, etc.). Each property (or Evaluation and Confirmation
variable) of an object has a value or a range of
values that a particular user prefers. Once it is The evaluation phase of the object agents is not
processed, it becomes a part of knowledge-base much different from that of human designers. The
for the setting modification of the object agent. solution state sought from the previous phase,
An object agent formulates a set of goal states and which is represented as a set of actions, is
their constraints by processing the user and activity evaluated by the object agent.
profile of the detected user.
In intelligent environments, the design decision
Since the initial problem space of each object agent of the object agent should be made within the
contains all the possible states (i.e., all the current context of the environment. In other words,
properties and their possible values) of the object, the object agent has to test the fitness of the
the process of goal formulation can significantly proposed solution to the goal state in relation to
reduce the search space of the object agent (i.e., other objects in the same environment before
individual users tend to have a limited number of making any modification to the object. This is
activities and preferred environmental conditions). because the objectives of the individual object
For example, when the object agent of ‘Chair-A’ agents are not totally independent of each other.
detects ‘User-A,’ it loads the portion of Chair-A When they are combined to modify the setting of
from the user and activity profile of User-A on its the intelligent environment as a whole, design
working memory. The loaded profile contains a decisions of an object agent to achieve its own
set of properties and their values (or ranges of goal may interfere with the goal of another object
values) that represent User-A’s preferred state(s) agent. Again in the previous example, the position
(i.e., the goal state(s)) of Chair-A. It may also and height of the chair object may be constrained
include relationship(s) with other object(s) (e.g., by those of a desk object in the same environment,
distance between the chair and a desk, etc.). When which may lead to a conflicting situation between
the agent perceives an activity of the user, the the two. Thus, the object agents must verify (or
search process for a setting the activity will follow check) the side-effects of their actions, before
thereafter. confirming them. If any conflict happens, an
appropriate modification should be made to
Solution Search and Synthesis resolve the conflict.
The design method that object agents use for their Uncertainty of the Design Problem
solution search and generation is means-ends
analysis. In the above example, when the user The ill-structuredness of design problems (Simon,
(User-A) initiates an action on the chair (Chair- 1984; Rowe, 1987) can also be found in the task
A), the first step for the setting modification of with which the object agent deals in intelligent
the object agent is to identify the type of the user’s environments. As ill-structuredness comes from
current activity. If the user’s current activity is the uncertain nature of a design problem, in an
‘Office Work,’ the agent looks up the activity intelligent environment, uncertainty is mostly
profile loaded on the working memory and induced by the dynamic nature of users and their
retrieves the goal state, the user’s desired activities. To cope with this uncertainty, the object
configuration (ends) of the chair for the given user agent requires more capabilities, in addition to the
activity. This process further reduces the problem problem-solving skills.

First, since users and user activities are not static agents exist as a form of an organization or a
elements (i.e., a new user can enter or a new society, social interaction through communication
activity can be initiated), the goal state (i.e., the networks is inevitable for agents to achieve their
needs of particular users and activities) that the own goals and the goals of their organization or
intelligent environment needs to satisfy is not society. Therefore, agents do not need to know
fixed. In other words, the intelligent environment everything, rather rely on other agents to know
must continuously gather information about the things required for their action generation
present state of the environment for its setting (Jennings, 1999). Communication protocols in an
modification by monitoring users and user organization of agents enable agents to exchange
activities. Second, such newly acquired and understand messages and to coordinate their
information (e.g., new user and activity, etc.) may behavior, resulting in systems that are more
keep modifying the goals and constraints, which coherent.
in turn may require changes to the solution (i.e.,
changes of the setting) (Figure. 2), much like a An intelligent environment is composed of a group
human designer keeps updating design solutions of objects of which each has particular role and
in relation to the goals and constraints that s/he behavior controlled by an embedded agent.
discovers through the design process. Thus each Although each object agent can generates its action
object agent has to be able to update its knowledge by monitoring the environment, information that
and behavior accordingly as the state of the is required for the successful action generation
environment changes. Third, different users may may not always be available to the agent. This is
have different needs, which require different because each object agent may perceive only a
environmental settings. These may conflict with part of the environment. For example, a lighting
the settings needed by other users of the same device in an office may require information about
environment, at the same time, thus requiring the current state of user activity from other
negotiation and compromise between different object(s) that the user activity is directly associated
object agents. In order to handle this uncertainty, with (e.g., chair). Moreover, some agent actions
collaboration and conflict resolution mechanisms may also be constrained by the state(s) or action(s)
should be established in the intelligent of other agent(s) (e.g., height constraint between
environment for environment-wide modification a table and a chair: IF the height of Chair-A is ‘a,’
and its consistency. THEN the height of Table-A is ‘b’). Thus a
communication network that contains
communication channels and protocols between
object agents has to be provided for object agents
to efficiently collaborate with other object agents.
Languages that are included in the communication

Figure 2. Design Process of an Object Agent
protocols are major means of communication
between object agents. They are used for encoding
and decoding information to be transmitted from
Design Collaboration of Agents: one object agent to other agent(s). In the proposed
Communication and Knowledge Exchange model of intelligent environments, the set of
profiles can also be regarded as a standardized
Knowledge sharing through communication language that facilitates agent collaboration. For
channels is one of the fundamentals of multi-agent example, when a user sits on a chair to switch from
systems. As computational systems become more work to rest , the change of user activity (i.e., from
complex, dynamic, and larger, the traditional ‘Work’ to ‘Rest’) detected by the chair agent will
Artificial Intelligence approach with a single locus cause the behavior modification of the chair agent.
of internal reasoning and control has shown Thereafter, this change of user activity, as a
limitations in building computational intelligence. message, will be transmitted to the lighting agent
Hence, as an alternative, the multi-agent systems through the communication channel, which in turn
approach with distributed reasoning and control will modify the setting of the lighting object (e.g.,
has been widely applied. Because of the nature of change illumination). This communication process
multi-agent systems in which multiple autonomous is described in Figure 3.

Figure 3. Topology of Agent Communication
Conflict Resolution of Intelligent Environments Object Profile Conflict
An environment is a combination of actors, their This type of conflict generally results from
activities, and the physical (or object) settings, perception or goal difference between object
which interact in various ways. In our model of agents (OAs). An object agent (OA), as a spatio-
intelligent environments, these interactions can be temporally and rationally bounded entity, can only
represented as interactions among the respective perceive a (small) part of the environment, about
profiles (Figure. 4). Although it is possible to which the agent thus has subjective knowledge as
observe six different types of conflict in the well as limited reasoning capacity. As a result,
environment, in this paper we will discuss only different OAs may interpret a same user activity
the three most prominent types of conflict: ‘Object differently. Furthermore, each OA has its own goal
Profile Conflict,’ ‘User Profile Conflict,’ ‘Activity which may be different from that of other OAs.
Profile conflict,’ and their resolving mechanisms. These perception and goal difference are a major
source of Object (Profile) Conflict. In order to
resolve conflicts between OAs, another type of
agent, the Behavior Manager Agent (BMA) —
which is similar to a human coordinator in a
teamwork — has been introduced. The main task
of a BMA is managing, mediating, and
coordinating the behavior of the OAs assigned to
it, by resolving conflicts among them.
User Profile and Activity Profile Conflicts
When two or more users are in a same zone (or

room) at a same time, there might be differences
in their preference over the settings of the zone
(or room), which may result in a User Profile
Conflict in an intelligent environment. Similarly,
whereas a single user normally performs a single
Figure 4. Three System Profiles and Types of Conflict activity for a certain time period, two or more users

may perform different activities in a same zone coordinate the behavior of its OAs based on the
(or room) at a same time, which may lead to an data about users, user activities, and objects in its
Activity Profile Conflict. For instance, in an office, assigned zone. By doing so, it can handle conflicts
the lighting and sound preference for the resting between OAs (i.e., object-level conflicts).
activity of one user may be conflicted with the
preference for the working activity of another user. EMA (Environment Monitoring Agent): The
Theses types of conflict arise due to the goal primary role of the EMA is to control the behavior
difference between users and may not be easily of a whole environment by monitoring the context
predicted in advance. Although these conflicts may of each zone. The EMA can identify the context
simply be resolved by referring them to the users of each zone through the data received from an
involved, this may be burdensome for the users. associated BMA (i.e., current state of a zone).
Therefore, an effort to resolve user and activity Based on the context of the zone and its relation
conflict should be made by the agents of an to the contexts of other zones, the EMA
intelligent environment. coordinates the behavior of BMAs, when the
modification may further change setting(s) of other
Layered-Agent Structure zone(s). Thus it can deal with zone-level conflicts
(i.e., conflicts between BMAs). The EMA is the
In order to efficiently resolve possible conflicts top level agent that controls the overall behavior
and maintain the consistency of environment-wide of the environment.
modification, we introduce a layered-agent
structure. Briefly, the primary task of upper level In general, organizations are expected to improve
agents is to coordinate the behavior of lower level their performance if their organizational structure
agents and resolve conflicts arise between lower matches their task structure (i.e., structural
level agents. The layered structure includes only alignment) (Carley, 1999). In this respect, the
three levels of agents, arranged in a hierarchical layered-agent structure in the proposed model of
structure (Figure. 5). Conflicts are resolved by an intelligent environment is well aligned with the
referring them to the agent next level up. The structure of the task environment, which is
function of each level of agents can be described hierarchical in its composition (i.e., Building Level
as follow: – Zone Level – Object Level) (Figure 6).
OA (Object Agent): A set of OAs control the Conflict Resolution

behaviors of objects in a zone (or room) of the
environment. When a user forwards an action to Resolution of Object Profile Conflict
an object, the OA of the object identifies the action
and responds to it based on the behavior criteria To accommodate changes in user activities (e.g.,
of the object, which are described in its object switching from work to rest), objects must detect
profile, in conjunction with the user and user the change, retrieve the stored data from the
activity profile. The OA passes the data of its activity profile, and apply it in a timely fashion.
current state to its BMA, which is a higher level However, the user’s change of activity is normally
agent as well as an intermediate agent which only evident through one or more objects that are
summarizes the current state of the assigned zone associated with that activity. Other, non-associated
and uses it to coordinate the behavior of the lower objects, may not know what the current user
level OAs. OAs are the lowest level agents that activity is, nor that it is different from an earlier
directly manipulate the environment. activity, because each agent is physically and
rationally bounded and can only perceive some
BMA (Behavior Management Agent): Each BMA part of the environment.
controls the behavior of an assigned zone (e.g., a
room, a lobby, etc.) that includes a number of OAs. Thus, as discussed in the earlier section, in order
A BMA keeps track of the states of users, user to overcome such limitations of individual agents,
activities, and objects within a zone that is under communication channels between agents are
its coordination. A BMA summarizes users and required. Through communication channels with
user activities at a given point of time based on other object agents (OAs), an OA that is not
the state data received from the OAs associated directly associated with a user’s activity can still
with the current user activities. Thus, a BMA can be informed about it. For example, when a user

Figure 5. Agent Structure and Communication of Intelligent Environment
Figure 6. Structural Alignment of an Intelligent Environment
starts her/his break, the change of activity may be Resolution of an Object Profile Conflict can be
triggered by the chair on which s/he sits (leaning done by a Behavior Management Agent (BMA)that
back vs. sitting straight up). The chair could then determines the user’s current activity by
communicate this activity through communication summarizing the current state of the room based
channels to other agents, which will change their on information gathered from OAs (e.g., which
settings for the ‘rest’ activity of the user. But during objects are currently involved in which user
the break, the user may grasp a magazine that s/ activity, the user’s location in the room, etc.) and
he bought that morning and start reading it. At other resources (e.g., time of day, user’s previous
this point, the “table” (agent) on which s/he puts activity patterns, etc.). Thus, a BMA can be
the magazine will detect this activity and reason considered a coordinator who acts at a level above
that the user stopped her/his break and returned the object agents.
to work. It will inform other objects of this change
of activity, and they will change their settings from Resolution of User Profile and Activity Profile
the ‘rest’ profile to the ‘work’ profile. However, Conflict
the “chair” agent may still think that the user is
resting. This means that the activity detected by When there are two or more users at the same time
the table does not match the activity detected by in the same environment, two or more different
the chair, generating a conflict between OAs. user profiles are active simultaneously, which may

lead to a User Profile or Activity Profile Conflict, 1. The model is built with a hierarchical
because one user’s preferences for the office organization of multiple agents that generate
setting may be different from another user’s actions through message transmission.
preferences (e.g., difference of music style, 2. The agents have both individual goals and
lighting illumination level, temperature, etc.). This shared organizational goals. The individual goals
type of conflict may simply be resolved by are typically action-oriented (e.g., lighting agents
referring it to the users. However, this puts the - keeping illumination level according to user
burden of controlling the setting of an environment activity), whereas the organizational goals are
on the human actor, which is precisely what an more abstract (e.g., zone-level modification) and
intelligent environment is intended to avoid. can only be accomplished by the combined result
Hence, an attempt to resolve a user or activity of agents interaction through the communication
profile conflict should be made by the agents network.
themselves. When such a profile conflict is 3. Bi-directional interaction between layered
detected, a mechanism of conflict resolution is agents about a task environment and generating
invoked within the BMA at the zone level. In short, actions on the environment, augmented by a
the BMA may select one of the two profiles, coordination and conflict-resolving mechanisms,
combine the two, or create new profile as an completes the organization of the system.
alternative (Figure 7). As the result of this process,
a Group Profile is generated for a group of users. The propose approach also provides a much
needed overall organizational methodology to
Summary and Conclusions developing intelligent environments. It is easily
scalable, and is expected to be more responsive to
Current approaches to developing intelligent its occupants needs than mechanistic approaches.
environments are based on the systems approach,
where the entire environment is considered a
‘machine’ whose output (i.e., the environmental
setting) must be adjusted to conform to some pre-
defined schema. We contend that such a
mechanistic approach is inadequate, because it
fails to recognize the changing behaviors and
activities of the users, and fails to respond to
conflicting needs of multiple users. A more flexible
approach, based on design collaboration, is
proposed. It comprises of a hierarchical structure
of largely independent agents, which can
communicate with one another and resolve
conflicts much like a (human) team-type
organization does. The fundamentals of the
proposed model can be summarized as follows:
Figure 7. Conflict Resolution: User-Activity Profile Conflict

References Holmes, A., Duman, H., and Pounds-Cornish, A.

(2002). The idorm: Gateway to heterogeneous
Arens, E., Xu, T., Miura, K., Zhang, H., Fountain, networking environments. In Proc. International
M., and Bauman, F. (1997). A Study of Occupant ITEA Workshop on Virtual Home Environments.
Cooling by Personally Contolled Air Movement. Huberman, B. A. and Clearwater, S. (1995). A
Energy and Buildings, Vol. 27, No. 1, 45-59. multi-agent system for controlling building Schön,
A. D. (1983). The reflective practioner: How
Benne, B. (2004). Techno-organizational models professionals think in action. New York: Basic
to support construction industry work processes. Books, Inc.
Ph. D. Dissertation, University of California,
Berkeley. Intel Research (2005). http://www.intel-
research.net/berkeley/Seminars.ASP?Archive=1
Boman, M., Davidsson, P., Skarmeas, N., Clark,
K., and Gustavsson, R. (1998). Energy Saving and Jennings, N. R. (1999). Agent-oriented software
Added Customer Value in Intelligent Buildings. In engineering. In Proc. 12th Int. Conf. on Industrial
Proc. PAAM98 (Nawana and Ndumu eds.), pp. and Engineering Applications of AI, Cairo, Egypt.
505-517.
Kalay, Y. E. (2004). Architecture’s new media:
Booch, G. (1994). Object Oriented Analysis and Principles, theories, and methods of computer-
Design with Applications. Benjamin/Cummings aided design. Cambridge, MA: MIT Press.
Publishing Company, Inc.
Krogh, P. and Grønbæk, K. (2001). Roomware and
Boulding, K. E. (1962). Conflict and defense. New intelligent buildings. In Proc. Architectural
York: Harper. Research and Information Technology, Århus
School of Architecture.
Carley, K. M. and Gasser, L. (1999).
Computational Organization Theory in Multiagent Le Corbusier (1927). Towards a new architecture.
Systems. In A Modern Approach to Distributed New York: Praeger.
Artificial Intelligence (Weiss, G. eds.), Cambridge,
MA: MIT Press. March, J. S. and Simon, H. A. (1958).
Organizations. New York: Wiley.
Coen, M. H. (1998). A prototype intelligent
environment. In Proc. 1st International Workshop Mozer, M. C. (1998). The neural network house:
on Cooperative Buildings (CoBuild 98). An environment that adapts to its inhabitants. In
Proc. AAAI Spring Symposium on Intelligent
Colley, M., Clarke, G., Hagras, H., Callaghan, V., environments.
and Pounds-Cornish, A. (2001). Intelligent
inhabited environments: Cooperative robotics & Negroponte, N. (1975). Soft architecture
buildings. In Proc. ISR 2001, Seoul, Korea. machines. Cambridge, MA: MIT Press.
Corkill, D. D. L. and Susan, E. (1999). Diversity Newell, A. and Simon, H. A. (1972). Human
in agent organizations. Object Magazine, 8, 41- problem solving. Englewood Cliffs, N.J.: Prentice-
47. Hall.
Galbraith, J. R. (1977). Organization design. Pohl, J. and Myers, L. (1994). A Distributed

Reading, MA: Addison-Wesley. Cooperative Model for Architectural Design. In
Knowledge-Based Computer-Aided Architectural
Hobbs, R.W. (1996). Leadership through Design (Carrara, G. & Kalay, Y. E. eds.), Elsevier
collaboration. AIArchitect Vol.3, pp. 11. Science Publishers, Amsterdam, The Netherlands.

Rittel, H. W. J. and Webber, M. M. (1973).

Dilemmas in a general theory of planning. Policy
Sciences, 4.
Rowe, P. G. (1987). Design thinking. Cambridge,

MA: MIT Press.
Russell, S. and Norvig, P. (2003). Artificial

intelligence: A modern approach (2nd edition).
New Jersey.
Schön, A. D. (1983). The reflective practioner:

How professionals think in action. New York:
Basic Books, Inc.
Sharples, S., Callaghan,V., and Clarke, G. S.
(1999). A self-learning multi-agent architecture
for intelligent building monitoring and control.
International Sensor Review Journal(19:1).
Simon, H. A. (1984). The structure of ill-structured

problems. In Developments in design
methodology (Cross, N. eds.) (pp. 146-166): John
Wiley & Sons, Inc.
Streitz, N., Geißler, J., and Holmer, T. (1998).

Roomware for cooperative buildings: Integrated
design of architectural spaces and information
spaces. In Proc. CoBuild 98.
Weiser, M. (1993). Some computer science

problems in ubiquitous computing.
Communications of the ACM. Vol. 36, No. 7, 75-
84
Weiss, G. (1999). Multiagent systems: A modern

approach to distributed artificial intelligence.
Cambridge, MA: MIT Press.

M. Cabrinha / From Bézier to NURBS
From Bézier to NURBS:

Integrating Material and Digital Techniques
through a Plywood Shell
Mark Cabrinha1
1
Rensselaer Polytechnic Institute
Abstract
The development of digital fabrication has reintroduced material processes with digital processes. There has
been much discussion about the tool and the objects of the tool, but little discussion of the implication of the
material process on the digital process. A brief historical review on the development of computer numerical
control and the origins of the Bézier curve reveals an instrumental fact: computer numerical controlled tools
necessitated advancements in computational surfaces which eventually led to NURBS (Non-Uniform Rational
B-Splines) surfaces. In other words, the origins of NURBS surfaces resides in its relation to material processes,
rather than many current approaches that develop free form surfaces and then force the tool onto the material
without regard to the material properties.
From this historical and mathematical review, this project develops toward more intelligent construction
methods based on the integration of NURBS differential geometry paired with material qualities and processes.
Specifically, a digital technique of developing conceptual NURBS geometry into piecewise surface patches
are then flattened based on the material thickness and density. From these flattened patches, a material
technique is developed to intelligently remove material to allow the rigid flat material to re-develop into
physical surface patches. The goal of this research is to develop digital and material techniques toward
intelligent construction based on the correspondence between digitally driven surface and digitally driven
material processes. The application of this technique as a rational and flexible system is to support the
dynamic response of form and material toward such performative aspects as structure, daylight, ventilation,
and thermal properties.
“Each idea must be related to the principle of a material system, simple and primitive though it may look, on
which a variable solution could be based.”
Gerald Farin, Curves and Surfaces for Computer Aided Geometric Design

Introduction programs, which is more akin to digital clay, is an

accomplice in obscuring the tectonics of these
The recent focus of digital fabrication in artifacts. (Serriano188-9) Although there is no
architecture has reintroduced material processes doubt evidence to support these two theses, is this
with digital processes. Since at least the early necessarily so? First, even if one only
nineties, the accessibility of both computer conceptually understands the mathematical
software and hardware motivated a formal principles behind NURBS, learning the software
renaissance in architecture propelled by glossy is a question of preferable interface, not of encoded
renderings. This formal emphasis has only been bias or proprietary tricks. 1 Secondly,
compounded with the ease of formal manipulation understanding the origins of the material processes
enabled by non-uniform rational b-splines that inspired NURBS development implies an
(NURBS). With the recent introduction of obvious link to material systems and processes,
computer numerical control (CNC) in architecture, bringing into the digital process a question of
the ability to build this formal manipulation is no material making. The argument presented here is
longer novel. Moreover, the approach to formal that there is a great deal more information
development and material fabrication is met with embedded within the NURBS surface than merely
a certain schizophrenia: formal development up representing a shape, but indeed in directly
to a point, and then the form is handed over to be fabricating it.
sliced up, often in orthographic sections, to then
be milled. Although NURBS were introduced to Briefly three built examples are given which form
architecture before CNC technologies, the origins the parameters of this particular research project.
of NURBS are tied directly to a material system. Then, a brief overview of b-spline principles are
introduced that lead to both a working vocabulary
and certain techniques that are a more materially
This research project was motivated by the fact informed approach to working with NURBS.
that Pierre Bézier, arguably the forefather of Specifically, a digital technique of developing
NURBS, wrote a book on numerical control nearly conceptual NURBS geometry into piecewise
ten years before publishing his slightly more well surface patches are then flattened based on the
known book on the Unisurf CAD System. (Bézier) material thickness and density. From these
Further reading reveals a startling fact: Bézier and flattened patches, a material technique is
others at the time developed the foundation of what developed to intelligently remove material to allow
is now NURBS as a result of the development of the rigid flat material to re-develop into compound
computer numerical control in the 1960’s. The curved surface patches. This technique is
irony is that early CNC technologies inspired the demonstrated through a rigorous and replicable
foundation for NURBS - not the other way around. computational process, and fabricated as a full-
In other words, the tool necessitated more robust size installation. The goal of this research is to
techniques to exploit the tool. While CNC develop a technique toward intelligent
technology gives the possibility to slice up any construction based on the correspondence between
shape and fabricate it, is there not a more robust digitally driven surface and digitally driven
approach to integrating form and material through material processes. The application of this
the relation between NURBS and computer technique as a rational and flexible system is to
numerical control? support the dynamic response of form and material
toward such performative aspects as structure,
The motivation behind this question is two-fold. daylight, ventilation, and thermal properties.
First it is presumed, and all too frequently the case,
that the software leads the development of form Precedent
through its encoded emphasis with the designer’s
direction and intentions filtered through that Examples of digitally fabricated structures are now
medium. For example, in one case study “Form ubiquitous, at least in the architectural media.
Follows Software,” it is argued that each software Despite their formal eloquence, these projects
interface either enhances or hinders the generally rely on similar techniques of description
development or 3-d alternatives beyond the default and development to fabricate these forms. The
primitives. (Serriano 187) Furthermore, it is evolution of materially informed processes on
argued that the formal flexibility of NURBS-based digital development occurs at a much slower

timeframe than the speed of which formal flexibility of the aluminum section, and the gentler
representation can be published through images degree of curvature in one direction, each
and publications. For example, a close inspection aluminum lattice line could be CNC bent in one
of Frank Gehry’s work from the Guggenheim in direction and curved into the second direction as
Bilbao, to the Experience Music Project in Seattle, they are fixed in place. The skin in this pavilion
to the Sosnoff Theater at Bard College, and the is simply fabric stretched between the lattice
Disney Concert Hall in Los Angeles reveals a slow openings.
but significant impact of material processes and
digital development revealed in the actual
experience of the work through smaller details that
are often glossed over in publications. Although
the Disney Concert Hall technically predates the
Guggenheim project, the long delay and
consequent redesign and redevelopment of the
Disney Concert Hall reveals the development and
knowledge based experience of both Frank
Gehry’s office and the fabricators that his office
works with, with the most obvious example,
among many, being the move from a curved
limestone skin to a titanium skin. Three more
succinct examples that illustrate an evolution of
technique are Bernard Franken’s BMW pavilions Figure 1. Orthogonal Slicing of Form
“The Bubble,” “The Wave,” and “The
Brandscape” developed between 1999-2000.
These pavilions are simple programmatically and
illustrate an explicit focus on developing material/
digital processes through these projects.2
The Bubble’s main structure was developed

through orthogonal slicing yielding a rigid egg-
crate representation of the form. The precise form
is then enclosed with acrylic sheets thermoformed
from CNC milled polyurethane molds. Despite the
egg-crate structure, the pavilion’s novelty resides
in its use of compound curved surface patches as
opposed to geometrically reduced ruled surfaces.3
While these acrylic patches directly correspond
to the digital model, they are clearly not structural. Figure 2. Thermoformed Acrylic Patches
The following year, Franken’s pavilion “The

Wave” and similarly developed “Brandscape”
focused on evolving the primary structure away
from the egg-crate approach. Instead of
representing the form through the approximation
of orthogonal slicing, the evolution in these
projects is using the iso-parametric information
embedded into the digital model to develop the
doubly curved steel pipe structure. However, the
CNC pipe-bending machine only curved pipe in
one direction, and therefore the doubly curved
structural section was segmented into
approximately 100 singly curved pieces. The
secondary structure similarly consisted of a lattice Figure 3. Iso-Parametric Structure
of double curved aluminum tubes, but due to the

Although these projects are now five years old, Through the understanding of a couple principles
the evolution these projects present appears to behind b-spline geometry, a more intelligent
have met an impasse, as there has been little further correspondence between digital and material
development in this direction. Specifically, despite process is possible. Of principle concern here is
the Bubble’s egg-crate structure, the thermoformed first the idea of piecewise construction exemplified
surface patches are developed from a direct through the evolution of the Bézier curve into the
translation of the digital form to material form, b-spline. Second, exemplified through the
albeit with the less than desirable intermediary development of Coons surfaces, which is
mold. In the Wave, the primary structure develops mathematically speaking a patch, outlines the
directly from the iso-parametric information differences, and potential, of distinguishing
embedded in the surface. The limitation of this between a patch and a surface. The combination
project is the compounded cost in assembly of the of these two principles leads to a materially
singly curved steel segments to build the doubly informed technique of rebuilding conceptual
curved iso-paramteric lines of structure, yet an NURBS surfaces into piecewise bi-cubic patches.
opportunity is also revealed through the natural
characteristic of certain materials to bend in at least Although many mathematicians were working on
one direction, such as the aluminum tubes. the same problem, the French mathematician
Combining the assets of both of these projects, is Pierre Bézier was the first to publish the
there not a way to combine the embedded mathematical basis of what is now known as the
information of the iso-parametric lines with a Bézier curve. (Farin XV) Generally speaking,
structurally rigid surface patch to develop a true the Bézier curve is the basic building block of b-
free-form self-supporting structurally rigid shell? splines and NURBS surfaces. A Bézier curve is
comprised of a series of control points, wherein
From Bézier to NURBS: Principles of B-Spline the curve only passes through the first and last
Differential Geometry control points, or end points. The lines that
connect these control points are known as the
It is not the intention of this historical review to control polygon. Despite their free form, these
follow the derivations of the differential calculus curves are the graphical expression of the
of these mathematical curves, but to uncover the parametric equation: C(u)= ni=0 Ni,p(u)Pi , wherein
implications of this mathematical basis on material {Pi} are the control points, and the {N i,p(u)}
form. It cannot be overstated that these advanced represents the pth degree of the Bézier curve.(Piegl
mathematical techniques were motivated by the 81) In the simple Bézier form, the degree of the
burgeoning methods of computer numerical curve follows the number of line segments in the
control in the automotive industries at Citroen and control polygon, or one less than the number of
Renault in Paris, and Ford and GM in Detroit. control points. For example, the illustration shows
(Farin 363) a cubic Bézier curve (3rd degree), which is the most
Figure 4. Cubic Bézier Curve and B-Spline Curve as Piecewise Cubic Bézier Curves

common type in computer graphic applications. Lofting between only two curves results in a linear
Alternatively, a Bézier curve of degree one (linear) interpolation between them or a surface generated
is a line segment despite being derived from the by straight lines or rulings, known as a ruled
same parametric expression above. A b-spline, surface. Coons developed his method of a surface
however, can be described as a piecewise defined by four boundary curves through a
construction of Bézier curves where the technique of bi-linear blending. Each opposing
intermediary control points with the line passing set of curves are lofted with the resultant grid
through them are defined as knots. blended into a continuous surface. While the
solution of bi-lineal blending this ruled mesh is
From these simple elements, a surface can be elegant, it is also problematic as a result of the
generated by continuously moving either Bézier linear interpolation from one curve to the next.
or b-spline curves through space, connecting For variable surfaces which rely on more than four
similar curves in the opposing direction to the boundary curves, as most do, the Coons patch is
control points of the generating curve, indicated not informed by neighboring patches thereby
as u and v directions. Mathematically speaking, creating creases at the boundary conditions.
this is known as the tensor product approach. Mathematically speaking, “cross boundary
(Farin 271) In lieu of modifying a given curve tangents along one boundary depend on data not
over time, a given surface can be modified through pertaining to that boundary.”(Farin 369) Although
its control net, or lattice. These types of surface there are blending functions to smooth out these
manipulation are so familiar to architects working differences, they are frequently undesirable as they
with NURBS that their mathematical origins are alter the defining boundary curves.
typically unknown. It is not the particular math
that is significant but the rationality in building This brief knowledge of Coons patches is useful
up the system that is fundamental. Where the b- for two reasons. First, although patches may not
spline curve is a piecewise construction of Bézier be as advantageous in conceiving form, there
curves with the knot marking the beginning/ primary benefit is to rationally reconstruct a
endpoint of subsequent piecewise Bézier curves, surface from a given network of curves
these knots become iso-parametric lines in the maintaining the logic of that network. Secondly,
surface. Note that a true Bézier surface would with the knowledge that these patches are bended
therefore have no intermediary iso-parametric from two ruled surfaces, it is possible to degree
lines, which is defined as a patch. Whereas the reduce these patches in one direction resulting
Bézier curve was seen as the mathematical back into a ruled surface.
building block of the b-spline, the Coons patch
can be seen as the mathematical basis of b-spline Digital Technique: Conceptual NURBS
surfaces. (Farin XV) Geometry into Piecewise Bi-Cubic Patches
The mathematician S. Coons4 is attributed for his

derivation of a surface from four boundary curves. While the subject of this paper is not formal
Most architects are familiar with lofting, forming conception, it does imply the introduction of
a surface from a series of curves or line segments. material systems and processes at much earlier
Figure 5. Coons Patch Development

stages of formal development. To this end, the More significantly, the piecewise construction is
technique presented here relies on simple not merely a different representation of the same
conceptual geometry, yet the ease of which these NURBS surface, but the spacing between the iso-
two can be integrated suggests an increase in parametric b-splines (isoparms) can be correlated
integration between form and material at the to the particular material unit sizes. In this
earliest stages of formal development. The very example, the conceptual NURBS geometry was
limitation of the patch, that “cross boundary rebuilt based on a standard 4x8 sheet size of
tangents along one boundary depend on data not plywood. These dimensions can be verified
pertaining to that boundary,” becomes the through flattening a particular patch (see below)
opportunity for this technique of rebuilding the or simply verifying the appropriate spline length.
conceptual NURBS surfaces into piecewise bi-
cubic patches. If developed from an appropriately In conceptual development, a minimum of iso-
formed network of NURBS, the boundary curves parametric lines will likely be desired to maintain
of the patch can be developed as a seam between global control. However, once the form is set, it
adjacent patches where the patch boundary edges is suggested here that adjusting iso-parametric line
separated by that seam do share the same tangent. density is directly related to product scale and
The result is the same continuous curvature of the detail when the surface is connected to a material
conceptual NURBS surface but separated as process. For example, the isoparm density would
patches. be different for say a stapler, than for a building.
It should be made clear that for representational
purposes this is nearly irrelevant. However, when
tied to material parameters and processes these
isoparms exhibit a potential much greater than
simply representation. The isoparms can be seen
as seams in a material process whereby adjusting
the isoparm density creates a correlation with a
material dimension and/or a material process. In
this way, the process of slicing is not necessary,
as the object or building can be broken apart at
the seams. For example, in an injection-molded
product, I might choose to split the surface more
or less down the center into two equal surfaces.
However, for much larger products, such as
architecture composed of many large pieces,
Figure 6. Conceptual NURBS Geometry from Torus distinct techniques are necessary.
After the given conceptual NURBS surface is

rebuilt based precisely on particular material
parameters5, the surface curves can be extracted
and reconstructed as bi-cubic patches in piecewise
form6 . These are termed “bi-cubic” as the b-spline
curves in opposing direction are both of cubic
degree. Although this is a basic and
straightforward approach, when NURBS are only
seen as digital representations of form the material
implications of this technique would not be
apparent. Furthermore, other approaches such as
triangulation have dealt with the rough
approximation of continuous curvature, but are
developed from the reconstruction of the entire
surface, globally as a mesh, thus dropping the
parametric expression of the generating curves.
Figure 7. Rebuilt Materially Correlated Surface Triangulation also yields a field of complex joint
angles that must be dealt with. By rationally

constructing NURBS geometry, continuous doubly curved, as in the bi-cubic patches.

curvature can be achieved while simultaneously Although both options require specific software
simplifying the assembly of joints as a result of solutions, option one rebuilds the surface into a
their shared tangents, and creating a digital/ developable surface such that it can be unrolled
material synthesis through the correlation of regardless of material properties. The challenge
material parameters and process through defined of flattening compound curve material is that the
iso-parametric seams. What was the specific iso- material is actually stretching, creasing, and/or
parametric line of the NURBS surface, becomes tearing and therefore the density and thickness of
the defining edge of the patch. the particular material must be taken into
consideration by the software.
Option One: Degree Reduction of Bi-Cubic

Patches to Developable Surface
Although ultimately this project proceeds with the

second option, this option results from the review
of the mathematical principles of NURBS and will
likely be invaluable in different applications.
While by definition, the bi-cubic patches (curved
in both directions) are not developable; a degree
reduction in one direction will yield a ruled
surface, or technically a dual degree patch (cubic
in one direction and linear in the other direction).
Figure 8. Piecewise Bi-Cubic Patches
Furthermore, as a result of the piecewise
construction of bi-cubic patches, each boundary
While the technique of piecewise bi-cubic patches curve shares the same tangent with the adjacent
simplifies the joint, the challenge resides in patch, and therefore through reducing the degree
economically fabricating these bi-cubic in one direction this technique will yield a
(compound curve) patches. Of course these developable surface. Pause must be given here to
surfaces could be milled from a solid chunk of stress that degree reduction in one direction of any
material, though this is obviously time and material NURBS surface will not necessarily yield a
consuming. Similar to the acrylic patches in developable surface, but will yield a ruled surface.
Franken’s Bubble, intermediary molds could be However, in this technique, as the conceptual
milled and thin sheets of material could be NURBS geometry is rebuilt through a piecewise
laminated to the mold. 7 However, the most construction of patches, this approach to degree
economically and materially desirable solution is reduction should yield a developable surface at
to flatten these patches utilizing flat stock material each patch.
and basic two and three axis CNC technology as
these are the most common in the industry. Although it would be easy to proceed with this
(Kolarevich 34) project in developable surfaces by unrolling them
through Rhino8, this formal approximation is not
At this point in the process, there are two options satisfactory compared to the intended compound
in flattening the patches to work with flat stock curve model. Care needs to be taken in which
material. The first option is to degree reduce the direction to reduce in degree as it not only will
patch in one direction into a developable surface, effect the appearance of the surface, but depending
and the second option is to use specific software on the geometry can also split the seams.
that can flatten compound curve material based Furthermore, projects which successfully use ruled
on the material thickness and density. A surfaces approach ruled surfaces at the conceptual
developable surface is a sub-set of ruled surfaces stage of design as a material/geometric restraint,
whose rulings are parallel (a cylinder) or such as the physical paper strips in Frank Gehry’s
concentric (a conic section). As a result, design process, and not as an approximation of a
developable surfaces can be iso-metrically desired compound curve surface. (Killian 78)
mapped, or unrolled onto a plane. A compound Scale is also an issue, as larger projects utilizing
curve is simply a curve in both directions, or smaller ruled surfaces may appear as compound

curved as Gehry’s work exemplifies. The attempt Lamina. The illustrated examples on the Lamina
to develop this project through ruled surfaces website are of smaller sculptural objects and
suggests that the successful use of ruled surfaces furthermore are objects formed by only the
is more a part of conceptual development, surface. This is problematic as architecture
accepting the material constraint of curving in only obviously deals with much larger pieces and is
one direction. always a composite of multiple systems. For
Lamina to work, surfaces must be rebuilt as a mesh
thereby dropping the iso-parametric information.
However, through the piecewise technique of bi-
cubic patches presented here, the boundary
condition is all that is necessary so long as it is
proportionate to the material process employed
(e.g. in this example a 4x8 sheet of plywood with
a 4x8 CNC router). Although this may appear as
merely procedural, all too often I have encountered
an indifference to surface construction stemming
from the digital model as only a representational
model of form expecting the software to do it for
you. In this way, this is not a critique of this
particular software, but supports the need for the
Figure 9. Option One: Developable Surfaces
digital process to be informed by the material
process – an intelligence from the designer, not
the software.
Material Technique: Routing Uniform B-

Splines at Lines of Stress
For clarity, this paper has presented a linear process

of moving from digital to material, with the
knowledge of the material parameters, such as a
unit size, informing the digital process. However,
as both the digital process and material process
were being tested at the same time, it cannot be
stressed enough that the material development
proceeded in tandem with the digital development.
While a great deal of research was geared toward
Figure 10. Option Two: Bi-Cubic Surfaces the digital technique of intelligently rebuilding the
NURBS geometry, an equal amount of physical
Option Two: Flattening Compound Curvature research was based on the ability of the material
through a Software Approach to form a compound curve as well as simply testing
the file translation process from digital to material
The software approach to flattening compound processes. This project proceeds with plywood
curvature was previously the purview of advanced as the material of choice for a number of factors,
aerospace, automotive, and aeronautical software but particularly for the ease of use of working with
and therefore presented a daunting learning curve wood, its relative inexpensive cost, and most of
to mention nothing of expense. With the recent all that I had unlimited access to a woodshop with
introduction of Lamina Design software, this has a 3-axis CNC router. Certainly other materials
changed 9 . This stand alone, single function could be used and their specific properties would
software is both inexpensive and easy to use. modify this process.
Unique to this software, is its ability to flatten
compound curvature based on material density and Test Patch
thickness. This is intended neither as a product
endorsement nor software tutorial, but this A small test patch was fabricated as an initial
direction would not have been pursued without demonstration of this process. Curvature analysis

was used in Rhino to locate the patch with the Through this move toward lamination, an obvious
most extreme curvature. From this analysis, a option of using several thinner sheets would
section of this patch and frame was flattened with eliminate the need for the grooves. Although this
Lamina. The frame was simple to fabricate and is a viable option, it was avoided in the attempt to
assemble with dados at the corners to register the minimize the number of pieces and the more
pieces. For the 3/4” structural skin, a technique complex apparatus required to laminate multiple
was necessary to selectively remove material at sheets together. Moreover, as thinner material was
regular intervals relative to the double curvature more flexible, it also meant an increase in density
desired, which was a particular challenge as all of support structure for fabrication possibly even
four edges were of unique curvature due to necessitating a continuous mold, which was to be
unrolling the compound curve patch. Although it avoided in the first place. Finally, the thinner
added for a complex workflow, the solution was laminations resulted in nearly double the material
easy: use the blending function in Adobe cost as thinner sheets are not priced proportionally
Illustrator. This function rebuilds these lines as a to their thickness. Therefore, it was desirable to
blending of uniform b-splines, whereas the develop a technique mating two skins each with
differential geometry of NURBS is formed by the an exposed exterior face, unlike the laminated
non-uniform basis of isoparms. These paths were sample above with 1/8" masonite. It was decided
then milled using a 3/4" ball endmill to a depth of to laminate two skins of 1/2" wood considering
about 5/8" of a 3/4" piece of plywood. A ball particle board, plywood sheathing, and veneer
endmill was used instead of a square endmill to plywood.
avoid stress failure at the internal corners. This
patch was then easily screwed to the frame. This
quick sample identified three issues for further
development. First, the blending lines were more
dense than necessary in one direction, but more
importantly were not dense enough in the second
direction creating a noticeable segmentation at the
edges. Second, the material does not naturally
want to bend in both directions, and therefore
formed a flat spot at the center of the
approximately 1' by 2' sample. Third, although
the skin is inherently stronger as a result of the
compound curvature, because of the routed
grooves it easily supported about 30 pounds, the
weight of a child, but began to crack upon my full
weight.
The structural integrity of the skin was of utmost Figure 11. Curvature Analysis
concern and two options were considered. The
first was to infill the cavity with polyurethane foam
which expands creating an insulated and
potentially rigid structure.10 The second was to
laminate a thin sheet of material to the grooved
plywood such that the grooves could not flex. A
sheet of 1/8" masonite was milled to the same
flattened path profile, glued to the plywood and
then screwed to the frame to dry overnight. After
removing the screws, the laminated skin
maintained the exact shape of the frame, and the
new shell would support my full weight without
cracking and minimal deflection.This introduced
a more robust approach toward a true structural
skin independent of an intermediary frame, and Figure 12. Test Patch Exterior
therefore this option was pursued in this project.

Figure 13. Test Patch Interior Figure 15. Blend Pattern
Figure 14. Laminated Skin Figure 16. Sample Fabrication on Particle Board
Frame and Skin

based on the same blending lines developed from
If the frame was not to be used in the final Adobe Illustrator, only now the blend was made
installation, could it be avoided all together? more dense to create a pattern of grooves and
Theoretically, if there was a way to register the alternating holes. The ability to register the two
two offset shells through pins or dowels, the frame laminations through dowels proved indispensable
could be avoided. It must be made explicit that in laminating the skins, however it proved
this would only work if the two skins were offset impossible to wrestle the wood into compound
in the digital model and precisely flattened taking curvature without the frame. The frame then really
into account there thickness resulting in the becomes a stretcher and also a physical check that
slightest dimensional variation. Yet as a result of the skin is at precisely the correct curvature. If
the surfaces being offset from each other, they the previously described option one, of using ruled
share the same surface normals - that is, any line surfaces was used, this same technique of hole
drawn perpendicular to the surface would likewise alignment and peg registration could be used
be perpendicular to the offset surface. Therefore, without a supporting frame or mold, as the wood
when these sheets are flattened, a hole drilled has no problem curving in one direction, so long
perpendicular to the surface would then become as there is minute movement between dowel and
normal to the surface when the pieces take their hole, including material flexure.
shape. More significantly, if the holes are precisely
placed, the holes should align only when the In addition to the need for lamination, the test patch
surface takes its shape and could be doweled fixing revealed two other issues, one regarding visual
their position. Holes were proportionately located segmentation as a result of the grooves, and

second, a flattening of the material at its Modular Skin: Design for Assembly
unsupported center. Numerous material studies
were made on 1/2” 3 layer cabinet grade birch The full size frame was left assembled as skin and
veneer plywood and 1/2” particle board adjusting joint options were tested on it. Through these
groove spacing and depth, testing for flexibility. numerous material and lamination tests, greater
As plywood is not an isotropic material, the grain attention was placed on its surface quality creating
direction of the laminations has a significant a screen-like object at a scale between furniture
impact on its ability to bend, and therefore deeper and architecture. As a demonstration project, the
groove depths were tested perpendicular to the grooves became articulated at the surfaces in
strength of direction. While particle board is more conjunction with their structural necessity.
iso-tropic than plywood, it has almost no flexural Routing all the way through the material in one
strength and therefore segmented or slightly direction, the material took the curvature as well
cracked at the grooves. Exterior plywood as expressing its construction. Similar to the
sheathing was briefly tested, but is rarely perfectly dowels normal to the surface, the grooves only
flat and therefore difficult to precisely mill the aligned as the panels took their shape.
grooves. Furthermore, plywood sheathing has
numerous knots, plugs, and internal voids that As a full size installation/demonstration project,
resulted in failure. For the project developed here, an additional constraint was to create an object
it was decided to proceed with 1/2” cabinet grade that would easily be assembled, demounted and
veneer plywood with grooves spaced at maximum re-assembled elsewhere. This placed emphasis on
of 3” on center, and 3/8” depth, leaving 1/8" of a removable joint between the nine prefabricated
material, or about one of the three laminations. panels. Several options were considered,
Further structural analysis may suggest alternating including doweling, puzzle pieces, and a fingered
the grain direction between layers. However, as spline joint. In the end, the simplest joint, a 1/2”
plywood generally comes with the grain strength lap joint was used. However, due to the size and
in the long direction, to alternate the grain geometry of these panels, a great deal of stress is
directions the piecewise construction of bi-cubic placed on the joint, and therefore the simple 1/2”
patches would have to be based on a 4’x4' max lap joint only became feasible when a system of
module and not the 4’x8' max module used here, post-tensioning was used in conjunction with the
doubling the number of pieces and therefore lap joint. By including vertical ball end grooves
compounding the number of joints. 11 The back to back, an approximately 3/4” round void
flattening of the material at its unsupported center was created, making it possible to run a continuous
was simply overcome by adding an intermediary dowel in that void for additional support.
vertical b-spline support in the final frame. However, a wooden dowel created too much
friction, and was therefore impossible to remove.
Frame Final Assembly This introduced a new material to the project: high
density poly propylene (HDPE). HDPE is
The frame itself demonstrates the direct translation relatively inexpensive, is very flexible, has a very
of iso-parametric information to material structure. low co-efficient of friction, with a high tensile
While orthographic slicing yields a series of flat strength, and is easy to fabricate. These HDPE
sections, utilizing the iso-parametric lines yields rods were tapped at the ends and used as tenons to
ruled surfaces. These lines were offset 4” for post-tension the panels which aided greatly in the
structural depth, and digitally modeled as a assembly of the structure.
modular series of nine frame patches. These
frames were then unrolled, and tested in a laser Skin Final Fabrication
cut scaled physical model. This physical model
indicated a sequence of construction, as well as The final fabrication of this project was no small
the need for curved corner bucks to give the affair, with the frame fabricated over spring break,
boundary of the frames their correct curvature, and and the laminated skins fabricated at the end of
was instrumental for assembling the full size the academic year when the wood shop was more
frame, suggesting that the model is not merely for or less vacant. Laminating each patch over the
representation, but a prototype of actual frame was a blend between choreography and
construction. The precise location of each chaos. Each patch required at least ten clamps
plywood spline was marked with dados. that were sequentially tightened to distribute the

Figure 17. Full-Size Frame Installation
Figure 18. Tenons in Assembly Figure 19. Post-Tensioning
stress as uniformly as possible. However, despite Conclusion

the rigorous digital process and numerous smaller
test pieces, the full size patches built-up enough The motivation of this research project is based
surface tension along the larger surface that each on the relation of the material process on the digital
piece buckled at one fairly consistent location. process. In particular, through the mathematical
Despite the frustration of this physical failure, and historical review of NURBS surfaces, it is
pushing the material to its limits is also indicative found that there is a strong correlation between
of the need for further material research in tandem NURBS and material processes moving
with the digital techniques presented here. One architecture beyond representations of form
of the three bays was able to be assembled as a toward physical processes informing form and
demonstration of the potential of this installation. enabling the design process. The emphasis of this
paper has been on the digital technique of
developing conceptual NURBS geometry into
piecewise bi-cubic patches informed by material

parameters and processes into a materially related

digital model. From this, two options are
presented: one in working with ruled surfaces and
one in working with compound surfaces. A
material technique of routing uniform b-spline
grooves was used to develop the directional
material properties toward compound curvature.
As a result of the material failures in final

fabrication, one of the conclusions of this research
is that working with ruled surfaces and compound
curved surfaces is more than simply a question of
degree reduction, mathematically speaking, but a
fundamental question of material resistance – a
parameter not included in the “material
parameters” of the digital model presented here.
While the digital techniques presented here are
an important development of integrating digital
and material techniques, the material techniques
require future work with emphasis placed on the
Figure 22. Final Assembly
Figure 20. Final Fabrication
Figure 21. Clamping Laminated Skin Figure 23. Detail

material resistance inherent in the material as a 9. see www.laminadesign.com.

result of the curvature. Material research needs 10. This remains a viable option for future work
to focus on material composition: iso-tropic vs. as well as product application, but this project
non-iso-tropic materials, material density, and moved toward the more demanding application
thickness of laminations in compound curvature. of a structural skin.
In addition to wood and metal, this would include 11. In hindsight, the largest pieces that approached
new materials such as the HDPE introduced in this 8’ were also the most prone to failure both from
project. For prefabricated assemblies, structural their own weight, as well as surface tension
testing is necessary through computational developed over the longer length of the piece, and
processes such as finite element analysis (FEA) therefore, alternating laminations on a 4’x4’ max
as well as physical destructive testing on entire module may be more successful.
assemblies including the joints between
assemblies. Building applications in polyurethane References
injected compound curved stressed skin panels as
well as compound curved concrete formwork are Bechtold, Martin et al, ed. (2000) New
two clear structural applications utilizing the Technologies in Achitecture: Digital Design and
techniques presented here. Manufacturing Techniques. Cambridge, MA:
Harvard Design Books.
Notes
Bechtold, Martin. (2004) Digital Design and
1. The ease of interface is the question of bias, as Fabrication of Surface Structures. P. Beesley (Ed.)
Serriano argues, to say nothing that NURBS could ACADIA 2004: Fabrication: Examining the
be said to be biased from the first. Digital Practice of Architecture. (pp 88-99)
2. To the degree that “Form Follows Software,” Toronto, Canada.
at issue here is the development from form to
fabrication and not the formal morphogenesis of Bézier, Pierre. (1972) Numerical Control:
these projects. See New Technologies in Mathematics and Applications. New York, NY:
Architecture and Digital Design and Wiley.
Manufacturing, for the morphogenesis of these
projects. Farin, Gerald. (1993) Curves and Surfaces for
3. Frank Gehry’s Conde Nast project uses a similar Computer Aided Geometric Design, 3rd ed. San
approach with its glass surfaces while using ruled Diego, California: Academic Press.
surfaces for the titanium interior. See New
Technologies in Architecture and Digital Design Killian, Axel. (2003) Fabrication of Partially
and Manufacturing for more on this project as Double-Curved Surfaces out of Flat Sheet Material
well. Through a 3D Puzzle Approach. K. Klinger (Ed.)
4. Ironically, I was first introduced to Coons ACADIA 2003: Connecting-Crosscroads of
surfaces through MasterCAM. Digital Discourse (pp 75-83) Indianapolis,
5. As a result of the scale of architecture, this Indiana.
typcially means an increase in isoparm density, if
a reduction of isoparm density is required, this will Kolarevic, Branko ed. (2003) Architecture in the
most likely alter the surface. Digital Age: Design and Manufacturing. London,
6. Alternatively, it may be possible to split the England: Spon Press.
surface at the isoparms as in Rhino. However,
the accuracy here is questionable. Trimming the Piegl, Les & Tiller, Wayne. (1995). The NURBS
surface is not an option, as trimmed surfaces Book. Berlin, Germany: Springer-Verlag.
merely mask the entire mathematical surface from
view. Schodek, Daniel et al., eds. (2005) Digital Design
7. See also Bechtold, Martin. “Surface Structures: and Manufacturing. New York, NY: Wiley.
digital design and fabrication” in Fabrication:
Examining the Digital Practice of Architecture. Serriano, Pierluigi. (2003) Form Follows
8. see www.rhinocerous3d.com, Rhino is well Software. K. Klinger (Ed.) ACADIA 2003:
known for its unique “unroll” command to unroll Connecting-Crosscroads of Digital Discourse (pp
developable surfaces. 185-205) Indianapolis, Indiana.

D. Barker, A. Dong / A Representation Language for a Prototype CAD Tool for Intelligent Rooms
A Representation Language for a Prototype CAD Tool

for Intelligent Rooms
Daniel Barker1, Andy Dong2
1
University of Sydney
2
University of Sydney
Abstract
Intelligent rooms are a type of intelligent environment which enhance ordinary activities within the confines
of a room by responding to human interaction using pervasive and ubiquitous computing. In the design of
intelligent rooms, the specification of how the intelligent room enacts intelligent behavior through computational
means is as integral as the geometric description. The self-aware and context-aware capabilities of intelligent
rooms extend the requirements for computer-aided design tools beyond 3D modeling of objects. This article
presents a Hardware as Agents Description Language for Intelligent Rooms (HADLIR) to model hardware in
an intelligent room as “hardware agents” having sensor and/or effector modalities with rules and goals. End-
users describe intelligent room hardware as agents based on the HADLIR representation and write agent rules
and goals in Jess for each hardware component. This HADLIR agent description and the requisite software
sensors/effectors constitute “hardware agents” which are instantiated into a multi-agent society software
environment. The society is then bridged to either a virtual environment to prototype the intelligent room or
to microelectronic controllers to implement a physical intelligent room. The integration illustrates how the
HADLIR representation assists in the design, simulation and implementation of an intelligent room and provides
a foundation technology for CAD tools for the creation of intelligent rooms.

Introduction room is designed and constructed by a pervasive

computing expert as a custom integrated hardware
The publication of Norbert Wiener’s Cybernetics and software system in situ.
in 1948 on the reciprocal nature of man-machine
symbiosis and the thesis of a (mathematically) A CAD package which contains pre-modeled
functional relation between human brains and drafting elements such as cameras and infrared
machines influenced architectural theory at least sensors is probably insufficient for an intelligent
since Gordon Pask’s (Pask, 1969) publication of room (architectural) designer. The computational
The architectural relevance of cybernetics. The elements drawn into the background of intelligent
notion that built environments and artificial objects rooms suggest that the type of computer-aided
within that environment should perform as design tool necessary to design intelligent rooms
embodied extensions of human minds and bodies is more than the 3D modeling and drafting
offered tantalizing potentials to conceptualize new currently available in software packages such as
types of experiences within those spaces. While AutoCAD. While a complete discussion of the
the range of actual experiences humans could requirements for a CAD system for intelligent
conceive is possibly too multifarious to be rooms is beyond the scope of this article, we
designed a priori, the proposition has been that believe that such a CAD system should enable the
those experiences can be designed for and designer to model and simulate the following three
manifested during operation of these "intelligent" aspects of an intelligent room, the geometric
built environments. However, architectural modeling aspects notwithstanding:
designers needed to wait until the technologies Sensing:
caught up with their visions for intelligent • to detect human activities including
architecture. speech, touch, presence, and movement
• to be self-aware within a finite bounded
Today, intelligent built environments appear in space
places ranging from research labs such as the MIT Thinking:
835 Intelligent Workspaces project to residential
homes such as The Gates Estate. One type of
• to behave as an intelligent system with a
intelligent environment, the intelligent room, is knowledge representation
described by Hirsh (Hirsh, Coen, Mozer, Hasha, • to coordinate sensors and effectors
& Flanagan, 1999) and Coen (Coen, 1998) as an Effecting:
integration of sensor and effector modalities to • to express audio and visual effects
enhance ordinary activities within the physical
boundary of a room. Intelligent rooms incorporate Thus, a CAD tool for intelligent room should
hardware for sensing and effecting such as satisfy at minimum the following requirements:
wireless sensors (e.g., "motes"), overhead
projectors, array microphones, smart whiteboards 1. A means to model arbitrary
and virtual reality equipment. Applications of computational (e.g., pervasive computing
intelligent rooms include intelligent conference components) hardware in the intelligent
rooms (Coen, 1998), automated homes (Intille, room
2002) and kids’ play rooms (Bobick et al., 2000). 2. A means to specify, model and simulate
the function and behavior of the
One hurdle in the design of intelligent rooms is computational hardware in a virtual
the lack of computer-aided design (CAD) and environment including interoperation
simulation tools, particularly those accessible to 3. A means to implement the software
non-computer scientists and electrical engineers. functionality of the intelligent room in a
Physical rapid-prototyping components such as physical environment
Smart-Its (Gellersen, Kortuem, Schmidt, & Beigl, 4. A means to model the geometry (i.e., 3D
2004) simplify the physical implementation of model) of the intelligent room and
computational objects into intelligent rooms, but associated hardware
do not offer a means to develop and test a virtual
prototype of an entire intelligent room plus its A 3D physical (structural) model of an intelligent
associated computational objects prior to room (Requirement 4) could be readily modeled
implementation. For the most part, an intelligent in a persistent, virtual environment such as Active

Worlds [www.activeworlds.com] or Second Life (plan) that may achieve a given goal or solve a
[www.secondlife.com]. However, these 3D virtual problem. An intelligent room must also review the
environments contain no native capability to results of the plan and compare those results to
specify and model the function and behavior of the desired goal. If the plan failed to achieve the
pervasive computing elements in the intelligent goal, alternative plans must be formulated and may
room or to implement the intelligent behavior of involve a request to other systems. As Hunt (Hunt,
the intelligent room in an actual physical 1998) argued, intelligent buildings should do more
environment. than sense and control the environment; they
should also "anticipate the behavior of its users
To address these deficiencies, and partially satisfy through knowledge based on experience."
the first and second requirements of an intelligent Believing that intelligent architecture consists of
room CAD tool, we developed the HADLIR more than the built environment, Gage (Gage,
representation. Our eventual aim is to create a set 1998) suggested deploying intelligent objects
of CAD tools which would allow a designer to throughout these spaces, objects which
specify the function, structure and behavior of the communicate and converse with each other and
computing hardware constituting an intelligent the occupants. Because the HADLIR
room and the intelligent, goal-oriented behavior representation can model these intelligent objects
of the room itself. Another research project at the as software agents, designers may prototype and
Key Centre of Design Computing and Cognition simulate in a virtual environment the technological
is investigating the design of an intelligent room means by which these objects act together with
as a type of curious agent, that is, an affective agent occupants before implementation.
whose behavior is partially determined by the
cognitive-behavioral emotion of curiosity. In This paper focuses on the development of the
support of the curious agents for intelligent rooms HADLIR representation as a part of foundational
research project, it is first necessary to create a technology for intelligent room CAD tools. The
CAD tool to specify, model and simulate the next section details the HADLIR representation
behavior of an intelligent room. The CAD tool and an interactive tool for describing hardware in
should make possible the specification, modeling intelligent rooms based on the HADLIR
and simulation of the behavior of the intelligent representation. Thereafter, a demonstration of the
room with the pervasive computing hardware HADLIR representation in modeling and
operational. Likewise, the specification and simulating a pressure mat in a virtual and physical
modeling should permit the designer to program intelligent room is provided. The paper concludes
intelligent behavior, that is, the "thinking" aspect with a discussion on the current limitations of the
of the hardware. For this research project, these system and future work.
objectives are addressed in part by modeling
hardware in intelligent rooms as (software) agents The HADLIR Representation
having sensor and/or effector modalities with a
descriptive language called the Hardware as Background
Agents Description Language for Intelligent
Rooms (HADLIR). Recently, ontologies have been implemented in
various intelligent rooms and pervasive computing
Deploying software tools for the design of environments as a potential solution to describe
intelligent rooms as comprised of sensing and the computational hardware constituting an
effecting hardware modeled as agents follows in intelligent room and the intelligent goal-oriented
the ideas by architects before us who proposed behavior of the intelligent room. Recent
agents as a way to implement "intelligent implementations include the ontology and
architecture." For example, Krueger (Kreuger, Semantic Web technology integrated into a
1996) proposed a biological model for intelligent CORBA-based infrastructure of an intelligent
and interactive architecture implemented as a goal- environment named GAIA (Ranganathan,
directed agent (Kreuger, 1998). Goal-oriented McGrath, Campbell, & Mickunas, 2004) and the
behavior is an integral characteristic of the COBRA-ONT ontology developed for context
HADLIR representation because intelligent aware intelligent environments (Chen, Finin, &
systems must call upon knowledge (i.e., operate Joshi, 2004). The GAIA intelligent environment
at a knowledge level) to formulate a set of actions primarily addresses the hardware configuration

problems; the GAIA technologies incorporate intelligent room as (software) agents ("hardware
ontologies to augment system-level services such as agent") having sensor and/or effector modalities
as configuration management and interoperation with rules and goals written in Jess. Many
of computational components. The COBRA-ONT hardware devices can be configured for intelligent
ontology, on the other hand, deals with the rooms; some examples include Web cameras,
description of the state and behavior of objects in pressure mats, lights, speakers, microphones, and
intelligent environments. The COBRA-ONT display panels. Regardless, each type of device in
ontology describes four distinctively related its interaction with the intelligent room
categories that include: ontologies about physical environment and occupants could be classified as
places, ontologies about agents (both human and either a sensor, an effector, or a sensor and an
software agents), ontologies about the location effector. Figure 1 describes the sensor/effector
context of the agents, and ontologies about the modality for some typical hardware devices found
activity context of the agents. While formal in intelligent rooms. Consequently, the function
ontology systems such as these offer the highest of the device as a sensor or an effector could be
levels of specificity and interoperability, they also characterized by one or more behaviors where
require the most up-front work and "ontological each behavior would include associated actions.
commitments" from the users and are rather
inflexible to subsequent changes.
For this research, we do not attempt to create a

complete ontology for intelligent environments for
the reasons pointed out by Chen (Chen et al., 2004)
and also due to the overhead in ontology
development, the federation of ontologies,
ontological commitments that make knowledge
sharing possible, and the requirement to update
the ontology as new hardware devices are added.
Instead, we propose to describe pervasive
computing hardware in intelligent rooms from the
standpoint of (software) agents.
Figure 1. Example of hardware devices as sensors, effectors
Representation of Hardware as Agents in HADLIR and sensors and effectors
Multi-agent systems appear to be the dominant We define the term "HADLIR agent description"
approach for creating intelligent rooms. Multi- as the hardware as agent description plus the rule
agent systems such as the ScatterBrain system base. The components of a "hardware agent" are
(Coen, 1997) are already found in operational the HADLIR agent description and the agent
implementations of intelligent rooms. Because software sensors and effectors written in a
agents are simple entities, agent based systems programming language such as Java. Software
relinquish the need for complex centralized control sensors and/or effectors allow the hardware agent
(Coen, 1998). As such, the key philosophy of the to interact with the agent environment. As
HADLIR representation is to minimize the illustrated in Figure 2, the HADLIR agent
descriptive overhead for each component of description is combined with requisite
hardware in the intelligent room to what is Environment Descriptors and a Communication
sufficient to characterize the piece of hardware as Bridge in order for the agent environment to
an agent. communicate with the intelligent room
environment.
Because of the predominance of the agent model
in intelligent room software architectures, the idea Because all practical intelligent environments
behind the HADLIR representation is to model would consist of multiple components of
hardware devices as (software) agents having hardware, once each piece of hardware has been
sensor and/or effector modalities with rules and represented as a hardware agent, the hardware
goals. The HADLIR representation describes the agent is then added into a device library and
pervasive computing hardware constituting an instantiated into a multi-agent society (MAS).

Within the MAS, the hardware agent may maintain We take an entity-relation approach. For clarity
awareness of itself and of other hardware agents in the following discussion, we provide the entity-
and the environment (Huhns & Seshadri, 2000). relation diagram in Figure 4 to illustrate the
For integration with a CAD tool, the MAS is relationships and organizational structure of the
connected to either a virtual environment for HADLIR representation. The behavior-type and
simulation or a physical room for implementation action child elements are used to describe the
as shown in Figure 3. actions a behavior performs. The behavior-type
child element type describes the form of data, for
example, force, distance, and time, that the
The hardware as agent description in the HADLIR
hardware agent senses and/or effects. The action
representation is comprised of a taxonomic
child element name is used in conjunction with
framework derived from an entity-relation model
behavior-type child elements name and type to
of the potential hardware devices to their
describe the specific sensors and/or effectors the
behavior(s) and action(s). The framework
hardware agent needs to be configured with. For
classifies behaviors and actions associated with
example, a pressure mat would have a behavior to
the hardware in a taxonomic hierarchy. The
sense if an object were present on the pressure
framework includes a set of representational terms
mat. For this behavior, the behavior-type name is
to describe a specific behavior and action.
sensor, the behavior-type type is force and the
Ranganathan (Ranganathan et al., 2004)
action name is measure. To a certain extent,
implements a SVO (Subject–Verb–Object) format
because the HADLIR agent descriptions adhere
in their ontological framework. Consequently the
to a taxonomic hierarchy and set of
structure of their predicates is Context Type
representational terms, one could classify the
(<Subject>, <Verb>, <Object>). For instance, the
HADLIR representation as an ontology (Gruber,
ontology declares that the location predicate must
1993). According to this definition of an ontology,
have a subject which belongs to the set of persons
the HADLIR representation is an ontology of
or things, a verb or preposition like ‘‘inside’’ or
hardware as agents in intelligent rooms rather than
‘‘entering’’ and a location, which may be a room
an ontology for intelligent rooms. The latter is the
or a building.
aim of the GAIA and COBRA-ONT ontologies.
Figure 2. Dynamic creation of a Hardware Agent

Figure 3. Agent Architecture of a CAD Tool for Intelligent Rooms
Figure 4. Entity relationships for hardware devices as agent descriptions

Furthermore, the HADLIR representation is to create an operational multi-agent society. This

intended to describe situated agents (Smith & multi-agent society enables the agents to
Gero, 2004). Situated agents are agents built using communicate amongst each other and with the
concepts from situated cognition. They achieve environment. The rules control the agent
goals through situated interaction rather than reasoning. Facts are obtained from the
algorithmic planning. The situated agent reasoning environment using sensors. Rules use the facts
model consists of five processes: obtained to determine the state of the environment.
Jess rules have the following structure:
1. sensation: senses the environment
2. interpretation: uses sensor data and (LHS (IF
expectations to interpret what the agent => =>
believes the environment to be RHS) THEN)
3. hypothesizer: observes interpretations of
the environment, and asserts goals based The LHS of the rule contains the conditions that
on the agent’s view of itself in the need to be satisfied to fire the RHS of the rule.
environment The RHS of the rule contains the action of the
4. action activation: reasons about the steps rule. Hence the Jess rules are structured much like
to achieve a goal and trigger the IF THEN statements. The rule base has MAIN,
effector(s) to make changes to the STRUCTURE, BEHAVIOR, and FUNCTION
environment modules for facts. These four categories describe
5. effection: effects (alters) the environment a priori facts about the hardware agent. The rule
base has INTERPRETATION, HYPOTHESIZER
Situated agents move away from traditional agent and ACTION modules for rules. The
communication languages such as CORBA INTERPRETATION module uses sense-data and
because they are constructive and interactive. expectations to interpret what the agent believes
Moreover situated agents are capable of reflexive, the environment to be. The HYPOTHESIZER
reactive and reflective action. Situated agents module observes interpretations of the
facilitate three types of actions: environment, and asserts goals based on the
agent’s view of itself in the environment. The
1. Reflexive action: directly initiated from ACTION module reasons about the steps to
sensor data achieve a goal and triggers the effector(s) to make
2. Reactive action: initiated by changes to the environment.
interpretations of the agent, no explicit
reasoning with goals and expectations Figure 5 describes the basic components of the
occurs situated agent architecture underlying the
3. Reflective action: hypothesizer explicitly HADLIR representation. One principal
reasons about expectations and requirement of the agent architecture is for the
alternative goals. HADLIR representation to have the capability to
configure an agent as a set of sensors/effectors and
Jess Rule Base a rule base. The agents in the agent package
(Smith, 2003) are reteagents. The reteagent is an
Situated hardware agents instantiated from the implementation of an agent in Jess excluding the
HADLIR representation also require rule-bases to sensors and effectors (Smith, 2003). The sensors
be able to sense, interpret, hypothesize, take action, and effectors are instead implemented in Java.
and effect the intelligent room. The second part Sensors are Java objects that sense an agent’s
of the HADLIR representation is the rule-base. environment and effectors act on that environment.
The implication of situated agents on the HADLIR Sensor data is stored in the Jess working memory.
representation is that each device must also have The Java-based sensors send their messages to
a rule-base to encode the device’s rules and goals. Jess, which stores the messages in its Working
Jess (Friedman-Hill, 2003) is a rule-based Memory. Messages from other agents are also
language extension to Java that supports the recorded in the Jess working memory by the
creation of software agents capable of reflexive, sensors. That is, each time new sensor data is
reactive and reflective action. We integrated Jess received, a new fact (actually, a Java Bean) is
with an agent environment package (Smith, 2003) asserted into Jess working memory. The Rete

Engine implements its goals through the Java-

based sensors. Both the sensors and effectors are
realized in the agent environment.
Figure 5. Architecture of ReteAgent
HADLIR Authoring Tool
The designer creates HADLIR agent descriptions

for all of the intelligent room hardware as agents
having sensor and/or effector modalities using the
HADLIR authoring tool. The HADLIR authoring
tool produces XML. Using XML permits the Figure 6. HADLIR Generator
semantics and structure of the HADLIR hardware
as agent description to be enforced with XML
Schema. The XML Schema ensures that:
• The element relationships in the device
descriptions are maintained
• All necessary elements are present in
each hardware as agents description
• The field content of elements is
appropriate.
• The elements and attributes are in the
correct order.
The JavaScript-based HADLIR authoring tool

shown in Figure 6 applies dynamic tables, inputs,
select fields and buttons to generate a HADLIR
hardware as agent description. Any number of
behaviors can be added to the hardware as agent
description. The HADLIR authoring tool produces
an XML file similar to the one shown in Figure 7
which was produced for the pressure mat example
described later in the paper. Once generated, the
hardware as agent description is combined with
the rule base and inserted into the collection of Figure 7. XML Description for a Pressure Mat
HADLIR agent descriptions for intelligent room
hardware devices and organized into a device
library.

Demonstration of the HADLIR Representation • Tell the pressure mat object to deactivate
when there is not an object on the
Using HADLIR to Describe a Pressure Mat pressure mat.
We will illustrate the implementation of the
• If the pressure mat has been told to
deactivate and we don’t currently have a
HADLIR representation with a simple sensor, a
goal to deactivate the pressure mat, then
pressure mat. A pressure mat is a pressure sensor
assert one.
which has no dependencies, that is, does not rely
on other hardware devices to operate. As will be
The pressure mat ACTION module has rules that
described in the Implementation in a Physical
do the following:
Environment section, the pressure mats provide
an array of sensors for the production of generative
digital media. Occupants walk on top of the • Satisfy the goal to activate the pressure
pressure mats (hidden underneath the carpeting), mat by changing the pressure mat state
thereby producing a sensation. The speed and in the world such that it is active.
direction of their movements affects the L-system • Satisfy the activate pressure mat goal
that generates the digital media. To the occupants, when the pressure mat is already active
the intelligent room appears to react to their by doing nothing.
movements without the occupants explicitly • Satisfy the goal to deactivate the pressure
"interacting" with the room. mat by changing the pressure mat state
in the world such that it is not active.
The prototype CAD tool for the demonstration
consists of the following components: An effector, such as a video camera, would have
similar modules except that the HYPOTHESIZER
1. The HADLIR representation and module’s rules would request activation/
authoring tool deactivation upon request rather than an
2. A multi-agent society software package environmental stimulus. Then, the HADLIR agent
(Smith, 2003) description is combined with appropriate
3. A virtual environment modeling system Environment Descriptors and transformed into an
(Active Worlds) agent definition file specific to our multi-agent
4. A microelectronics control system (Teleo society software package using XSLT rules. The
and Max) agent definition file contains the properties
necessary to instantiate the agent into the multi-
First, the designer creates the HADLIR hardware agent society software package. The XSLT rules
as agent description using the authoring tool. Next, include two elements: the specification of the
the designer creates a HADLIR agent description software sensor(s)/effector(s) associated with the
by combining the previous HADLIR hardware as HADLIR agent description and the Environment
agent description with an appropriate Jess rule Descriptors. For the pressure mat, the software
base. The Jess rule set is dependent only on the sensor (the hardware agent’s sensor) detects
hardware, that is, it is constant across the virtual location and pressure. The Environment
and physical intelligent room environment. The Descriptors express how the agent is to be
pressure mat INTERPRETATION module has presented in the (virtual or physical) intelligent
rules that do the following: room environment. For example, in the virtual
environment, the Environment Descriptors would
• Locate the pressure mat in the prescribe the appropriate 3D model of a device in
environment and load the rule modules. a VRML file. In the physical environment, the
• Collect local sensor data. Environment Descriptors would prescribe the
required device drivers and the communication
The pressure mat HYPOTHESIZER module has port on the microelectronics board.
rules that do the following:
Virtual Environment Simulation
• Tell the pressure mat object to activate
In the virtual environment simulation, the XSLT
whenever an object is on the pressure transformations additionally produce an object
mat.

definition for the pressure mat in Active Worlds. effectors that would be suitable for a range of
The object definition is read by Active Worlds to hardware devices.
instantiate the pressure mat as a 3D object with
avatar capabilities in Active Worlds. Once For purposes of illustration, a wall object was used
instantiated into Active Worlds, the designer can as a 3D representation of a pressure mat. The
move and modify the geometric properties of the activation of the mat is visualized in the virtual
3D object as in a standard 3D modeling package. environment through a color change. As illustrated
The hardware agent running in the agent in Figure 8, when an avatar (virtual representation
environment communicates with the pressure mat of a person) walks towards the wall object
in Active Worlds via a Communication Bridge. representing the pressure mat, and the avatar is
The Communication Bridge is essentially a "hook" inside the proximity of the object, the activation
into the event notification engine in Active Worlds. of the pressure mat is signaled by a color change.
The hardware agent’s software-based sensors,
written in Java, pass and receive messages from Implementation in a Physical Environment
the Communication Bridge as a Java Bean to allow
the hardware agent to sense and interact with the The primary difference between the virtual
Active Worlds virtual environment. Active Worlds environment simulation and the physical
sends events as messages to the hardware agent environment is the requirement of additional
through the Communication Bridge; the hardware and software to process the electronic
Communication Bridge parses these messages for data signals from the pressure mat. In other words,
relevant messages, that is, messages of interest to the hardware agent continues to operate within the
the software sensors. In response, the agent sends multi-agent software package as before with the
actions through the Communication Bridge to same HADLIR agent description. However, the
Active Worlds. Unfortunately, the physics of Communication Bridge is configured to send
pressure is not defined in the Active Worlds. location and object presence messages to and from
Consequently proximity was used to model the physical environment. For the physical
pressure. Simulations of other hardware devices environment, we instrumented Teleo
in the virtual environment would require the use [www.makingthings.com] microelectronic control
of different software-based sensors and/or modules and Max [www.cycling74.com] to
effectors. We are currently working towards the process data signals from the pressure mats as
development of a generic class of sensors and schematically illustrated in Figure 9.
Figure 8. Simulation of the pressure mat hardware agent in the KCDCC AW Agent Package

Figure 9. Version that uses a physical environment sensor Figure 11. Teleo Module
to produce an action
Figure 12. Max application for a pressure mat at loc. D6

Figure 10. Pressure Mats in the Sentient
The HADLIR representation was implemented for The t.intro.din retrieves the data signal from the
a physical intelligent room called The Sentient. pressure mat and sends it as a message into the
The Sentient is an intelligent room project being toggle. The toggle sends the binary number as a
developed in the Faculty of Architecture at the message to the number object, which displays the
University of Sydney. The Sentient provides a data for debugging purposes. The data is then sent
physical test bed for the HADLIR and the from the number object to the mxj myClient object.
HADLIR-based CAD tool. Underneath The
Sentient lies a series of pressure mats as shown in The mxj myClient object is a Java-based network
Figure 10. Each of the pressure mats is connected socket that sends the binary value to the
to a Teleo module (Figure 11); several Teleo Connection Bridge. The Connection Bridge is
modules are connected to create a society of simply a network socket which listens for
pressure mat hardware agents. messages from Max, sent from myClient, and then
passes the message as a Java Bean to the pressure
Figure 12 illustrates the data flows for the Max mat’s hardware agent. Recall that in the virtual
application developed to send a binary number environment version, the Communication Bridge
from the pressure mat to the hardware agent to handled message passing between Active Worlds
indicate the state of the pressure mat. When an and the hardware agent’s software sensors. The
object is placed on a pressure mat, the signal is pressure mat’s software sensor is the same as
passed through the Teleo module to the hardware before.
agent through the following data flow. The Teleo
module receives a digital signal from the physical Results of Implementation
pressure mat. If the binary value received is 1, then
there is an object on the pressure mat; if it is 0, Figure 14 illustrates the output from the CAD tool
then there is no object on the pressure mat. during operation in a virtual and a physical

intelligent room. First, the CAD tool loads the Conclusions

HADLIR agent description
(pressuremat.properties and pressuremat.xml). The HADLIR representation is founded around
Then, the pressure mat hardware agent instantiates the concept of modeling hardware in an intelligent
itself into the multi-agent society called "Office." room as agents having sensor and/or effector
modalities with rules and goals. This high level
Figure 13 illustrates what is seen in the simulation representation of the hardware in intelligent rooms
in Active Worlds. In the images of Figure 13, the captures the attributes and relationships that
perspective shown is from a first-person (avatar) determine hardware behavior and actions while
view. Thus, the left hand image is "from a distance" the Jess rule-base forms a basis for the degree of
whereas the right hand image is "up close." The intelligence that can be embedded into the
hardware agent performs the color changes based hardware. The HADLIR representation of
on the sensor data it receives. When the avatar is intelligent room hardware as a hardware agent,
distant, the pressure mat is inactive (color green). comprised of its description as an agent, a rule
When an avatar moves close to the pressure mat, base, and Java-based software sensors/effectors,
the pressure mat hardware agent, based on the rule is suitable for prototyping intelligent rooms in
base, activates. For demonstration purposes, we virtual environments and implementation in a
visualized the activation by having the hardware physical intelligent room. The physical
agent send the message "action create color white" implementation is the same as the virtual
to Active Worlds. If the avatar moves out of the environment simulation apart from the fact that
proximity then the color of the wall object the hardware agent accesses sensor data from a
representing the pressure mat is changed back to physical hardware device. We were able to
its original color. demonstrate that a CAD tool could allow a
designer to prototype and simulate an intelligent
In the physical intelligent room, an object is placed room in a virtual environment and then implement
on the pressure mat. At that instance, the message the room using the same description. The
is passed through the Teleo module to the hardware demonstration offers evidence that the basic idea
agent. As it is received and parsed, the hardware of describing hardware as agents in intelligent
agent displays the message "input value 1" in the rooms is a workable model for building a CAD
output screen. The rule triggers and the message tool for intelligent rooms.
"goal ON asserted" is displayed. When the object
is removed, the sensor displays the message "input At present, the HADLIR representation and
value 0" and the rule triggers; the resulting authoring tool caters for the description of devices
message is "goal OFF asserted." without dependencies. An example of a "device"
Figure 13. Images of the pressure mat simulation in Active Worlds. The pressure mat hardware agent changes the color of the wall
object representing the pressure mat

with dependencies would be augmented reality. The creation of intelligent environments presents
The hardware for augmented reality includes a a new level of complexity in architectural design
printed pattern, a digital video camera, and a raster because, in addition to the description of the state,
display device. We are currently experimenting the architect-designer must also describe the
with the relative merits of including these function and behavior of the computational objects
dependencies into the HADLIR representation that enact the environment’s intelligent behavior.
itself or into the rule-base. Others have defined Thus, a new type of CAD tool is necessary in
these dependencies in the ontology itself. For which the specification of how the intelligent
example, the Semantic Space ontology (Wang, environment operates is incorporated into its
Dong, Chin, Hettiarachchi, & Zhang, 2004) geometric description. The HADLIR
includes friendOf and colleagueOf in describing representation presented in this paper describes a
relationships between people in a context. This way to incorporate "machine instructions" into the
idea could be extended into the HADLIR geometric description and to use the description
representation. On the other hand, these for both virtual prototyping and physical
relationships may also be included in the rule set implementation.
as goals to allow each hardware agent to partner
with any number of devices which could satisfy
the goal, rather than a pre-defined partner. The
ontology could encapsulate functional
dependencies which could be satisfied by a class
of devices – for example, a digital video camera,
rather than a light, could satisfy the functional need
for illumination by projecting a white image. We
believe these issues warrant closer scrutiny.
Figure 14. Output from HADLIR implementation of a pressure mat

References Intille, S. S. (2002). Designing a Home of the

Future. IEEE Pervasive Computing, 1(2), 80-86.
Bobick, A. F., Intille, S. S., Davis, J. W., Baird, F.,
Pinhanez, C. S., Campbell, L. W., Ivanov, Y. A., Kreuger, T. (1996). Like a second skin, living
Schütte, A., & Wilson, A. (2000). Perceptual user machines. Architectural Design, 66(9-10), 29-32.
interfaces: the KidsRoom. Communications of the
ACM, 43(3), 60-61. Kreuger, T. (1998). Autonomous Architecture.
Digital Creativity, 9(1), 43-47.
Chen, H., Finin, T., & Joshi, A. (2004). An
ontology for context-aware pervasive computing Pask, G. (1969). The architectural relevance of
environments. The Knowledge Engineering cybernetics. Architectural Design, 7(6), 494-496.
Review, 18(3), 197-207.
Ranganathan, A., McGrath, R. E., Campbell, R.
Coen, M. H. (1997). Building Brains for Rooms: H., & Mickunas, M. D. (2004). Use of ontologies
Designing Distributed Software Agents. The Ninth in a pervasive computing environment. The
Conference on Innovative Applications of Artificial Knowledge Engineering Review, 18(3), 209-220.
Intelligence (IAAI-97), (pp. 971-977), Providence,
Rhode Island: AAAI Press. Smith, G. (2003). KCDCC AW Package. Sydney:
Key Centre of Design Computing and Cognition.
Coen, M. H. (1998). Design principles for
intelligent environments. The Tenth Conference Smith, G., & Gero, J. (2004). Describing situated
on Innovative Applications of Artificial design agents. In J. S. Gero (Ed.), Design
Intelligence (IAAI-98), (pp. 547-554), Madison, Computing and Cognition '04 (pp. 439-457).
Wisconsin: AAAI Press. Dordrecht: Kluwer Academic Publishers.
Friedman-Hill, E. (2003). Jess in action: rule- Wang, X., Dong, J. S., Chin, C., Hettiarachchi, S.
based systems in Java. Greenwich, Connecticut: R., & Zhang, D. (2004, July-September 2004).
Manning Publications Company. Semantic Space: An Infrastructure for Smart
Spaces. IEEE Pervasive Computing, 3, 32-39.
Gage, S. (1998). Intelligent interactive
architecture. Architectural Design, 68(11-12), 80-
85.
Gellersen, H., Kortuem, G., Schmidt, A., & Beigl,

M. (2004, July-September 2004). Physical
Prototyping with Smart-Its. IEEE Pervasive
Computing, 3, 74-82.
Gruber, T. R. (1993). A translation approach to

portable ontologies. Knowledge Acquisition, 5(2),
199-220.
Hirsh, H., Coen, M. H., Mozer, M. C., Hasha, R.,
& Flanagan, J. L. (1999). Room service, AI-style.
IEEE intelligent systems, 14(2), 8-19.
Huhns, M. N., & Seshadri, S. (2000). Sensors +

Agents + Networks = Aware Agents. IEEE Internet
Computing, 4(3), 84-86.
Hunt, G. (1998). Architecture in the ‘cybernetic

age’. Architectural Design, 68(11-12), 53-55.

M. Fineout / The Tower of Babel
The Tower of Babel:

Bridging Diverse Languages with Information
Technologies
Matthew G. Fineout1
1
Carnegie Mellon University + EDGE Studio
Abstract
New digital tools or information technologies are providing the means for architects to realize unprecedented
architectural creations. Unfortunately, the promise these technologies hold is far from their potential expression
in the built physical environment. A contributing cause to this disjunctive state is the multiplicity of languages
and knowledge sets employed by the various team members or actors engaged in a building project. From the
cost models of the owners to the shop drawings of the fabricators, each actor views the project in terms
specific to their individual discipline. In order to successfully engage the building process, these new
technologies must account for this condition and develop means in which to span across traditional boundaries.
This paper will examine the disjointed and fractured nature of the building project and identify opportunities
for the deployment of information technologies to bridge boundaries, ultimately providing for and delivering
architectural projects of unparalleled precedence. Specific aspects inherent to these technologies will be
examined to understand where their application may benefit the building process. The key attributes this
paper will focus on include: visualization tools, centralized database, cross discipline platform tools and
novel forms of information representation. A case study of an architectural project will serve as the means in
which to study the successful implementation of these attributes and their resulting impact on the design
process and building project. This study will demonstrate how information technologies can be implemented
within the multifaceted framework of conventional building projects to yield a project of unprecedented
form.

Introduction information technologies were instrumental in its

development; how these technologies facilitated
There has been much written lately on the promise collaboration between a diverse group of team
new digital technologies hold for the architectural members engaged in the building project and at
and building industry. For the most part these what points these technologies were introduced
writings have focused on individual technologies to enhance the process as well as the project. This
isolated in effect from the broader architecture, paper will demonstrate that through the successful
engineering and construction industry and the implementation of information technologies the
processes involved in a building project. Instead project benefits through collaboration and
of focusing on a specific technology, this paper efficiencies gained through the work process
will examine how information technologies as a resulting in added value to the end product.
whole can be implemented within the overall
building scheme to enhance the process and the Constraint Modeling
end product. To facilitate this study, the process
involved with the development of an architectural The commission for this particular project was
project will be examined to highlight where won through a competition. Prior to the initiation
opportunities lie to employ these advanced of the competition the project had been developed
technologies. through 30% design documents. The winning
architects proposed an unconventional scheme that
literally opened the subway line and station to the
adjacent urban surroundings. This strategy
revealed the inner public workings of the station
and linked them to a broader urban public context.
Prior to the competition only two-dimensional
forms of representation had been generated for the
30% design document phase; there was not a
comprehensive 3D computer model of the project
or of the site to that point. Due to the complexities
of the site in terms of urban infrastructure networks
and the desire to link the public functions of the
station to the broader public of the surrounding
context, the architects felt a 3D computer model
was necessary to gain a better understanding of
the existing conditions in place.
The ensuing 3D computer model initiated by the

architects modeled both the physical features and
associated constraint envelopes of the site.
Constraint envelopes refer to spatial zones whether
occupied or not that are inherent to specific urban
networks. For instance, not only was the rail line
modeled but also the dynamic clearance envelope
associated with the rail line. This envelope maps
the spatial zone associated with the train car upon
Figure 1. Computer model screen capture illustrating which nothing can encroach. This strategy for
constituent components that form the mass transit project.
modeling extended to the streets and associated
vehicular traffic zones, pedestrian zones and even
The paper will use as a point of reference the case to utility infrastructures and what were termed ‘no
study of a mass transit building project. This fly zones’, areas where potential utilities could be
project is a subway station that forms one inserted in the future. The site consists of multiple
component of a comprehensive urban layers of urban infrastructure networks that require
infrastructure project for the extension of an mediation for the functionality of the station. With
existing mass transit rail system. Several aspects the physical features of the site and its constraints
of the project will be examined to illustrate how modeled, it became clear where opportunities lie

for the station to engage the adjacent surrounding

context. Initially the constraint model served as
a means to identify vertical circulation zones to
move people from street level to platform level.
As the model developed it was instrumental in
organizing and articulating pedestrian flows in and
around the station. Not only did it describe how
to interface with the surrounding public networks,
but also how the project could articulate and
enhance these public networks. This process
continued which ultimately led to strategies for
the organization and configuration of the station.
The constraint modeling process was also

instrumental in gaining consensus among the
various actors engaged in the project. It enabled Figure 2 . Screen Capture from the computer model.
Horizontal tubes are dynamic clearance envelopes associated
discussions between the surrounding stakeholders, with the train cars.
the client and engineers with respect to various
site and project development issues. Each team
member came to the project with their specific
interests and concerns and a bias towards
understanding particular aspects of the station.
The interests of the stakeholders lie in the
amenities and benefits the station could provide
to the surrounding area while the engineers
focused on the functional aspects of the project
and the client straddled both concerns. The
computer model was instrumental in describing
various issues related to the station in terms that
were clearly understood by all parties. The
computer model was able to illustrate how issues
inherent in the site and the functional requirements
of the station would precipitate particular design
strategies. In turn these strategies were developed Figure 3. Screen Capture from the computer model. Vertical
through collaboration with the various team volumes represent pedestrian access between street and
members. This enabled the team members to platform level.
participate in an active dynamic relationship as
opposed to passive bystanders which is typically
the case (Boland, Lyytinen and Yoo, 2002). In resulted in an increase of engineering disciplines
this manner the process somewhat differed from engaged on the project. This condition placed a
the conventional project approach where the critical emphasis on the ability to share
architect imposes their vision upon the project information between team members to enable
team. Instead the computer model fostered an project development. Initially this process was
atmosphere of collaboration between team cumbersome and prohibitive in some respects, due
members that allowed the project to be borne from to the techniques used for information exchange.
the constraints of the site, the expectations of the There are two factors that led to difficulties for
client and stakeholders and the functional sharing information between engineering
requirements of the station. disciplines; two-dimensional forms of
representation and site coordinate systems specific
Centralized Database to each discipline. Two-dimensional forms of
representation tend to generate gaps in information
As the design for the station progressed the project (Boland, Lyytinen and Yoo, 2003). Their form
grew greater and greater in complexity. This results in discontinuities of understanding within

in overall continuous body of information. Instead as a centralized database that gathered information
of describing an overall form, information is only generated from each engineering discipline.
identified on a two-dimensional plane. In Information between disciplines was treated in a
addition there is a systematic methodology that consistent manner which eliminated any biases in
drives where slices of information are taken or its description. The 3D capability of the computer
where two-dimensional planes interface with the offered the advantage of modeling physical objects
object of inquiry. Typically these planes or section or systems in their entirety, eliminating partial or
cuts are oriented perpendicular or parallel to the piecemeal descriptions inherent to 2D
major axis of the work assembly. If the associated representations. This was a major asset in terms
work between two separate engineering disciplines of understanding the various systems and their
is not in alignment these section cuts will not be relationships. With the entirety of all pertinent
coplanar. physical systems modeled in a uniform coordinate
system, inquiries into relational issues could now
This last issue couples directly with the use of be performed. This was instrumental in
coordinate systems that are specific and unique to understanding potential conflicts or interferences
each engineering discipline. This uniqueness does between engineering work assemblies. This
not allow for a common language or syntax in capability also provided for a much tighter
which to share information, but instead views work integration among the various systems of the
associated with each discipline in terms specific station. This approach enabled the computer
to that discipline. Roadway engineers used the model to act as a boundary object; introducing a
centerline of road as a reference in which to common language or syntax that provides for the
generate and understand their work with respect sharing of information across boundaries (Carlile,
to the site. The centerline is divided into equal 2002). Now, instead of multiple information sets
intervals. Locations and measurements are then particular to each team member, one database
understood in terms of intervals along and could be shared in a collaborative manner for the
perpendicular to this centerline. It’s as if one took development of the station. The data could be
a three-dimensional Cartesian grid and warped it accessed and modified as required. It could be used
to follow the road. This system is then also used to analyze and coordinate relationships between
in determining the orientation of two-dimensional various engineering work assemblies.
section cuts. To be consistent with the system all
sections are cut perpendicular to centerline of road.
The tunnel engineers employed the same
methodology for describing information as the
roadway engineers, however their reference
system was based on the centerline of track. At
this particular site centerline of road and centerline
of track did not fall into alignment, nor were they
parallel. Section cuts by the roadway engineers
did not correspond to nor were they coplanar with
section cuts by the tunnel engineers. This made
sharing information between the two disciplines
problematic. Even though the methods of
describing information were consistent between
disciplines, the reference system employed by each
differed which led to difficulties in the exchange
of information. Since the scope of work associated
with a specific discipline was developed within
their own unique coordinate system,
understanding relational issues between work
products was difficult at best.
The computer model was instrumental in spanning

Figure 4. Plan drawing of site, dashed gray line is centerline
these boundaries. The computer model served of road, dotted lines are centerline of trac.

cross platform sharing between disciplines.

Instead of each team member generating a
database from scratch specific to their discipline,
the existing Rhinoceros database could be easily
translated into terms that were meaningful to each.
This represented a major time saving procedure
and made information sharing between disciplines
a seamless matter.
The use of the Rhinoceros software was very

instrumental in the development of the project.
This ability to easily share data between various
disciplines had two effects: 1), an iterative process
for the development of the station unfolded in
contrast to the conventional linear process and 2),
Figure 5. Right, Section cut drawing through proposed transit the team members collaborated in a much more
tunnel and existing utilities. participatory dynamic manner than the passive
neutral relationship which is typically the case
Cross Platform Tools (Boland, Lyytinen, Yoo, 2002). The iterative
process allowed for design strategies, in contrast
With the development of the project came the to set and unyielding design solutions to tackle
greater need for more specialized information the many issues encountered with the development
which resulted in an increase of engineers of the project. In this way it differed from the
representing diverse fields of interest. To satisfy conventional design approach which is
criteria with respect to each engineering discipline, characterized by a linear process of development.
it was necessary to view the project in a variety of This linear process of development implies that
terms meaningful to each. Therefore, the project the outcome of the design is known in advance
takes on a number of different meanings that and that the development of the project merely
correspond to the various disciplines. This consists of the motions of moving from point A to
condition can be considered a difference in point B. This method of development however is
semantics among various parties viewing a very unforgiving when unforeseen conditions are
common language. Whereas in the previous encountered. This process precludes possibilities
example a common language allowed for the or opportunities that may arise over the course of
exchange of information amongst a diverse development for the project. It also preempts the
audience, semantic differences account for ability of team members to actively contribute to
different interpretations of that common language the design development process. On the contrary,
(Carlile, 2002). For instance, as the project an iterative process allows the design to evolve in
developed, smoke evacuation became a critical relation to a number of factors influencing the
factor for the functionality of the project. This project, whether they stem from the physical
meant that the station needed to be viewed in terms attributes of the site, the functional requirements
of its behavior with respect to fluid dynamics. This of the project or various concerns of the team
represents a very specific set of terms and criteria members. As issues or conditions arise over the
for the evaluation of the station. These terms course of development, design proposals can be
contrast sharply for instance with the terms a quickly modeled and in turn translated for analysis
structural engineer employs to evaluate the station. and evaluation by the respective engineering
disciplines. This process continues until the
In the conventional method of project various conditions or issues associated with the
development, sharing information between project play themselves out and ultimately reach
disciplines would ultimately mean recreating the a state of resolution.
project from scratch in terms specific to a
particular discipline. The computer software For this iterative process to occur, the team works
that was employed for this project avoided this in a much more integrated and collaborative
scenario. The software used for the project was manner; team members are actively engaged
Rhinoceros which lends itself extremely well to instead of neutral bystanders (Boland, Lyytinen

and Yoo, 2002). This allows for the various

disciplines to influence the design of the project.
As issues arise or conditions develop, they can
easily be evaluated, assessed and rolled into the
design of the project with this process. For
instance, if through the fluid dynamic analysis,
information is gained that requires the
modification of particular surfaces to aid in the
removal of smoke, this information can be input
back into the design model for alteration. Once
the design model is altered, these modifications
can then be brought back into the fluid dynamic
model for furthering testing and analysis until the
model performs within acceptable criteria limits.
This process continues in each discipline until
resolution is reached for the various issues Figure 7. CFD model screen capture illustrating design fire
affecting the station. The greatest benefit this and associated temperature ranges.
process yields is a much more integrated project
between the various systems that constitute its
form. Since the project essentially is a compilation describe information in a number of different
of systems, this iterative process allows for a much formats. Just as a spectrometer can identify
tighter integration of these various systems. This different element compounds invisible to the
is the added value the iterative process contributes naked eye, so too can digital technologies
to the project. The cross platform sharing of foreground particular information that is relevant
information between disciplines allows the project to a project. This becomes extremely useful in
to reach a much higher and sophisticated level of conversing with diverse groups of people. Again
development. it allows for a common understanding of
information among various team members. This
ability became extremely important for this
particular project. Due to the complexities of the
site and program and their interrelationships,
issues of project configuration and constructability
became driving factors for the station. It was
incumbent upon the station to mediate a number
of conditions that varied from street level down
to platform level. These conditions suggested
forms that are not conventional in the construction
industry.
The construction industry favors forms that are

orthogonal due to a number of factors stemming
from industry standards of manufactured products
(straight runs of steel) to conventions of assembly
in the field. These factors coalesce to create a
Figure 6. Computational fluid dynamic (CFD) model screen bias in the construction industry towards
capture illustrating smoke layer with smoke exhaust system rectilinear forms. The complexities of the site and
off. the misalignment of the street and rail line below
did not lend itself well to a rectangular form.
Several problems were encountered with the
Visualization Tools ability of a rectilinear form to negotiate the
specifics of the site. To satisfy these conditions a
In addition to the tools afforded by digital simple geometry in the form of a cylinder was
technologies noted earlier, the computer also offers proposed. The cylinder served as the controlling
unique visualization capabilities. It can filter and geometry for the exterior envelope of the station.

What made matters unique is that only a very small area than a corresponding rectangular form, and
portion of the overall geometry comprised the therefore less materials and weight associated with
envelope of the station. In addition that portion the structure. The provision of visualization tools
that constituted the envelope did not follow the by digital technologies allowed this to be
dominant ruling lines of the geometry but was demonstrated across a broad audience, from the
situated off axis. It’s as if someone took a client through the engineers. These visualization
cardboard tube and cut a small piece out that was tools enabled people from various backgrounds
not perpendicular or symmetrical to the length of and knowledge sets to understand the value
the tube; the resulting piece would appear skewed afforded by such an approach. Again it is through
and irregular. When the station envelope was the effective deployment or application of
viewed alone it appeared complex since the information technologies which aid in spanning
simplicities of the overall controlling geometry across knowledge sets of diverse team members,
were not evident. Typically a proposal like this enhancing the project development and hence end
would meet resistance due to the unorthodox use product.
of a unique geometry and the associated
constructability issues. Issues of costs in relation
to unique framing members, forms of fabrication
and assembly would immediately be cited.
However the visualizations tools provided for by
the computer were able to circumvent these
arguments. When the constructed portion of the
frame was shown to be part of a larger regular
geometry, it could be demonstrated that the
framing members were consistent and based on a
single radius. Not only were the framing members
consistent but also all the cladding and connection
details. This proved to be not only feasible but
cost effective, due to efficiencies gained through
the use of a cylindrical form which has less surface
Figure 9. Computer model screen capture. Resulting station

envelope constructed as a portion of the overall cylindrical
geometry.
Conclusion
These examples have shown how information

technologies associated with the new digital tools
of the computer are instrumental in the
development of a project. As can be seen, the
process of design for the station was enhanced
through a collaborative team effort created through
the use of information technologies resulting in a
better end product. The successful deployment
of these technologies helped in the restructuring
of the conventional working practices of the team
calling for tighter relationships between individual
Figure 8. Computer screen capture illustrating the cylindrical team members (Boland, Lyytinen and Yoo, 2002).
geometries that form the envelope of the station. This collaborative process afforded by information

technologies results in a greater stake in the project

by individual team members. This raises the
expectations of the parties involved and
consequently the design of the project.
Acknowledgments
I would like to thank all the team members that

participated in the collaborative effort on this
project especially Jerry Marinzel from the Port
Authority of Allegheny County; Greg Yates, Paul
Diez, Judson Herter, Peter Mauritz and Rick
Goddard from DMJM Harris, Dr. Peter Gehrke
and Dr. Nicholas Agnew from Maunsell Australia
Pty. Ltd., John Schneider and Gil Taylor from
Atlantic Engineering Services; James O’Callaghan
from Eckersley O’Callaghan Structural Design;
and Gary Carlough, Erik Hokanson, Rob
Pfaffmann, Carl Bergamini and Tim Roos from
the Light Motion Collaborative.
References
Boland, R. Jr., Lyytinen, K., Yoo, Y. (2003). Path

Creation with Digital 3D Representations:
Networks of Innovation in Architectural Design
and Construction. Sprouts Working Papers on
Information Environments, Systems and
Organizations, Vol. 3
Boland, R. Jr., Lyytinen, K., Yoo, Y. (2002).

Appropriations of Information Technologies to
Improve the Performance of Doubly Socio-
Technical Systems. Weatherhead School of
Management, Case Western Reserve University.
NSF Grant Proposal.
Carlile, P. (2002). A Pragmatic View of

Knowledge and Boundaries: Boundary Objects
in New Product Development. Informs,
Organization Science, Vol. 13, (No. 4), 443, 444.
Rhinoceros, NURBS modeling for windows,

Version 3. Robert McNeel & Associates, copyright
1993-2002

T. Modeen, C. Pasquire, R. Soar / Design Ground - An Iconic Tectile Surface
Design Ground - An Iconic Tactile Surface

Thomas Modeen1, Christine Pasquire2, Rupert Soar3
1
Loughborough University
2
3
Abstract
This paper forms an intermediary summary of a project which aim is to suggest an alternate methodology for
utilizing additive Rapid Manufacturing (an evolved rendition of Rapid Prototyping), for the conceptualization
and fabrication of design and architecture. It plans to do so by establishing a methodology that is innate and
a direct reflection of the additive RM production process. The project also aims to address the seemingly
divisive discrepancy between the process of digitally conceiving a design and the intrinsically somatic way
we perceive it.
Such aims are explored through a surface design that is not predominantly guided by visually derived nodes
but instead relies on a form of ‘tactile iconography’ as a means for expressing and amplifying various qualities
and elements found in its vernacular. The resulting design would be very difficult, if not impossible, to make
by any other means.

Introduction surface around the neighborhood of London’s

Exhibition Road in South Kensington, an area
“The entire body may ‘know’ a dance. In such occupied by number of the city’s main museums
cases, knowledge is all the more likely to be and galleries. However, it wasn’t necessarily the
physically inscribed, without an overtly more, call it, ‘highbrow’ aspects of this locale that
intellectual component.” were of interest, but the more peripheral marks of
human/ urban occupation – the blots of discarded
Quote from Malcom McCullough’s book, chewing gum that seemed to saturate the
‘Abstracting Craft: The Practiced Digital Hand.’ streetscape. There must have been thousands, tens
of thousands, hundreds of thousands of these,
The project was instigated based on two defining predominantly off white, specks carpeting the
parameters. main segments of the road and flanking sidewalks.
Firstly, to produce a sensorially inclusive design. Observing these blots closer it was interesting to
To produce a design that ‘looks’ beyond visual, note how one could decipher their approximate
or visually derived, stimuli. To allow the haptic location on the road simply by their shapes and
(touch), auditory (sound), kinesthetic (motion), the patterns they formed. Very dense around and
and all the substrata of such perceptual just outside the sidewalk curb. Dense, but more
formulations (temperature, vibration, pitch, treble, pronounced, by pedestrian crossings; less dense
rhythm, obliqueness, texture, friction, etc.) to play and elongated by the main or central footpaths;
a more involved role in the conception of a design. and sparse by the inside (building side) of the
To produce a design framework within which such sidewalk.
aims can be achieved…
These patterns, somewhat suggestive of Braille
Secondly, to use Rapid Manufacturing (RM) in (the tactile reading and writing method used by
the formulation of the designs conceptual the visually disabled), provided the initial clue for
foundation as well as for final physical fabrication. how to proceed.
To, through the conception of a design, suggest a
way in which this quite remarkable, and still What if, instead of merely passively interpreting
evolving, technology could be formulated in a way these aforementioned patterns, one would provide
that reflects and utilizes its intrinsic qualities them with meaning, immerse them with
instead of merely mimicking alternate fabrication performative attributes that could inform one of
methods… their immediate vernacular on a more planned and
specific level? Could one, by manipulating such
Project Catalyst patterns’ surface and consistency, create an
additional strata of, non-intrusive, sensory stimuli
“Wittgenstein liked to say that the most difficult that, ideally, would not only enrich the streetscape
problems are the ones right in front of our eyes, on a purely perceptual and affective level, but craft
the ones we don’t see as problems. Those are the something that could also act as a functional tool
ones we have to struggle to perceive.” for conveying empirical information to those with,
Quote from James Elkins’ book, ‘The Object Stares and without, a sensory disability? Could the
Back - On the Nature of Seeing.’ surface be made to perform an amplifying role in
clarifying the features of its immediate environs?
The project was initially constructed around a
competition brief provided by the Swedish Design Implementation
flooring company ‘Pergo’. The brief was left
somewhat undefined, requesting its participants “As the most ancient and largest sense organ of
simply to design something ‘innovative’, and that the body, the skin enables the organism to learn
such novelty should happen in/ on the plane below about its environment. It is the medium, in all its
ones feet. differentiated parts, by which the external world
is perceived.”
How such an open brief was developed for this
particular project was initially prompted by the, Quote from Ashley Montagu’s book, ‘Touching -
to some extent peripheral, qualities of the paved The Human Significance of Skin.’

Figure 1. The project was catalyzed by the ‘field’ of discarded chewing-gums on London’s Exhibition Road. These
patterns were used to suggest how the consequent patterns of the surface should be implemented
How such concerns were put into practice required (as, say, the entrance of a building or an
that the aforementioned gum-patterns be approaching pedestrian crossing).
conceived within a more regimented framework.
This entailed that the initially more randomly One could even, by manipulating the consistency
distributed gum-blotches were conceived as a fish- of the fabrication material (something that is
scale like array of small tactile components that possible thanks to the minuteness and accuracy at
allow for a surface with a binary and directional which one can control the composition of a
pattern (the surface is smoother in one direction, material through this technology), change its
and ‘rougher’ in the other). tactile and even auditory qualities. I.e. by making
the surface more or less dense, and by patterning
This surface could consequently be ‘enriched’ by such consistencies according to various utility and
applying simple binary ‘deformations’ to the core performance based configurations, one can make
pattern. The scales of the surface could be made the surface ‘feel’ and even sound different (when
(gradually, as on a sliding scale) smaller or larger, walked upon) at various locations.
thinner or wider, more or less pronounced, etc.,
according to various predetermined and intuitive How such features are actually applied is explained
haptic iconographies. This fish-scale pattern could further in the following chapter, however here it
also be curved towards, and obliquely angled, to is worth noting that the paradigms within which
show and insinuate various surrounding features the aforesaid notions have been conceived in have

Figure 2. Illustrations of how one can amplify the affective qualities of the surface by simple binary means – making
it more or less pronounced, wider or narrower, smaller or larger, etc. The scale pattern can also be made to curve or/
and angle towards various relevant nodes in its vernacular
become somewhat arbitrary – for to fabricate the

design(s) it is its mass and volume, not its
complexity or the repetition of a set formulae, that
matters. Here there is no need for a fabrication
template (entailing a set of identical pieces) as for
the RM machines it is equally laborious to produce
two (or more) alike pieces as it is to produce a
number of, roughly equal in mass and volume, but
still drastically different components. This is both
the processes advantage and its crux, as the
seemingly almost infinite freedom this means of
production seemingly entails, it still needs to be
defined and thus confined, according to some form
of predetermined parameters – a protocol that is
Figure 3. An example of how the surface properties could be still in need of further revisions before a set, almost
affected – here by setting up a truss matrix which constraints
were determined by the properties of both the fabrication
innately settled, modus operandi has been
process and its accompanying material. Here this entails that established.
the make-up of the designs truss based structure can be
adjusted according to a responsive algorithm that can be
attuned to reflect particular tactile traits, such as softness,
Design Application
elasticity, or a variety of different textures… This aspect of
the project was developed in conjunction with Sean Hanna “Smooth surfaces invite close contact, while rough
and Siavash Haroun Mahdavi from University College
London, the Bartlett (school of architecture) and Department
materials such as hammered concrete generate
of Computer Science movement in wide radii around corners and more
careful, tentative movement through corridors.
Changes of texture often signal special events and
all been guided by a framework set by what the can trigger a slowing or quickening of pace. It
relevant technologies (both software and would be possible to generate a whole
Hardware based) allow for. This entails that the choreography of movement through the
size, patterning and modulation of the designs composition of textural changes alone.”

Quote from Kent Bloomer’s & Charles Moore’s • More solid surface consistency than
book, ‘Body, Memory and Architecture.’ along the main path.
• Slightly convex slant (surface
How such aforementioned interests could be obliqueness) towards the building
applied in a design is illustrated in the example entrance.
below. • Surface sound ‘B’ (a ‘harder’ sound
than sound ‘A’).
In a generic streetscape four key nodes of a
pedestrian sidewalk are recognized: the main (or
• The curb:
‘spine’) path of the sidewalk, the entrance to a • Directional, skewed, fish-scale pattern.
building, the curb, and the area of a pedestrian • Higher, more pronounced, texture.
crossing. • Surface level slightly higher than along
the main path (the surface is oblique
Applying the aforesaid principals to each of these down towards the main spine of the
four key locales the various surface properties sidewalk).
could differ as follows: • More solid surface consistency.
• Surface sound ‘D’ (more ‘hollow’ than
• The main path of the sidewalk: surface sound ‘A’).
• Directional (fish-scale) pattern applied • The pedestrian crossing:
along the main spine of the pedestrian • The fish-scale pattern fades into blister
passage. paving.
• ‘Medium’ porous consistency.(1) • Slight, gradually increasing, concave
• Level, horizontal, surface. surface tilt towards the crossing.
• Surface sound ‘A’ (generic surface • ‘Softer’ surface base consistency.
sound). • Surface sound ‘D’ (more ‘muted’ than
• The entrance to a building: surface sound ‘A’).
• Directional fish-scale pattern towards
the entrance of the building.
Figure 4. A generic streetscape in which four key nodes are recognized

Figure 5. Breakdown of the variations in perceptual properties between the four key nodes of
the generic streetscape
Even by the application of such a reduced Quote from Juhani Pallasmaa’s book, ‘Polemics
taxonomy of iconic nodes an almost infinite - The Eyes of the Skin.’
variety of interpretative options and variations of
patterning cold be applied to the surface. The As insinuated above the project has also explored
inherent flexibility of this parametric process also additional means for how the principals examined
allows for the adjustment of the surface for a could be adapted to serve a variety of other
variety of different formats, be these for interiors purposes. These include developing the Ground
(entailing a ‘toning down’ of the surfaces Design in more, call it, ‘expressive’ ways to, say,
coarseness), to even applying these principals to change its color when the surface get dangerously
a wall surface (as briefly outlined below). hot (which would be useful when in the vicinity
of a stove), or, by making subtle indentations in
A number of physical prototypes have also been the surface, make it form, when it rains, ‘Designer-
fabricated, both through the use of a Selective Puddles’ in the shape of images letters or paths. It
Laser Sintering (SLS) machine, and a could also be conceived in the format of a wall
Stereolithography (SLA) machine, allowing for surface, a continuous organic relief of sorts, that
the testing both the tactile surfaces operational, would allow one to perceive or read things through
wear & tear and function related, practicalities, as the use of ones palms and fingers. Even further
well as the surfaces more ethereal characteristics, applications are being explored – the potentials
how it actually ‘feels’ and how easy or intelligible for such parametric conceptions seem to be almost
the surfaces intended expressive qualities actually limitless…
are.
Conclusion
Additional Adaptations of the Design
“The hands of the sculptor are independent In conclusion, the aim of the project is to establish
organisms of recognition and thought; the hands a way of designing something that utilizes and is
are the sculptor’s eyes.” instigated by digital (CAD) based means, here

Figure 6 B. segment of the Tactile Wall-Scale Surface

prototype, made through Stereolithography
or just purely pleasurable stimuli that would

provide an additional enriched strata of embedded
data that could reflect and interact with its users
and environment without adding further to the
already saturated clutter of the urban fabric.
Figure 6.A ‘Rhino’ modeled, sectioned Selective Laser The intent of the project is to suggest a somewhat
Sintering (SLS) model of the Tactile Paving
deviant approach to how architecture can be
perceived and conceived. Its aim would not be to
through the use of Rapid Manufacturing, but does exclude the other senses from the equation but to
so without omitting the more innate sensory create a more comprehensive and attuned
qualities of what usually makes one respond and understanding of how, what we usually define as
appreciate a design (its physical presence, things ‘touch’, could be included more inseparably into
such as its texture, apparent weight, or even the the grammar and vocabulary according by which
subtle ‘sounds’ it makes). By considering all the we read, define, design and build our ‘architecture’
various tactile, auditory, in addition to the more by. By utilizing something as immediate as touch
obvious visual, aspects of the design already in as a catalyst for the brief the hope was to form the
the initial (digital) conception stage of the project, foundations of a more personalized and distinctive
along with attempting to tie such considerations approach for how to understand and define our
in with the process used for actually fabricating built environment. An approach that transcends
the design, hopefully a more accomplished way beyond the purely pictorial (conceptual) and
for how one could achieve a more sensorially ocular drive that seems to saturate so much of what
provoking and invigorating end result is implied. we today conceive of as architecture into a realm
more in tune with how we actually sense things,
The design is still in the process of being both within and outside ourselves.
implemented. Currently we’re in the process of
scaling-up, testing and fine-tuning further the
aforementioned assumptions. But hopefully the
final realization of the design will act as a non-
intrusive, yet, for those actively seeking or are just
generally more perceptually aware, subtly inciting
additional source of either empirically informative

Notes
(1) Here terms such as ‘medium’, or ‘sound ‘A’,

are used as purely, hermetically comparative,
expressions that are only referential to each other
instead of some more generic (haptic or tone
based) standard.
References
Bloomer, K., Moore, C. (1977). Body, Memory,

and Architecture (p.71). New Haven, CT: Yale
University Press.
Elkins, J. (1996). The Object Stares Back – On

the Nature of Seeing (p.58). New York, NY:
Harcourt Inc.
McCullough, M. (1998) Abstracting Craft: The

Practiced Digital Hand (p.5). Cambridge, MS:
MIT Press.
Montagu, A. (1971) Touching – The Human

Significance of the Skin (p.5). New York,NY:
Harper & Row Publishers.
Pallasmaa, J. (1996) Polemics – The Eyes of the

Skin - Architecture and the Senses (p.40). London,
UK: Academy Editions.

T. Yeu et al. / Smart Wall for Thermal Comfort
Applying Scientific Simulation to Integrate

Thermoelectric Conductor Module into
Architectural Design – Smart Wall for Thermal
Comfort
Tsou Jin Yeu1, Chan Yi Lee2, Mak Kwok Pui3, Du Ru Xu4, Liang Jian 5,
Yeung Kim6
1
The Chinese University of Hong Kong
2
3
4
5
6
Abstract
This paper presents the innovative architectural design concept, which is to integrate the new material and
technology into the building design to achieve the thermal comfort and at the same time reduce the energy
consumption of the building by making use of the renewable energy, including solar and wind energy. The
system is developed based on the idea of regional thermal comfort in building. The advantage of the system
is the environmental friendly approach, costless operation, reliability, flexibility, scalability and adaptability
for the integration to the building design. With the design concept, we tried to do two application designs in
two virtual sites. One is a badminton court for the 2008 Beijing Olympic Games and the other is a cooling
pond in a shopping mall. We will introduce how computational simulation can contribute to the prediction of
the performance of the design. We will also discuss how the computation simulation can help in the design
optimization process. Through the development of the new design integration of the material to the building,
we would like to feedback to the material industry to encourage further collaboration and development in the
material enhancement, so that both industries and the society can benefit from the advancement.

Introduction—Thermal Comfort and Smart The “smart material’ here means the
Wall Thermoelectric Conductor Module (TEC) (Figure
1), which is well known for its ability to convert
Usually thermal comfort is achieved by robust electrical energy into thermal energy. When there
ambient temperature control by centralized HVAC is a current passing through the Smart Wall, one
system. The weaknesses of conventional HVAC side will be cold and the other will be hot. Through
system are high energy consumption and the intelligent control and system integration in
complexity of system. In this project, we are going the building design, the Smart Wall can achieve
to explore the new way of thermal comfort and the regional thermal comfort.
energy saving by applying scientific simulation
to integrate the thermoelectric conductor module Advantages of Smart Wall
into the architectural design.
a) Environmental friendly approach. The Smart
Smart Wall for Thermal Comfort wall does not produce any pollution during the
operation. And at the same time, it makes use
Have you ever imagined a column is no longer of the renewable energy such as the wind and
merely a structural support that blocks the space solar energy for its operation. The operation cost
but a cooling oasis that people will love to get could be reduced.
close to and stay for a while or a gateway not only b) Reliability. There are no movable parts and
magnificent in gesture but also welcome you by hence, will not wear or fatigue. This makes the
real warm ambience? Corridor is always system very reliable. In case of malfunction, the
considered as a transition, cooling down and module design makes it easy to repair / replace.
sanctifying journey before entering an sacred
c) Flexibility. The system can be integrated with
space, now the effect could be achieved not only
the partition wall, ceiling or even be mobile. It
by architectural design but enhanced by regionally
can be integrated into different areas depending
cooling. So the main idea here is regional thermal
comfort, overcoming the limitation of robust
on the needs of regional thermal comfort
HVAC for sensitive thermal delight. through architectural design.
d) Scalability. The cooling load can be easily
Rationale adjusted through the intelligent control system
depending on the thermal requirement of the
Smart Wall is new architectural design integration occupants. Also the size and the scale of the
for the provision of thermal comfort in the indoor system can be changed up to the volume of the
environment. The Smart Wall project aims at space. It gives large flexibility to the design
integrating the smart material, advanced integration.
technology and innovative architectural design in e) Adaptability. The system can adapt to
order to achieve regional thermal comfort and also different functional requirement. Since the
energy saving in buildings. working rationale is through radiation and/or
with low speed fan assisted, it can adapt to strict
requirement such as the badminton court, where
requires no air movement in the playing field.
Computational simulations
The computation simulation plays an important

role in the design process and the optimization
process. To achieve the thermal comfort, we do
consider the air movement and the temperature
distribution in the space. In order to predict the
system performance, computation simulation is a
must to be applied due to its high precision.
Figure 1. The ration rationale of the TE module

Computational Fluid Dynamics (CFD) involves The water circulation system plays an important
analysis of fluid flow, heat transform and other role in Smart Wall. Compared with the
related phenomena by means of computer-based conventional forced air convection, it has much
simulation. The technology is powerful and better heat / cool dissipation capability. At the
spreads widely in industrial and non-industrial mean time, it is also used for energy storage. For
application areas. In our research, we will apply future designs, it is also possible to add solar panel
CFD to predict the internal flow and temperature and wind turbine to the system, so that more
distribution. Through the CFD simulation, we can, renewable energy can be used.
first verify the performance of the Smart Wall, and
second, further improve the architectural design Intelligent control system
by knowing more about the flow and temperature
distribution pattern. The intelligent control system is an important
component, which manages the system operation,
To verify the design options, we set up CFD such as choosing the power source (switching
models and physical model to evaluate their between the battery and the rectifier), and
performance. We also explore the possibility to controlling the flow rate of the water circulation.
integrate the Smart Wall with other green systems In order to provide precise temperature control,
and grey water system. This will be discussed in some intelligent control methods can be adopted.
the latter part of this paper. Under disturbance of ambient temperature and
heat load, the control system can adjust the input
Design experiment current to guarantee the required temperature. In
addition, the control system can also be integrated
In order to explore the possibility of using the to modern building control system for more
Smart Wall in the building design, we have chosen functions, such as remote monitoring and control.
two virtual sites for the design. The first one is a
badminton court for the 2008 Beijing Olympic Integrating the Smart Wall with
Games. And the second one is the shopping mall. Architectural Design
According to the different site contexts and design
criteria, we make use of the characteristics of the We identify the role of Smart Wall which is
Smart Wall to facilitate the function of the regional thermal comfort. We recognize the
buildings. The main concept is the regional cooing weakness of Smart Wall is its low COP (about
which aims to achieve the thermal comfort for 0.45) which is in no way competitive with
designated region in an enclosed space. traditional AHU (about 2.5-3). However, there are
rooms for Smart Wall to emerge. The main idea at
New Material & Technology this stage is: AHU targets ambient cooling; Smart
Wall targets regional cooling. Also, Smart wall has
Design of Smart Wall some unique characteristics and advantages that
AHU cannot achieve.
The thermoelectricity was first discovered about
180 years ago. The thermoelectric materials could The flexibility and the scalability are two of the
operate as a heat engine to convert heat into most powerful and useful characteristics. The size
electrical energy (Seebeck effect), or convert of the Smart Wall can be customized. It can be
electrical energy into thermal energy for cooling installed and fixed on the wall or ceiling, or be
or heating (Peltier effect). installed as a sculpture or a panel for dividing the
space. With the effective heat dissipation system,
Smart Wall consists of two primary components, the limitation of the installation can be eliminated.
TE modules and intelligent control system, and So the Smart Wall can be put closely to the human
two auxiliary components, a power supply (which and/or activity area. It is easier to achieve the
include a rechargeable battery and an AC power regional cooling since the space to be cooled can
rectifier) and a water circulation system. It has be minimized. Also, the cooling load of the Smart
three functions: heating, cooling, and power Wall is adjustable. It can be easily adjusted
generation. depending on the occupant’s needs. Through the
intelligent control system and sensor, the smart

Figure 2. Smart Wall units for cooling of the athletic field (left); the auditorium is cooled by low-speed-large-volume fan; the athletic
field is cooled by cold air units and radiation from Smart Wall (right)
wall can even “self-adjust” to incorporate with the As it is known that badminton is very sensitive to
occupant’s activity. For example, we can integrate air movement, Smart Wall is very suitable to tackle
the moving sensor to the Smart Wall in an activity the strict requirement for the badminton court.
room. Once it senses the human movement which
implies the activity starts, the control system will Since the court is required to accommodate certain
allow more current supply to the Smart Wall. So amount of audience, our design strategy is to
that it can interact with the human behaviors. Other divide the court into few different regions –
sample such as an “intelligent door handle” (for auditorium and athletic field. We use low-speed-
winter time) can also be designed based on the large-volume cooling strategy for the cooling of
characteristic and the concept of the Smart Wall. the auditorium (Figure 2). For the athletic field,
But we will not further discuss here. large area Smart Wall units will be mounted on
the wall and mobile Smart Wall units will be put
The efficiency and cooling capacity of the smart on the side of the field for the regional cooling for
wall are not competitive as AHU for large space the athletes (Figure 2). By low speed fans and
cooling. In some buildings, where part of the radiation, the playing region will be cooled down
enclosed space does not need the same amount of by the Smart Wall units so that regional cooling
cooling, presumably, Smart Wall can reduce the can be achieved without creating obvious air
overall energy consumption by cooling regionally movement (Figure 3).
only spaces where users occupy. As Smart Wall
can operate with unsteady current, it gives
flexibility and opportunity in utilization of
renewable energy. Solar panel and wind turbine
can be easily integrated into the system. The Smart
Wall can contribute to reduce the energy
consumption by the HVAC system. Even in the
space where temperature should be well
controlled, with the integration of the intelligent
control system, precise temperature can be
achieved. The intelligent control system can also
allow automatic switch to different energy source
(battery or electricity net) or adjust the current in
order to save energy.
Space cooling for badminton court of 2008 Beijing

Olympic Games
Figure 3. CFD simulation verifies that by using Smart Wall
This is a virtual site in Beijing to build the and low-speed-large-volume fans, the air movement in the
stadium can be well controlled
badminton court for 2008 Beijing Olympic Games.

By the intelligent control system, the Smart Wall “Cooling pond” for thermal comfort in seating of
units can be driven by two kinds of energy. The shopping mall
first one is the renewable energy (solar and wind
energy). During the sunny day, the solar panels Further to the first design prototype used in the
outside the stadium can collect the solar energy badminton court, we further investigate the
and convert it to electricity to supply the Smart potential of the Smart Wall for regional cooling.
Wall units. During the non-match days, the The idea of this design project is to install cooling
electricity can be stored in the battery for further ponds in shopping mall to provide a resting region
use. Wind energy is used and stored in the same for users.
way. During the cloudy or rainy days, the control
system will automatically switch to the battery or The resting area is regionally cooled by the Smart
the in-house electricity supply to maintain the Wall system. The essence design is the utilization
service of the Smart Wall units. of renewable energy from the wind turbine. This
solely overcomes the weakness of TEC and further
The hot sides of the units are cooled down by a revolves it into advantages over other cooling
water recirculation system to maintain the devices.
efficiency of the units. The water comes from the
pond outside. By this design, multi-function of the Wind turbine as a renewable energy source is not
water pond can be achieved: to facilitate the widely used because the fluctuating nature of its
efficient operation of the Smart Wall and energy supply which means there must be a storage
aestheticize the external environment of the battery for a steady electricity outcome. For typical
stadium. We also make use of the shelter of the use of wind turbine, it takes two steps of energy
seats. Solar panels are installed on top of the shelter transformation before the energy can be used,
so that during the sunny days, the shelter can act which is, from the turbine to the battery and then
as shading devices and the solar energy collector. from the battery to the appliance. Each
transformation indeed dissipates a considerable
The advantage of using the Smart Wall in the amount of energy.
badminton stadium is, first eliminating the air
movement; second making use of renewable With Smart Wall, this energy loss can be avoided.
energy to achieve cooling; third making use of the We do not need a storage battery because we do
surrounding facilities such as seating area and not need absolutely steady current to power up
water pond to facilitate the operation of the units. Smart Wall. Instead, we use a piece of thermal
Figure 4. Operation mechanism of “Cooling pond” in shopping mall

Figure 5. Thermal mass in front of the Smart Wall unit to regulate the temperature (left); CFD simulation, contour of static temperature
(right)
mass to regulate temperature (Figure 5). And we achieved by multi-means. With the CFD
use the wind energy directly by connecting the simulation, we can optimize these detail design
TEC to the turbines through the intelligent control such as the sizing and dimension. Figure 5 shows
system (Figure 4). the simulation result of temperature distribution
within the seating area.
Figure 4 showing the design concept of the cooling
pond in the shopping mall. The two sides of the Wind turbine and optimization
mall are shops and the concave resting area in the
middle is the cooling pond. Surrounding is the One of the interesting points of this project is the
seating area where the Smart Wall is installed. The multi-system optimization. Since the system is
dome shape roof can collect the hot air from the making use of the wind energy, the quality of the
human beings and exhaust to the exterior. It can wind energy is crucial for its performance. We
eliminate the accumulation of hot air nearby the have interesting investigation on how to optimize
people. The shops are ventilated by conventional the efficiency of the wind turbine in order to get
HVAC system while the resting area is by the smart as much wind energy as we can.
wall.
We know that the building design, especially the
As mentioned, the smart wall will directly connect geometry of the blocks, affects the wind
to the wind turbine through the intelligent control environment a lot. We have done simulations on
system. The hot side is cooled down the water three different building geometries (Figure 6). We
circulation from the water pond. On the cold side, also design the special feature – the wind catcher
the temperature is regulated by the thermal mass on the building edge – to enhance the wind speed
to confront the fluctuation of the wind energy passing through the wind turbine so that more
supply (Figure 5). The thermal mass is installed energy can be generated.
in front of the Smart Wall. It could be a piece of
thick concrete or stone. Since the Smart Wall can Through this exercise, we are demonstrating the
operate 24 hours with the energy supply from the process of parameter optimization by architectural
wind turbines, the thermal mass can store the cold design and CFD verification. 4 parameters are
/ heat generated from Smart Wall. By this way, optimized: 1. the pressure difference created by
first, we can avoid the use of battery, which may the wind catcher. 2. Wind catcher with the building
cause energy loss, for energy storage. Second, the edge design, including the geometry of building
temperature nearby can be maintained at and how much the wind catcher cantilevers out of
designated level. Besides, we use low speed fan the building edge. 3. Flow regime nearby the
between the thermal mass and Smart Wall to building edge. 4. Installation detail, such as the
provide cool air at daytime when the seating area size and installation location of the catcher and
are occupied, so that the thermal comfort can be turbine.

The convection coefficients on the cold and hot

sides are designed as 30 and 300
when cooling, and 20 and 30 when heating. The
experiment and simulation results of the
temperature in the prototype are indicated by solid
and dotted curves respectively, as shown in Figure
9.
From Figure 9, it is seen that in the experiment,

when Smart Wall is working in cooling mode
without heat load, it can achieve the lowest
temperature Ts = 16 °C in about 6 minutes, 8
°C lower than the ambient temperature. Under
the three cases above, COP from simulation results
can be given as 0.3, 0.42 and 0.56 respectively,
which are not as good as expected. Because the
heat insulation is not good and temperature
distribution in the prototype is not uniform, the
experimental COP values can’t be provided
precisely, but can be approximated as 0.2, 0.357
and 0.47, based on the measurement and
calculation.
Figure 6. Models for simulation of wind environment
surrounding different building geometries (left); simulation Comparing the experiment and simulation results,
results of wind environment (velocity vector) (right) the error is significant. These may be caused by
several factors. First, in the experiment, the
Scale Model for Design and System temperature of the input water is 28oC, 4oC higher
Installation Verification than the ambient temperature, and hence, the heat
dissipation ability is not very good. As a result,
The Design of Prototype the convection coefficient is lower. Second, the
force air comes from several small fans with the
In order to validate Smart Wall concept, a small total energy of 24 W on the cold side, and hence,
prototype is built, as shown in Figure 8. Note that the cool dissipation ability is not very good. As a
Smart Wall does not need to cool / heat the entire result, the convection coefficient is actually lower.
space. Instead, it only needs to provide regional And third, the prototype is made of wood, and
cooling / heating at where the cool / heat is needed. hasn’t been well insulated. Hence, a large amount
The prototype model is made of wood, and the of heat exchange exists between the prototype and
dimension is approximate 1.8 m x 1.2 m x 0.4 m. ambient, which is estimated as 10×(Ta x Ts) W.
The heat load is provided by six light bulbs with
maximum power of 40W ´ 6 = 240 W. The control Findings and Challenges
system is developed by using LabViewÒ with a
sampling rate of 1 second. Findings
Experiment & simulation results and discussions After we completed the two design experiments
in badminton court and shopping mall, and also
A number of experiments are performed under the simulation for verifying the design options,
various conditions to investigate the temperature some findings can be drawn:
distributions in the prototype. During the
experiments, the temperatures are measured using Integration method with different disciplines. The
thermo couplers placed near Smart Wall, and the Smart Wall is a collaborative project with the
test results are displayed on a monitor for real time Department of Automation and Computer Aided
tracking and analysis. Engineering. With the innovative design thinking,

Figure 7. CFD simulation, velocity contour at the building edge (left); design of wind catcher to optimize the operation of the
wind turbine (right)
the semi-conductor can be integrated into the Challenges

architectural design.
a) Computational simulation is powerful in the Although the design process is very interesting
design integration process. It helps to predict and full of joy, we still face a few challenges:
the performance. Design optimization can be a) The way to improve the performance of the
achieved after rounds of simulation and design total system. The individual system needs to be
modification. integrated better to eliminate the efficiency loss
b) The Smart Wall aims to provide regional at the interface. For example, the water
cooling instead of space cooling due to the circulation and the thermoelectric module
limitation of the cooling capacity of the integration determine how efficient the heat is
material. Smart Wall can complement to the dissipated.
HVAC system b) The performance of the material. The wind
c) The low COP of the semi-conductor is a limit catcher, the water circulation system, energy
for the project. But by improving the COP of storage media and the thermoelectric module
Smart Wall, the efficiency of the system can be need to be better design to optimize the
further enhanced. performance. Also, to improve the efficiency of
Figure 8. The experimental prototype of Smart Wall

the wind turbine and solar panel is another The integration of the Smart Wall can be a driven
challenge. force to further develop and utilize the TE module.
c) The low COP of the thermoelectric module. The Smart Wall design concept is derived from
It affects the efficiency of the smart wall. It the TE. It shows a new way that how the TE can
makes the smart wall difficult to compete with be anchored to the building industry. With the new
conventional HVAC system until now. This is development direction, the TE can be further
the limitation from the production industry. improved. This will create a two-way development
between the TE industry and the Smart Wall. Both
sides can be undoubtedly benefited.
The Smart Wall is one of the new technologies.

The Smart Wall project demonstrates the thinking
of how to integrate the “new” material and
technology in the architectural design by
collaborating with different disciplines to improve
the quality of the built environment, and at the
same time, to raise the potential of other existing
technologies. The concept of the project is
expected to provoke other new thinking of
utilization of new material and technology in the
building industry.
Figure 9. Experiment and simulation results of temperature:

(1) Cooling without heat load; (2) Cooling with heat load;
Acknowledgement
(3) Heating without heat load.
The work described in this paper was partially
Conclusion and Future Work supported by a grant from the Research Grants
Council of the Hong Kong Special Administrative
The Smart Wall project is expected to provoke new Region. (Project no. CUHK4199/02H)
design thinking in the building industry. First is
to rethink about how the new building form can
facilitate the operation of wind turbine to collect
the wind energy. Second is to use grey water
system (the water circulation system) for cooling
of the Smart Wall. Third is to reduce the use of
conventional HVAC system in the public area such
as the lift lobby in housing so that the energy
consumption can be cut down.
Moreover, the smart wall can be an multi-

functional element in design. It can be a
complementary to the entire HVAC system to
provide thermal comfort. Also, we can integrate
the smart wall to the architectural design. It could
be a decorative feature such as a sculpture.
Besides, we can investigate the possibility to
integrate it into the structural system design. In
the future, the smart wall can be really a “wall” in
the house. With the intelligent control system, it
has the potential to integrate into the intelligent
house. Through the sensor and control system, it
can adjust the performance so as to satisfy the
thermal comfort of the occupants. That will be the
real “Smart” Wall.

References:
Y. J. Dai, R. Z. Wang, & L. Ni (2003).

Experimental Investigation and Analysis On A
Thermoelectric Refrigerator Driven By Solar
Cells, Solar Energy Materials & Solar Cells,
77(4), 377-391.
Lertsatitthanakorn, C., Hirunlabh, J., Khedari, J.

& Scherrer, H. (2001). Cooling Performance Of
Free Convected Thermoelectric Air Conditioner,
Proc. 20th International Conf. on Thermoelectrics,
(pp. 453-457), Beijing, P. R. China: IEEE Press.
Joel H. Fergiger & Milovan Peric (2002).

Computational Methods For Fluid Dynamics. (3rd
Edition). Berlin , Germany: Springer.

Chapter 3
Applications
Gregory A. Luhan
Modern Translations, Contemporary Methods: DL-1_Resonance House®
Nimish Biloria, Kas Oosterhuis, Cas Aalbers

Design Informatics
Jordan Brandt
Skin That Fits: Designing and constructing cladding systems with as-built structural data
Madalina Wierzbicki-Neagu
Unfolding Architecture – Study, Development and Application of New Kinetic Structure
Topologies
Bob Sheil, Chris Leung

‘Kielder Probes’ – Bespoke Tools for an Indeterminate Design Process

G. Luhan / Modern Translations, Contemporary Methods: DL-1_Resonance House®
Modern Translations, Contemporary Methods:

DL-1_Resonance House®
Gregory A. Luhan1
1
University of Kentucky
Abstract
As the first design-build-fabricate-assemble experiment at our school, the intent of the studio was to design a
framework from which to examine a “lived space” through digital-to-digital processes. Moving from digital
models and physical stereo lithographic models to hand-fabrication and digital assembly allowed the students
to move from creation to completion. As part of our holistic design process, the studio fabricated almost all
components for the project. These elements include the wood flooring, the copper and wood skins, the building’s
structural panels, and the two-story light vortex.
This single-family, in-fill house is located within an historic downtown neighborhood and is subject to historic
district zoning regulations, design guidelines, and Board of Architecture Review approvals. The project is
analogous to design challenges presenting themselves in historic districts throughout the United States including
the Savannah, Georgia site for the 2005 ACADIA Conference. The scale of the project relates well to the
horizontal nature of this context and after a formal, televised review process with the local Board of Architecture
Review, the project represents a dynamic, yet sympathetic architectural dialogue with the surrounding buildings.
The project develops simultaneously from the exterior and interior resulting in two courtyards that mediate
the urban “front door” and the private “terrace.” The students designed these areas through a series of two-
dimensional axonometric drawings, three-dimensional physical and digital models, and four-dimensional
time-based animations. The building massing separates into two core elements: gabled copper volume and
wood screen volume. These elements maintain their conceptual purity by using the same types of modulations
on their skins. The copper form with its deep-cut reveals and proportionally placed light scoring patterns
reflects the horizontal datum lines of the floor, sill, threshold, and ceiling. In contrast, the wood volume
reflects these same lines as applied “shadow screens” which create depths that seamlessly tie together the
side, rear, and front facades.
The hinge point of the house is the light vortex. Designed in Rhino, translated in Catia, fabricated out of
aluminum, and clad in stainless steel, this two-story sculptural element will literally wrap light around its
surfaces. Like a sunflower, the light vortex, with its angel hair stainless steel finish, responds to the incremental
differentiation of light throughout the day. Photosensitive floor-mounted lights designed to augment the volume
of natural light will provide a continuous light rendition on the sculpture. The project, scheduled for completion
at the end of the 2005 summer session, is at the time of this submission about 60% complete.

In the last ten years, the profession has embraced

explorations in perception linked to the emerging
architecture associated with digital technology. As
architects trained to synthesize a wide array of
constraints and facts, when working with the
digital we must work at the conceptual level where
a framework exists. In order for new forms of
practice to emerge, we must unite educated
knowledge with trained skill. With the great
reduction in cost of software such as Rhino, Catia,
and Lamina as well as the increasing integration
of BIM (Building Information Management)
software, it is no longer necessary to translate
three-dimensional digital models back into two-
dimensional drawings to produce “conventional”
working drawings. These types of software allow
a very direct synthesis to occur between the Figure 1. Synthetic Landscapes Studio, Publication Cover
original idea and its formal application with a high
level of precision and accuracy. The sites for
digital architecture are synthetic, ultimately
providing the architect a place to improvise within
spatial and temporal limits. I posit, that by
augmenting the traditional ways of knowing with
the more contemporary and technologically
advanced ways of seeing, architects will
undoubtedly develop the capacity to visualize and
understand spatial relationships that would
otherwise be too time-consuming to discover. The
digital design-build studio illustrated in this paper
brings this digital intersection into the field of
construction.
Software, technology, and industry are leading and

shaping the architecture profession as never
before. In response, I established an all-digital
design studio in 1999 at The University of
Kentucky (then the College of Architecture, now
the College of Design-School of Architecture), to
Figure 2. Synthetic Landscapes Installation at Founder’s
combine the strong tradition of handcrafting in the Square in Louisville, Kentucky. Architecture’s In-between
existing design program with those Appearance and Reality
technologically sophisticated tools shaping the
profession for the 21st century. Our primary goal
was to afford the students a greater flexibility in
design, and eventually as future practitioners, In creating this digital studio at the University of
develop innovative solutions for design efficiency, Kentucky, I have forged relationships throughout
affordability, and sustainability. The materials the College of Design, the College of Engineering,
presented here reflect a logical next step of a the Center for Robotics and Manufacturing, and
sequence of comprehensive digital projects the Center for Visualization. Within these
produced under my direction from 1999 through interdisciplinary connections faculty from a wide-
2004. These projects varied in scale, site, context, range of disciplines intersect to develop and further
and material and represented a broad spectrum of research initiatives. It is from this collaborative
design execution, from experimental installations context that this project, DL-1_Resonance
to permanent inhabitations. House®, emerged.

architectural response to public and private lived

Fall 2004 | DL-1 Resonance House: Infill
space that was bridged by light.” This mission
Designs for Historic Neighborhoods
statement provided a generative discourse that
produced a set of design development drawings
for a 2,800-sqft house, a series of digital and
physical models, and formal presentations at
design reviews attended by local practitioners and
the local Board of Architectural Review. The
studio separated into three distinct, yet intertwined
components: design and detailing, fabrication and
assembly, and documentation.
Figure 5. From Manual to Digital – Carving Foundation,

Figure 3. Virtual Arrival and Departure Points within the Publication Cover
installation (Student design by Lisa Neidhardt)
Figure 6. Hand-carving Alabaster Workshop with digital re-

creations and visualizations used to translate the actual stone
Figure 4. Spaces for Rave within the installation (Student through point cloud data into digital models. The final digital
design by Aaron Anderson) outcome was a sequence of stereolithographic models used
as mock-ups for the design of full-scale structural bases for
exhibiti
The goal of this fifteen-week design studio project
was to generate a framework for fabricating infill-
housing solutions in historic neighborhoods. This As the first digital-to-digital design-build-
collaborative studio comprised of historic fabricate-assemble experiment at our school, the
preservation, architecture, interior design, and intent of the fall studio was to examine a “lived
engineering students proposed, “A volumetric space” through digital-to-digital processes.

manufacturer discussions, and hand-sketches were

actualized.
Introduction
This project is a work-in-progress that offers

students from architecture, historic preservation,
interior design, and engineering the chance to
explore a variety of digital software media and, in
a collaborative framework, address issues of
design, fabrication, and full-scale assembly.
Reflecting the realities of the profession today, this
real world, single-family house venture actively
integrates advances in software into the design
process, allowing students to move beyond basic
representation and documentation of the design
concept to an advanced analysis and understanding
of fabrication concerns.
Figure 7. The Deep-Time Probe, Investigations into Light
Architecture. A full-scale installation and digital laser The studio sought expertise for these projects from
projection event
the professional design community and forged ties
with local and national industries. Practitioners
were involved not only in the formal jury reviews
but dynamically integrated into the monthly design
charrettes. Students developed their designs in the
context of a professional design team, which
included structural engineers, fabricators,
contractors, architects, theorists, code enforcement
officers, and civic officials as well as members of
local governing organizations, including our local
architectural review board. Students displayed
their work at a variety of exhibitions and at
national and international design and technology
Figure 8. Digital-to-Digital Studio, Publication Cover conferences to contribute to and learn about
emerging ideas in digital design and to witness
Moving from digital models and physical stereo firsthand how design affects communities.
lithographic models to hand-fabrication and digital Participating practitioners also learned potential
assembly enabled actualization of the project. The uses for a variety of materials and gained greater
students created a series of flexible, open-wall insight into how increased and innovative uses of
structural panels, which by the end of the spring technology could transform their own designs and
semester were ready for placement on the site, careers.
tilting into position, and assembly. As part of the
holistic design process, the studio fabricated While one emphasis of this all-digital studio was
almost all components for the project. These to expand the students’ creative and visualization
elements include the wood flooring, the copper skills, another was to develop technical knowledge
and wood skins, the building’s structural panels, through allied seminars and elective classes that
and the two-story light vortex. To understand centered on the fabrication process. Ultimately,
better, how the pieces of the construction puzzle the goal of this course of study was to develop a
would fall into place, the students produced a comprehensive design/build opportunity that
sequential animation of the construction assembly would demonstrate that contemporary design
and a ½”=1’-0” scaled, laser cut, wood structural could not only enhance the historic context but
model. For many of the students it was the first also create a dynamic resonance between the
time that their incremental digital drawings, historic fabric and the methods of building today.

Figure 10. 1/2"=1'-0" Scale Structural Mode
Figure 11. Design/Review Charrette with local practitioners

and industry experts, 02.18.2005
Figure 12. Historic Analysis and Growth Patterns of the

Figure 9. ½”=1’-0” Scale Structural Model as Exploded Western Suburb Historic District
Axonometric

teams found a common ground in November 2004

after they attended the 2004 ACADIA National
Conference in Toronto, ON. The project that
emerged by the fall semester’s end touted a design
mission statement of “a market-driven,
architectural response to public and private lived
space bridged by light.” The students presented
this common proposal to the Lexington-Fayette
County Urban Government (LFUCG) Board of
Architectural Review (BOAR) for schematic
design review and then to the College of Design’s
all-school review. The lessons learned from these
external reviews shaped the design proposal.
Site
The location of this single-family, in-fill house is

within a historic downtown neighborhood and is
subject to historic district zoning, design
guidelines, and Board of Architecture Review
(BOAR) approvals. The project is analogous to
design challenges presenting themselves in
historic districts throughout the United States
including the Savannah, Georgia site for the 2005
ACADIA Conference. The scale of the project
relates well to the horizontal nature of this context
and after a review process with the local BOAR,
it represents a dynamic, yet sympathetic
architectural dialogue with the surrounding
buildings.
Design Process
The process of moving from digital-to-physical

Figure 13. Photograph of 1/8”=1’-0” Scale Context Model models to stereo lithographic models to hand-
fabrication and digital assembly was important to
our studio investigation because it enabled the
students’ opportunities to integrate quickly
Schematic Design different design solutions. The necessity for this
digital-to-digital (D2D) approach stemmed from
In the fall 2004 semester, a comprehensive College the narrowness of the existing site, the site’s
of Design studio encompassing students from the immediate adjacency to surrounding historic
Schools of Architecture and Interior Design and structures, and the temporal weather patterns that
the Department of Historic Preservation often plague construction schedules in the early
collaborated on various market-responsive Kentucky springtime when the house was due to
proposals for an in-fill site within the downtown begin construction. The strategy that emerged from
Lexington area. After resolving a sequence of the D2D process series of assemblies that could
complex issues that ranged from understanding be pre-constructed in the shop, temporarily stored
local building codes to implementing historic off site, easily transported to the site, and then tilted
overlay guidelines, student teams formulated into place as needed.
designs for presentation that offered a proof-of-
concept, real world construction experience for
the students. The competitive aspects of the twelve
students and the divergent nature of the design

orientation now responds in part to two forces-

the shift of the surrounding buildings to the railcars
that served the Western Suburb in the mid-1800s
and the prevalent solar angles in the region. While
equally important, the vitality proposed by these
opportunities affected the building design in
uniquely different ways. Like the storefronts that
line the street, the rotation of the building opens
the front façade to oncoming traffic. In fact, our
building, anchors one of the Kodak moments along
the historic streetscape at the intersection of Old
Georgetown Street and Ballard Avenue. Likewise,
the building’s placement exposed the building to
a wider degree angle of southern exposure, which
in turn allowed the students to examine the uses
of applied shading devices and passive green
systems for blocking the sun’s rays during the
cooling months while allowing the sun to pass
Figure 14. Images of Student Presentation at the Lexington- through during the typical heating months.
Fayette Urban County
Over the ensuing weeks, a more formal design
development package that responded to the
reviewer’s comments began to take shape,
including the building massing, the design of the
facades, and the selection of building materials.
The building massing separates into two core
elements: a gabled copper volume and a wood
screen volume. These elements maintain their
conceptual purity by using the same types of
modulations on their skins. Reflecting the
horizontal datum lines of the floor, sill, threshold,
and ceiling as deeply cut reveals and
proportionally placed light scoring patterns, the
copper form responds to both material limitation
and spatial configuration. In contrast, the wood
volume reflects these same lines as applied
“shadow screens” which create depths that
seamlessly tie together the side, rear, and front
facades. The project simultaneously developed
from the inside toward the outside and from the
exterior through the interior, resulting in a design
proposal with two courtyards that mediate the
urban “front door” and the private “terrace.” The
students shaped these areas through a series of 2-
dimensional axonometric, 3-dimensional physical
Figure 15. Exploded Axonometric of Building Components and digital models, and 4-dimensional time-based
animations. At the beginning of the spring
semester, the studio presented their complete
A thorough analysis of the historic patterns of design development package work for final
development in the area proved that the current approval. The LFUCG televised this formal
in-fill zoning guidelines were incompatible with critique and broadcasted it locally as a matter of
the immediate context. The studio presented these public record. The BOAR unanimously approved
findings to the BOAR, which in turn, supported the scheme and issued the required Certificate of
our contextual assessment. The building’s Appropriateness (COA) on January 18, 2005.

Spring 2005, Design Development: Permitting- of Architecture. In this class, the methods of
Fabrication-Assembly making and realizing the project were pre-
determined from our initial research. As a
Upon COA issuance, the structure, DL- collective team, the studio completed a design and
1_Resonance House®, began to develop. The fabrication schedule, and (using the BIM software
studio focus shifted from conceptual design and program REVIT) produced a comprehensive set
design development to permitting, constructability, of working drawings and digital visualization
and fabrication. Early in the semester, we wrote a models within the first five weeks of the spring
legal easement between the adjacent properties semester.
gaining off-street parking for both 147 Old
Georgetown Street and our site at 151 Old Our original BIM models were further engineered
Georgetown Street. To facilitate the project, the into wood and steel fabrication shop drawings after
studio also enrolled in a construction methods consultation with Peyman Jahed, Ethan Buell, and
course entitled Building Systems Integration Gyles Winkler of Buell, Fryer, and McReynolds
taught by Professor Bruce Swetnam in the School Structural Engineers. Using these documents as a
Figure 16. Studio Visualizations, View From Rear Courtyard (top); View From Front
Along Old Georgetown Street (bottom)

set of shop tickets, the students digitally the site and poured the concrete foundation walls
dimensioned, cut, and assembled standard and slab, while another group of students was
components into open-wall wood structural panels sourcing the next set of building materials.
under the supervision of Carroll Fackler, the
Director of the University of Kentucky, College
of Agriculture, at the Department of Forestry’s
Wood Utilization Center. From our digital models,
the local steel fabricator, Fab Steel, fabricated the
required bearing plates, columns, angles, support
clips, and the structural tubes and beams for our
curtain wall.
Figure 18. Images of CNC Milling scribing the Wall

Templates and Views from the Wall Panel Fabrication
Figure 17. Panel Fabrication Shop Tickets for the First Floor
and Second Floor Structure
In our initial planning schedule, I had allotted two

weeks to complete the 50+ panels, but once the
floor templates were established and cut sheets
produced, the assembly took only three days to
complete. The students transported the results of
these efforts to a staging warehouse located less
than one mile from our site. As part of our holistic
design process, the studio fabricated almost all
components for the project. These elements
include the wood flooring (four days) and the Figure 19. Roof Truss Fabrication
wood skins (four days). While one group of
students was fabricating these materials,
professional contractors surveyed and excavated

Figure 20. Photographs of Footing and Concrete Formwork
Summer 2005, Project Erection
Site Construction
Even with the extremely tight tolerances of the

digital-to-digital process, the choice to use poured-
in-place concrete rather than pre-cast concrete
foundation caused a slight delay in the erection of
our project. Fortunately, the modulation of the
building was able to absorb the discrepancies
without negatively affecting the design. To
understand better the deviations from the original
dimensions, the students digitally measured the
Figure 21. Photographs of the “Stone Slinger” and Concrete building and began developing an alternative as-
Reinforcement Mesh Inlay built digital model to track the changes.
In the six-week gap between spring and summer

Millwork and Interior Design sessions, six of the original twelve students
continued to develop the project. We treated the
The students collectively detailed the millwork foundations, corrected surface irregularities,
drawings for the kitchen, bathrooms, study, and sealed the basement walls, poured the basement
bedrooms. The details and layout of these elements slab, set the steel columns, laid the TrusJoist Silent
emerged from conversations with Ann Dickson, floor system and Parallam beams, glued, and
Director of Interior Design; Jennifer Eaton, an screwed the OSB sub floor, backfilled the
instructor in the School of Interior Design; and excavation, and rough graded the site.
local practitioners Brooks Meador (Brooks
Meador Interior Design, Inc.) and Carey Spalding On June 9, 2005, the summer session began in
(BC Woodworking). Accompanying these detailed full, with the tilting of the first floor wall panels
drawings, were another set of cut sheets that and framing of the interior walls. By June 15, the
allowed the students to pre-manufacture the boxes, second floor sub floor was in place in the rear
doors, and pre-drill the hinges of all of the volume and framing the stair tower was underway.
cabinetry. This process started on April 11, 2005 By June 17, the “bridge between the two volumes”
and by the end of the spring semester three weeks was set and work on the front volume’s second
later, the students had constructed 75% of the floor was beginning. We anticipate that the finish
millwork and 80% of the structure for the house. framing and that the enclosed black box will be

Figure 22. Photographs of Concrete Slab Pour Prior to Finish
complete within two weeks. As of this writing,

the framing continues and the students are “Metal is the material of our time. It enables
preparing fabrication documents for the building’s architecture to become sculptural; it also expresses
copper skins for a local fabricator. Local technological possibility as well as the time-
professional electrical, mechanical, and plumbing honored characteristics of quality and
contractors will complete work during late June permanence.” Frank O. Gehry
and early July. As these contractors work inside,
we will finish fabricating the interior millwork, In September 2004, my colleague, Robert Ogle
detailing the material finishes, installing the from the Center for Historic Architecture and
redwood and copper siding, and fabricating the Preservation (CHAP) at the University of
hinge point of the house - the “light vortex.” Kentucky, and I formed Design Lab, Inc®. Design
Light Vortex Lab is a private-not-for-profit company whose
Figure 23. Millwork Shop Tickets for Kitchen Cabinetry

mission statement is to “enhance our built

environment through research and education” by
enabling design/build projects, such as the DL-
1_Resonance House® for the University of
Kentucky and other schools of architecture. To
elevate the digital-to-digital processes that I have
been developing over the last seven years and to
expand our links to industry beyond the region
into the national level, we visited the fabrication
plant for the A. Zahner Company, who has
fabricated a majority of Frank Gehry’s work as
well as Steven Holl’s Turbulence House among
others. I wrote a proposal for them to assist us in
fabricating, engineering, and installing of a portion
of our project, a 24-foot tall sculptural light
element, and the light vortex. A small section of
this sculptural element was on exhibit during the
2005 ACSA National Conference in Chicago and
then installed at the all-College of Design
exhibition in Lexington.
Designed in FormZ and Rhino, then translated into

Catia before fabrication, this two-story element,
structured with aluminum fins and substructure
and then clad in stainless steel, literally wraps light
around its surfaces. Like a sunflower, the light
vortex, with its angel hair stainless steel finish,
responds to the incremental differentiation of light
throughout the day. Designed in collaboration with
the Cuban lighting designer, Joe Reybarreau,
photosensitive floor mounted light fixtures
augment the volume of natural light so that its skin
will appear to change colors over the course of
the day. In addition to lighting, the form creates a
figural focal point enclosure that contains the
fireplace box, and acts as a mechanical plenum
that conceals all of the mechanical ducts and
electrical conduits that rise to the second floor.
Due to its placement, this object anchors the entry
sequence and is visible in a variety of angles both
inside and outside of the house.
Opposite the vortex, a two-story curtain wall

marks the edges of the overall light void. In plan,
the L-shaped curtain wall inversely relates to the
adjacent single-story house on the south end of
the site, while in section, it provides striated visual
access from the second floor. A terraced area
beyond this glazed wall extends the axis of the
kitchen and dining room beyond the living room
out toward the rear of the site. A water-jet cut
copper screen wall at the property boundary
terminates this alignment. Figure 24. Photographs of Basement and First Floor
Framing

Figure 25. Design and Fabrication Process of the Light Vortex with the A. Zahner Company (Kansas City)
Conclusion between students, industry, and the profession has

allowed the College of Design to provide
Over a six-year period, this all-digital design leadership for practicing architects, to create a
studio has developed from a pedagogical model dialogue between industrial and design
for developing new ways of seeing and making professionals, and to integrate design pedagogy
architecture to a “proof-of-concept” experience with the technological applications that will shape
that blends state-of-the-art visualization the future of architecture practice. The resulting
techniques with contemporary expectations of effect has been a series of virtual studies with real-
practice and construction. Creating these links world applications and an increased role for

students in shaping a new reality for practice

through advances in technology and industry.
Much strength associated with this multi-

disciplinary partnership has enabled a rethinking
of the traditional teaching of design studio so that
it can align with the pursuit of research, funding
opportunities, and a deeper understanding of the
market place that would not have otherwise been
possible. Further, this relationship has celebrated Figure 26. Design Lab, Inc. Logo
the dynamic relationship between visualization,
representation, and fabrication-both by hand and
digital making without compromising either one.
An innovative partnership between Design Lab,
Inc®, a private not-for-profit company, and the
University of Kentucky facilitated the DL-
1_Resonance House® design build studio.
Founded to facilitate design build projects for
schools of design and historic preservation Design
Lab, Inc.’s mission to enhance the built
environment through research and education
allows for a unique approach for students to work
on real world design, fabrication, and construction
solutions while receiving academic credit. Design
Lab’s professional volunteers manage all of the
business end of the transaction with the client and
help source financing for the project through
corporate sponsors and other donations. Design
Lab also provides grants to their partner schools
and scholarships to students of design and
preservation. The creation of Design Lab, Inc®
and its affiliation agreement with the University
of Kentucky, has given the students invaluable real
world learning experience that narrows the gap
between the design drawing process and
actualization methods. It has allowed architecture,
interior design, and historic preservation students
to form sets of teams with a wide range of
practitioners, industry-specialists, and researchers
(inside and outside of academia) and it has
afforded them access to technologies that we
currently do not have within our program. This
process will continue to develop as this project
nears completion at the end of the 2005 summer
session in mid-August.

N. Biloria, K. Oosterhuis, C. Aalbers / Design Informatics
Design Informatics
Nimish Biloria1, Kas Oosterhuis2, Cas Aalbers3

1
Technical University of Delft
2
Technical University of Delft + ONL, Rotterdam
3
ONL, Rotterdam
Abstract
The research paper exemplifies a novel information integrated design technique developed at ONL (Oosterhuis
and Lenard), Netherlands, specifically appropriated for envisaging complex geometric forms. The ‘informed
design technique’, apart from being highly instrumental in conceptualizing and generating the geometric
component constituting architectural form in a parametric manner, is also efficiently utilized for precise
computer aided manufacturing and construction of the speculated form. Geometric complexities inherent in
contemporary architectural constructs and the time spent in appropriation of such topologies, fueled the
‘informed design’ approach, which caters to issues of timely construction, precision oriented design and
production (visual and material) and parametric modeling attuned to budgetary fluctuations. This design-
research approach has been tested and deployed by ONL, for conceiving ‘the Acoustic Barrier’ project, Utrecht
Leidsche Rijn in the Netherlands and is treated as a generic case for exemplifying the ‘informed design’
technique in this research paper. The design methodology encourages visualizing architectural substantiations
from a systems perspective and envisages upon a rule based adaptive systems approach involving extrapolation
of contextual dynamics/ground data in terms of logical ‘rules’. These rules/conditionalities form the basis for
spawning parametric logistics to be mapped upon geometric counterparts exemplifying the conception. The
simulated parametric relations bind dimensional aspects (length, width, height etc.) of the geometric construct
in a relational manner, eventually culminating in a 3D spatial envelope. This evolved envelope is subsequently
intersected with a ‘parametric spatio-constructive grid’, creating specific intersecting points between the two.
The hence extorted ‘point cloud’ configuration serves as a generic information field concerning highly specific
coordinates, parameters and values for each individual point/constructive node it embodies. The relations
between these points are directly linked with precise displacements of structural profiles and related scaling
factors of cladding materials. Parallel to this object oriented modeling approach, a detailed database (soft/
information component) is also maintained to administer the relations between the obtained points. To be able
to derive constructible structural and cladding components from the point cloud configuration customized
Scripts (combination of Lisp and Max scripts) process the point cloud database. The programmed script-
routines, iteratively run calculations to generate steel-wireframes, steel lattice-structure and cladding panels
along with their dimensions and execution drawing data. Optimization-routines are also programmed to make
rectifications and small adjustments in the calculated data. This precise information is further communicated
with CNC milling machines to manifest complex sectional profiles formulating the construct hence enabling
timely and effective construction of the conceptualized form.

ONL and the notion of Multi-disciplinarity merger (of design ideologies and multiple
disciplines) for the case of the ‘Acoustic Barrier’
ONL, a multidisciplinary office directed by Prof.ir. project (Figure 1) developed at ONL.
Kas Oosterhuis and visual artist Ilona Lenard
performs as a design-research body driven by Form Finding
contemporary Information communication
technologies, focusing upon issues of ONL in its attempt to decipher ‘form’ dwells into
collaborative design in a media (digital and diverse processes of utilizing digital to analogue
electronic) augmented spatial environment. The means of mapping contextual dynamics onto
notion of visualizing a context embedded design generic geometrical compositions. The stylization
solution, at ONL is conceived through building a process ranges from hand drawn curvilinear
generic connectivity between geometric styled- geometry, which is eventually digitized and
prototypes (spawned by existent spatial scenarios) parameterized, sophisticated digital simulation
articulated with parametric relations and a based generative geometry to tactile conceptual
corresponding demographic data base of their prototypes (physical), which through a process of
contextual settings. Such inclinations allow one reverse engineering are translated into the digital
to simulate emergent spatial behaviors through real realm. A relatively intuitive approximation of sets
time data exchange and a networked nature of the of curves, surfaces and masses is hence formulated
architectural grammar constituting corresponding as the conceptual root at this stage. Eventually,
physical prototypes. ONL, in order to manifest the notion of developing a generic connectivity
such an agenda, embodies a synergistic merger of between the real and the virtual is employed over
the expertise offered by architects, visual artists, these conceptual spawns by computational means.
web designers and programmers, who work These computational means range from
together and join forces, practicing the fusion of developing inherent connectivity of geometric
art, architecture and technique on a digital conceptions with database structures, visualizing
platform. The notion of fusing information initial sketches as continuous curves and surfaces
(context driven data scapes): soft component with (NURBS), developing generic relations between
the physical materiality of architecture: hard geometric components by means of parametric
component to generate a co-evolving spatiality design and deploying indigenous scripts focusing
drives the design-research ideology at ONL. The upon extraction of digital data required for direct
research paper focuses upon a design strategy: the file to factory processes.
‘informed design’ exemplifying this synergistic
Figure 1. The Acoustic barrier project (with the Cockpit/Hessing showroom): ONL

For the purpose of this research paper, we will between the NURBS curves. This relational set
specifically concentrate on the Acoustic barrier up is specifically defined (in this case) owing to
project and will hence elaborate upon the design the manner in which the acoustic barrier will be
development and form finding phase deployed in perceived by the commuting mass. A relational
visualizing the projects complex curvilinear, rule that satisfies issues of scale, surface-continuity
almost reptilian form. and smoothness (non-distracting) hence will
eventually substantiate styling, visual perception
The Acoustic Barrier: Parametric Set-Up and form generation of the construct. The barrier,
(Conceptual Resolution) a “one mile building” seen from the perspective
of the highway, (considering the above mentioned
The project based in Utrecht Leidsche Rijn in the criteria) derives its reptilian form on the basis of a
Netherlands aims to combine a 1.5 km long context driven rule: the length of the built volume
acoustic barrier with an industrial building (the of the Cockpit emerging from the acoustic barrier
cockpit/Hessing showroom) of 5000m2. The will be 10 times more than its height.
conceptual underpinning for the project is laid by
means of articulating sets of NURBS curves, This parametric relation regulates the linear form
suggestive of a relation between height, width and of the barrier to generate transversal sections,
the length of the barrier. These curves are stretched which are smoothly transformed from concave
along the 1.5 km stretch of the highway and form towards convex faceted surfaces with occasionally
the above-mentioned intuitive spatial guideline for emerging sharp longitudinal folds. This parametric
the project (Figure 2) relation once set, and mapped onto the sets of
curves yields a relatively smooth curvilinear
surface with an equally smooth transient bulge,
which houses the cockpit/Hessing showroom
space (Figure 3). This ‘informed geometry’, which
creates the three-dimensional skin for the acoustic
barrier not only operates as a ‘form generator’ but
also proves to be a ‘form regenerator’, owing to
the geometrically relational (parametric)
dependence of the generic curves. Any parametric
alteration made to the curves, consequently leads
to a regeneration/re-appropriation of form in
accordance with the context based, basic rule
(which induces the relation between the
dimensional aspects of the 3d form) hence
reflecting a new, yet controlled spatial
configuration.
Figure 2. Set of related curves defining the topology of the

Acoustic Barrier
Figure 3. Set of related curves with the parametric relation
mapped onto them resulting in the bulging topology
Subsequently, the deployment of computational
logic to the abstract sets of curves is contextually
derived with respect to the speed/flow of passing Parametric Set Up (Finer Resolution: The Point
traffic. The swarm of cars streaming at a speed of Cloud)
120 km/h along the acoustic barrier site lays the
rationale for deriving parametric rules, specifically In order to derive a finer degree of control over
linked with developing generic geometric relations the obtained (conceptual) three-dimensional form

(from the network of curves), a ‘parametric Generative and (Re) Generative Design by
structural grid’, which obtains its dimensional Scripting
logic from an optimal construction, oriented
perspective (e.g. dimensions of glass panels) is The “point-cloud” is a crucial model fostering
mapped onto the surface of the conceptual generation and [re] generation of all point-data,
construct. This intersection results in the extraction parameters and the relations between the points
of a distinct series of nodes/points, collectively (constructive nodes). However, in order to develop
called the ‘point cloud’ (Figure 4). a constructive spatial structure and to manufacture
the glazing and cladding material for the acoustic
barrier a novel application is programmed.
(Scripting and programming refer to the process
of writing a simple program in a utility language
to orchestrate behavior. It consists of a set of coded
instructions that enables the computer, to perform
a desired sequence of operations). This
application, programmed in diverse scripting
languages [MAX-script, Auto Lisp] connects to a
database system developed for handling all point-
data and their relations.
The developed scripts operate on a simple rule:

Figure 4. Point cloud generated from the conceptual 3D all points should look at and analyze their
envelope
neighbors (in terms of co-ordinates and proximity).
Such a rule-based interaction is akin to the notion
of Flocks: Flocking behavior and Boids, as stated
The point-cloud represents a parametric set-up: it by Craig Reynolds. Boids, replicated in the case
describes the volume by points and establishes of the digital model by points/constructive nodes,
spatial relationships between them: by serving as are active members of a flock, calculating their
a generic information field concerning highly position in real-time in relation to each other. Each
specific coordinates, parameters and values for Boid, locally, extends the principle incorporated
each individual point/constructive node it by Flocking mechanisms of computing a limited
embodies. The sound barrier contains set of simple rules, towards scripting the Point
approximately 7000-point objects, whose relations cloud behavior. This behavior of localized and
are administrated in a database. These relations limited computational performance by parts of an
are directly linked with precise displacements of entire system, bring about complex reactions at a
structural profiles and related scaling factors of holistic level. These simple sets of rules, can hence
cladding materials. This linkage is further be interpreted as the behavior producing genes of
extracted from the point cloud body by running the nodes (junctions in the prototype) and these
specialized ‘Scripts’ developed at ONL. These will behaviors in-turn, are directly related with the
be exemplified in detail in the next section (1.4). formal articulation of the prototype: a bottom up
approach directly inducing top down performance
Apart from creating a precision oriented geometric determination.
configuration, working with parametric models
also creates an excellent communication space for The programmed script-routines, based on such
the stakeholders in the building design process and flocking principles, when applied on the point-
enables one to discuss varied dimensions cloud iteratively run all the calculations to update:
composing the quality of the proposed space. Such
an approach also releases the design process to • Steel-wire frames with its databases
collaborative engineering opportunities during the
• Steel-lattice-structure including all the execu-
execution phase of the project and hence creates
tion drawings
an open framework for generating meaningful
interactions between clients and users. • Dimensions and Execution drawings of glass
plates.

The scripting computational component operates the distance parameter in the initial
at three levels, each component embedding within point-cloud configuration
it a series of iterative operations. Exemplification • Projection of a re-configured point-
of the three scripting levels in relation with their cloud version > point-cloud + alter-
operational performance is as follows: nated displaced points normal wise
aligned towards the orientation of the
Script 01 point (with neighboring points) is ini-
tiated through a scripted iterative pro-
Basic operation: Loads the Rhino generated cess.
.DWG files containing the point clouds > Makes • A possibility for adjusting the Thresh-
a single mesh out of them > Offsets this mesh by old value for searching for points can
the r brace value (radius of the braces conceived also be adjusted at this stage.
by the glass manufacturer that will be used for the
assembly of the glass plates) > Creates a series of
spheres cantered to the vertices of this mesh that Script 02
represent a second point-cloud to be used
exclusively for the glass plates The second Script based operation is responsible
for segmentation of the entire point–cloud body
Script 01: Overview of Input parameters: into bays of 9.33 m. This generation of segments
global fmin=01 dissects the barrier into three bays with 118 points
global fmax=44 each (Figure 05a.) and derives its logic from the
global rbrace=61 sequence in which the foundations for the
global threshold=250 construct have to be laid. This basic dissection of
allthepoints=#() the volume apart from being appropriate for
allcount=0 Physical construction also proves to be beneficial
in terms of CPU usage and data handling and hence
• Data administration phase: the operation intends to be much more efficient and performative
volves a methodological extraction of data in the long run. Each segment contains a group of
from the body of the point-cloud at three sub- points and its corresponding mesh. The meshes in
levels namely turn describing the glass plates and the amount of
• Defining ‘f min and f max values’ displacement needed by the extracted glass plates
(the range of segments to iteratively in between adjoining segments.
generate the requested data) and 4.0.3. Script 03
hence subscribing a dimensional as-
pect for limiting the administration Basic operation: Builds the axis of the steel
process to a given length. profiles that form the structure > Projects the
• Logistically naming and re-naming planar surfaces generated between the points,
of the points (to be administered) in defining shape and position of the glass panels.
the point cloud
• Formulating an Array wise database The third script operates at two levels > Generating
of the points (based on X, Y, Z co- steel construction (Figure 6) elements and
ordinate orientation) Generating Glass plate elements (Figure 5b.)
• Mesh generation phase: after the data admin- • The script geometrically generates steel
istration phase, the script generates a mesh, construction elements in a wire-frame
where each face in the mesh embodies a face mode and exports the file in a specified
of a glass plate including the Scaling of the protocol format for the production
glass plate. process (to be communicated to steel
cutting machines). This protocol, set up
• Scaling for the glass plates is defined
by ONL and Meijers Staalbouw (the steel
by the ‘r brace value’ (radius of
manufacturing company), presents the
braces/scaling distance for fixing the
obtained data-file in several layers, colors
glass plates: provided by the glass
and named elements. The layers are set
manufacturer). This value re-defines
up for different purposes, for example,

describing the steel-profiles in a • Drawing 3d splines representing the axis of

hierarchical fashion: horizontals, tubes, the steel profiles (structural) and the contours
diagonal profiles scaled or non scaled, of the glass plates
highway side elements, industry side • Unwrapping the glass surface and placing it
elements, etc. on a horizontal plane as separate triangulated
• Besides the steel sections, the generation elements
of glass plates also happens in a similar • Naming each element with quotations
manner but these are eventually flattened
down on an X, Y plane for control and
• Assigning Layer numbers to each element
visual judgment purposes and to enable according to the sequence of construction
one to check the script wise generation • Saving these elements (according to the
of data. specified protocol) in separate files segment
• This geometrically generated data for the after segment:
glass plates is further exported, arrayed, • One .dwg file for the steel construction
logically named and positioned in a (to be exported to the steel company)
complex Excel data sheet, which is • One .dwg file for the glass plate
directly utilized by the glass manufacturer
manufacturer for precision based • One .txt file for the glass plate
production purposes. manufacturer containing entries with
• This sheet is also used for further data essential data for every glass plate
manipulation, like optimization or individually (for glass manufacturer)
correction routines. • Two separate files containing all the pairs
of plates that form an angle higher than
10 degrees (for glass manufacturer)
Script operation overview • One .max file containing the complete
segment of 18 m length and 118 points
The three scripts combined together present a
• Splitting of the last generated point cloud
methodological approach towards efficient
segments (script 1, generates segments of
translation of conceptual form to precise geometric
18m length constituting 118 points) into two
and information rich entity. This sequential
9 m segments constituting 59 points (required
translation after offsetting the Rhino based point
by the steel manufacturer bearing in mind the
cloud (through script 1) through scripting can be
foundation stages) and saving them as
listed down in the following manner:
separate .max files for the steel manufacturer.
• Reading one after the other the .max files with
the double point clouds
Figure 5a. 3d wire frame model (above) displaying a segment of the acoustic barrier and the relations between the points in the
point-cloud

This comprehensive and precise data, processed The protocol developed for storing information
via the Scripting and Generative design in the database at subsequent stages of the design
components is further communicated to the process is also directly linked with the manner in
manufacturing units for computer aided which CNC machines would process the design
manufacturing purposes (CAM). data. However, as a generic outcome of the
computational processes mentioned above, one
can extract three basic strategies deployed over
architectural form to reach the production process:
• Conversion from point cloud to steel-
wire frame model and administration of
all its parameters in a database
• Conversion from steel-wire-frame model
to steel-lattice-structure and generating
execution drawings (Figure 7)
• Conversion from point cloud to glass
plate manufacturing and administration
of all dimensions, codes and specific
values plus generating execution
drawings.
Figure 5b. 3d wire frame model displaying a segment of the
acoustic barrier and the administration of all unique glass
The excel database which stores the data in a
plates [generated by script]
numeric array corresponding with the generated
execution drawings and 3d segments is bundled
together and further communicated to the
CAM techniques manufacturing units as a concise production
schema. This assists in speeding up the production
CAM strategies are dealt with in a rather coherent process and hence results in accomplishment of
fashion throughout the design and development complex projects within the specified timeline. The
stages of the acoustic barrier project. A parallel parametric design conception, filters down to the
development and maintenance of a database smallest detail (the point/construction node) and
system, which stores the script-generated data is results in the development of two generic details
seen as a generic process through this project and to mount either glass-plates or expanded steel-
can be easily deployed for a variety of complex plates towards the steel-structure. These details,
spatial topologies.
Figure 6. 3d model displaying a construction node [= point in point cloud] of the acoustic barrier and the steel profiles, steel plates
and welded joints [generated by script]

being parametric in nature, efficiently adapt to the named/numbered parts (more like a kit of parts
dimensional and orientational (towards the steel scenario) in a sequential manner to produce a
structure) variation prevalent in each mounted holistic topological marvel.
glass-plate or steel-plate and hence proves to be a
vital performative aspect, when conceiving
complex spatial topologies. The database, which
embeds these variations in numeric arrays is
subsequently communicated and executed by the
manufacturing units to produce customized details
with utmost ease and precision (Figure 8).
The efficiency and speed involved in the

production process, provides both, the architect
and the engineers to erect and test/analyze 1:1
prototypes (crucial portions of the construct) for
spatial and structural purposes at a relatively early
stage, hence deploying corrective measures and
speeding up the realization of such complex Figure 8. Assembly of unique construction nodes and vertical
projects. The assembly phase (Figure 9) is hence frame at Meijers Staalbouw Factory, from parametric 3d model
to mass customized production
reduced to an exercise of connecting precisely
Figure 7. 3d steel lattice model and its corresponding execution drawing
Figure 9. Perspective view of the assembled façade, (right) perspective displaying the parametrically generated steel structure and
glass plate cladding

Conclusion References
A design-informatics hybrid, the multi- Bossomaier, TRJ and Green, DG. (2000). Complex
disciplinary techniques exemplified in the research systems. UK: Cambridge University press.
paper, focuses upon a synergetic merger of
technology, art and architecture to efficiently Bier, H. (2004). Digital Design Strategies.
manifest much-speculated complex spatial International DESIGN Conference, Dubrovnik.
constructs. Such ‘informed-design’ techniques,
while promoting a parametric mode of operation, Capra, F. (1982). The turning point: science,
which enables one to communicate smoothly with society and the rising culture. New York, NY:
three dimensional models and the project database, Simon and Schuster.
inherently involve a collaborative design
approach, entailing derivation and appropriation C.W. Reynolds. (1987). Flocks herds and schools:
of diverse tools and techniques (programming/ A distributed behavioural model. Computer
scripting, graphic design, architecture, engineering graphics 21, SIGGRAPH’87.
and CAM) towards manifesting spatial constructs.
Oosterhuis, K. (2002). Architecture goes wild. The
The acoustic barrier (looked at as a generic Netherlands, Rotterdam: 010 publications.
example) validates the remuneration of a rather
structured manner of data exchange between Oosterhuis, K. (2003). Hyperbodies - Towards E-
geometric and text based/numeric arrays (in an motive Architecture. Basel: Birkhauser.
excel sheet) of contextual and spatial aspects and
promotes the possibility of controlling and Zellner, P. (1999). Hybrid Space, New forms in
optimizing all the points/construction nodes from digital architecture. UK, London: Thames and
a datasheet. The datasheet format also makes it Hudson Publications.
easier and faster to apply optimization and
correction routines to (by means of application of
scripts in an iterative manner directly to the data,
instead of geometrically intervening and
intuitively tweaking a 3D model), which directly
update geometrical data, and text-based
information. Working with parametric models also
creates an active communication space for the
stakeholders in the building process to discuss the
qualities of the proposed environments. It opens
up the design process for collaborative engineering
in the phase of the execution of the project and
hence promoting the design process as a
meaningful medium of interaction with the clients
and the users.


J. Brandt / Skin That Fits: Designing and Constructing Cladding Systems with As-Built Structural Data
Skin That Fits: Designing and Constructing

Cladding Systems with As-Built Structural Data
Jordan Brandt1
1
Harvard Graduate School of Design
Abstract
An awkward interface exists between the structure and skin of complex architecture. The primary structure is
typically allowed much higher tolerance ranges than that of the cladding industry, due primarily to the delicate
nature of the building envelope which, above all, must prevent water penetration and meet the aesthetic
requirements of the architect and client. As architecture has integrated advanced design and fabrication
techniques to realize increasingly complex shapes, this problem has been aggravated because of the tangency
requirements for high gloss curved finish surfaces and the larger variations found with rolled steel columns
and undulating concrete forms. To date, most innovations in this area have been focused upon mechanical
connections that can be adjusted and shimmed, thus requiring increased design engineering and on-site labor
costs for effective implementation. It would be preferable to manufacture cladding components that are properly
adjusted to the actual site conditions, negating the need to predict and accommodate potential dimensional
variation with complex connections. The research provides a model for implementing long distance laser
scanning technology to facilitate a real-time parametric BIM, herein called an Isomodel.

Introduction resides in an environment without gravity, rain,

cold, and, most onerously, less than perfect
The modern face of architecture comes humans. Until all of these variables are properly
prepackaged as curtain wall cladding units. When factored into the BIM, or until the entire building
assembled, these disparate components synergize is constructed in the confines of an
into an epidermal edifice and thus perform as a environmentally controlled and automated
complete building skin. The requirements of assembly plant, the as-built structure will never
curtain wall cladding cover a broad spectrum of equal the design intent. Therefore an entire
performance criteria: structural, environmental, language has been formalized to quantify this
and aesthetic. It is, therefore, imperative to design unavoidable and most problematic of construction
and construct skin that fits its skeleton. As this dogmas: tolerances.
paper will demonstrate, the dimensional criticality
inferred by the term ‘fit’ is driven primarily by Tolerances are, by definition, the allowable
the latter aesthetic criteria. The precise nature of dimensional limits of any material or the location
tolerances involved when detailing cladding of that material in an assembly. The range of
systems will be addressed in addition to the permissible tolerances and the precision of their
dilemmas that such tolerances, and the disparity definitions vary greatly between industries, and
thereof, have presented. The hypothesis seeks to are generally extracted from a reasonable
resolve these antagonizing construction problems consideration of quality, cost, and construction
through the use of an Isomodel. An Isomodel is a feasibility (Ballast, 1994). The construction
parametric BIM that receives real-time coordinate industry allows the largest variation of tolerances;
input from the actual building as it is constructed. from thousandths of an inch for prefabricated
Therefore it is possible to fabricate the cladding window systems to several inches for on-site
system to fit a variable structure without the costly concrete work.
need for adjustability.
This disparity has been aggravated by the
This paper presents a method to (1) parameterize increased geometric complexity of building
the BIM for metrological input and (2) implement enclosures. Reflective and curvaceous envelopes
current long range laser scanning technologies for require a high level of continuity, approaching the
obtaining real-time as-built data to support the level of sophistication employed by the automotive
fabrication, erection, and quality control of industry.1 Failure to maintain tangency can lead
cladding envelopes. Finally, the overall impact to undesirable visual results, such as stepped
of an Isomodel segues into a larger speculative surfaces. In an effort to keep pace, the steel and
environment that considers the vastly more concrete contractors, who are notorious for being
theoretical potentials of as-built data relative to out of tolerance on rectangular frame structures,
the ongoing discourse of digital tectonics. are trying to convert old mills into modern
precision plants to accommodate the complexity
Problem Background of contemporary designs. This difficult process
is reflected in the typical allowable tolerances of
The state of the art in building technology is milled steel sections which are even greater for
dominated by top-down processes wherein a radius sections. Curved beams and columns are
computer model simply feeds data to a CNC device finding an increasing use in contemporary designs,
for production and provides a theoretical but the resolution of their fabrication has not
framework for erection locations. In order for the developed accordingly. CNC cutting the webs can
architect to truly assume the role of the master produce relatively accurate steel members, but the
builder, a means to integrate construction feedback cost often precludes this option. Rolling steel
into contemporary CAD/CAM processes must be members into a radius is much cheaper, but shape
established. Building Information Modeling limitations and the standard deviation from the
(BIM) has come a long way to provide detailed target dimension is quite high. Radius tolerances,
information that allows for custom components in fact, are not even established in the AISC Steel
to be designed, fabricated and erected to fit within Design manual. In practice, they are often in the
a precision building assembly. These components range of 80mm for 8 meter lengths rolled the hard
are, however, fabricated to fit a digital design that way in a 450 mm deep section.

Whereas the language of tolerances can merely cladding is erected, and then the anchors are post-
define and quantify the problem, the dimensional adjusted to get the panels properly aligned. This
disparity must still be accommodated. When process of adjusting cladding panels after erection
connecting the envelope to the structure, it would often leads to skewed window openings which
be desirable to employ one common rigid anchor then require further shimming (NAAMM, 1960).
for all cases, however rather complex solutions
have been perpetually reinvented since the advent The means to verify the design intent of
of curtain wall cladding systems (Figure 1.) The increasingly complex building envelopes has
conundrum is such that maximum flexibility is become paramount to many architectural practices.
desired until the precise location is set, at which Long distance coordinate measuring systems have
time maximum structural capacity becomes the been applied in a limited number of recent building
critical factor. Furthermore, locating the cladding projects to more accurately position traditional
units is itself a tedious process. In typical high anchorage and to ensure the quality control of the
rise construction, which has thousands of anchors, finish surface, but have yet to be seamlessly
the anchors are roughly pre-adjusted by relative integrated into an as-built Building Information
measurements from the supporting structure. The Model to obviate such tasks. Furthermore, the
Figure 1. Typical Cladding Anchors

thrust of current research into long range laser Case Study: M.I.T. Stata Center
scanning technology is focused on a top-down
approach that seeks to provide holistic information Gehry Partners’ M.I.T Stata Center, recently
of construction progress, part locations and completed in 2004, offers a good look at the state
automated defect detection, for project of the art in designing, fabricating, and erecting
management or automating erection and complex curtain wall systems.2 As much as 25%
excavation equipment (Boukamp, 2004 & of the design time for the structural metal cladding
Witzgall, 2001). Such methods typically on the Stata Center was spent on a deterministic
concentrate on the comparative analysis of large tectonic level developing two connection systems
point clouds (with hopes that advances in object (for steel and concrete) that could adjust in three
recognition can help to automate this process) to principle directions to connect to the primary
determine areas of concern, however no course structure, but still accommodate seismic
of corrective action is discussed (Shih, 2004 & requirements and shear forces due to gravity and
Gordon, 2003). The difficulty, however, in terms more significantly, wind.3 The tolerances for the
of construction, does not wholly lie in obtaining concrete contractor were in the range of ±1",
surface data (although a system to quickly obtain however some areas were out by as much as 3"
and analyze complete surface data on a job site due to settlement. The structural steel had a similar
does not yet exist) but rather in analyzing the as- tolerance, but by the time the columns were rolled
built data relative to the design intent. If two to the specified radius and subsequently erected,
surfaces, such as the top face of a steel beam and they were wildly inaccurate. Loose tolerances
the bottom face of a concrete slab, are supposed were not allowed for external cladding, however,
to be coincident and are not, how is this disparity where environmental protection and aesthetic
quantified? There are means of analyzing surface considerations do not allow for large gaps.
data built within certain software environments
(Figure 2), however there is not a protocol for An ArcSecond Inc. Vulcan© long-range laser
communicating the distance between two surfaces, positioning system was used to locate the as-built
or a means for corrective action. anchor points and adjust the anchorage systems
in-situ. This system consists of four primary units:
Whereas the as-built location of every component two laser transmission stations, a locating wand
might be of interest to the general contractor, the with two receivers at each end, and a PDA that
cladding manufacturer is simply concerned about records the location points. The laser transmitters
the points where the cladding connects to the are set up and calibrated to the coordinate system
supporting structure, and the architect just wants of the job site, which is congruent to the digital
to ensure that the finish surface meets design model. The transmitters send a modulated infrared
intent. The requisite technological horsepower can signal to the location wand, and the 3d position is
be greatly diminished by integrating a bottom-up triangulated according to the angle and distance
approach. to the transmitters. It thus requires line of sight
operation between the locating wand and both
transmitters. Metrological devices such as this are
quite accurate, with a standard deviation of 2-10
mm over 100m, (Lindfors, 1999) but require set
up time to calibrate the two laser transmitters
(which must be successively positioned around the
job site to ensure line of sight operation) and a
human operator must physically contact the point
to be measured (Figure 3). Considering that there
were over 1200 anchors on the Stata Center, and
each anchor required at least two location points,
this operation was obviously tedious. The as-built
anchor coordinates were recorded in a separate
as-built digital model, however they were never
integrated into the active master model.
Figure 2. Surface Deviation Analysis

Figure 3. Measuring Coordinates Figure 5. Cladding Units
Figure 4. Steel Anchors
Unfortunately, many of the anchorage locations

on the steel superstructure were unreachable by
this means, therefore accurate position data could
not obtained to precisely adjust the anchorage
systems employed (Figure 4). Because the finish
surface was clad in high gloss white powder coated Figure 6. Visible Discontinuities
aluminum, the lack of tangency and curvature
continuity between the constituent cladding panels
was visually apparent (Figures 5, 6).

Integrating As-Built Structural Data The anchorage for the cladding components can
then be greatly simplified to a fixed design because
The process of designing and building cladding no further adjustment will be necessary to ensure
systems with as-built data begins with the B.I.M. a smooth, continuous surface per the design intent.
(Figure 7). This model must be parameterized in The surveying instrument can also serve as a
a rather specific way to facilitate the Isomodel, quality control tool to verify the surface location
which is later presented. Generally, the B.I.M after the cladding is erected. Other immediate
facilitates the design, fabrication, and erection of possibilities include the use of the as-built data to
the primary building elements. A long range laser analyze deviations, and determine if tolerance
scanning instrument (Ladar, Lidar, Total Station, allowances have been exceeded (Akinci, 2004).
etc.) is located on the job site, and can be used to This could provide critical data for the architect,
locate the structure as accurately as possible within contractor, client, and the sub-contractors in any
the given tolerance. The surveying instrument then pending litigation.
records point data from the structure, which are
the nodes to which the cladding panels will The Isomodel operates on two primary concepts:
eventually attach. The coordinates of these points
are uploaded to a database, from which the 1. Minimize the required physical data.
Isomodel receives its parametric input. The 2. Integrate the physical data as the
Isomodel performs as an interface between the real parametric datum in the BIM.
and virtual; all subsequent assemblies that attach
to the primary structure can be fabricated to fit Looking at the tolerance problem from an
the as-built conditions (as far as lead times might application specific domain is imperative. If the
allow.)
Figure 7. As-Built Structural Data Implementation

cladding attaches to a wide flange steel column, sphere of space, the radius of which equals the
then it is not important to have complete surface maximum allowable tolerance. Therefore a servo-
data of the column, but rather only the location of actuated Lidar system employed on the building
the connection node. With this conceptual step, sight can be programmed to automatically scan
the required physical data can theoretically be only those areas and return the surface data of the
reduced to the three coordinate variables of a connection node.
single geometric point. Such a reduction is
achieved because there is information coming With the raw surface data, the goal is then to
from both the top-down (digital to physical) as extract the geometrical center. At the Digitizing
well as bottom-up (physical to digital). The basic Laboratory at the Harvard Design School, steps
parameters for digitally modeling cladding were taken to automate this rationalization process
systems are the exterior envelope (supplied as utilizing a Minolta Vivid 700 laser scanner. The
surface geometry in the BIM) and the physical spherical portion of the anchor was painted a
connection nodes. Because the actual anchorage contrasting color in an effort to isolate only the
has location and orientation (six degrees of desired surface (Figure 8). The hypothesis was
freedom), a local coordinate system must be that the points within the point cloud surface data
established at each unique connection node and could be attributed with their associative color
any coordinate system can be reduced to three value (in this case, an 8 bit value ranging from 0-
points (whether in the form of two intersecting 255) and the points outside of the given range
lines, a vector and a plane, etc.) Therefore if the would be eliminated. The specular highlights and
as-built location point is provided in the Isomodel, shadowed areas proved to be problematic
a vector normal to the exterior surface through however, so this step was conducted manually.
that point creates one line, and a vector from the With the spherical area isolated, a NURBS surface
as-built point in the direction of gravity (i.e. the was mapped onto the raw data in order to provide
dead load of the cladding) provides another line, a smoother interpolated surface (Figure 9).
thus establishing a local coordinate system. In
the case of curved cladding systems, there could Given a smooth hemisphere, any line drawn
be multiple possible lines normal to the surface normal to the surface in the direction of negative
through the as-built point. For such occasions, a Gaussian curvature with a known radius value will
simple least distances script could determine the obtain the center point. Considering that there is
best option. The important concept here is that inevitable deviation that occurs with the physical
the process has been reduced to its theoretical data, more accuracy can be obtained by using
minimum requirements and is no longer restrained multiple points. Reference geometry is projected
to a linear flow from digital model to construction, onto the surface to establish three geodesic points,
but incorporates a true bi-directional design/build from which the normal lines are drawn to the
process. radius value. The three endpoints provide more
accuracy without a significant increase in
Of course this approach necessitates a specific yet complexity; they all inevitably lie on a common
standardized hardware design for the anchorage plane and there are only two possible connections
system to support the data acquisition and erection between each point. This ensures a non self-
process. The envisioned design resembles a ball- intersecting polygon, in this case, a triangle. A
hitch commonly found on pickup trucks, used to simple trigonometric script then determines the
tow light trailers. In the case of a steel structure, center of the triangle and thus establishes the point
this connection node is welded directly to the connection node.
supporting structure in the steel mill. Such
methods are commonly employed by the simple These points, the dynamic variables coming from
means of pulling a tape measure along the length the physical world, comprise the datum by which
of columns and beams, marking the locations the parametric hierarchy of the building skin must
(typically supplied in BIM derived drawings), and be rigorously defined. Therefore when the B.I.M
subsequently welding the anchors in place. Thus is first modeled, all geometries related to the
when the steel components are shipped to the site cladding system must be defined exclusively by
and erected, the connecting nodes are already in the connection nodes (as defined in the database)
place. The absolute coordinates of the nodes can and the virtual surface geometry. Then, when the
be predicted by the BIM to exist within a given as-built connection node data is uploaded to the

database, the cladding in the B.I.M. model will Future Research

automatically update to the actual conditions.
The method presented for obtaining as-built
connection nodes proved effective in the
laboratory, however the Lidar technology
necessary to accurately scan such small connection
nodes at ranges of 50+ meters is quite expensive.
Beyond this, any means to avoid point cloud data
would be preferable in order to reduce the post
processing requirements. For these reasons, an
alternative method is being developed. This
method employs a servo-actuated Total Station,
which is already a familiar tool on job-sites, and
it has been shown to provide the cheapest per
project and complete life-cycle costs (Lindfors,
1999). It is able to obtain accurate single point
data at long ranges and can operate autonomously.
In this scenario, location stickers that consist of
three reflector targets would be placed upon the
primary structure before erection (at least in the
case of precast concrete and steel), similar to the
process of the connection nodes. The automated
Total Station would then proceed to scan until it
registered three reflector targets, which establish
Figure 8. Raw Scan Data of connector node
the connection point and plane of each connection
node. Further research into this method will be
conducted on project sites.
Conclusions and Theoretical Remarks
Ultimately it is clear that knowing the precise

location of the as-built supporting structure, and
integrating that data into an Isomodel to produce
rigid anchorage that needs no further adjustment
can yield the following benefits:
• Reduce the complexity of designing

cladding anchors
• Reduce the complexity of fabricating
cladding anchors
• Eliminate on-site adjustment
• Increase erection speed of cladding
• Eliminate the need for window frame
shimming due to panel distortion
• Increased erection speed of steel, because
the tolerances, at least for benefit of the
cladding, can be more forgiving.
• Reduction of clearance space between
Figure 9. Smoothed Surface with normals drawn the cladding and exterior face of
structure.

More important, however, are the larger CAM techniques could be informed and perhaps
speculative questions that arise in the context of improved by the process of craft. The complexities
craft and CAD/CAM. There seems to be a of haptic feedback, as-built adjustments and
premature obsession with precision when complex material interface are becoming critical concepts
geometries are concerned, and indeed it is in realizing an increasingly articulated
undeniable that the precision of the enabling CAD/ morphological architecture.
CAM technologies have made complexity
feasible. It must be stressed, however, that a Notes
building is only precise as its most inaccurate part.
Thus the assembly of precise components requires 1. In reference to G0 (coincident), G1 (tangent), and
precise techniques, because custom assemblies are G2 (curvature) continuity standards for Class “A”
not conducive to efficient modifications on site. Automotive Surfaces.
Craft traditionally functioned as an iteration of 2. The author worked for A. Zahner Company as a
multiple processes, each carried out by successive design engineer developing the metal cladding systems
trades. This natural feedback process allowed for for the M.I.T. Stata Center.
variations to be accommodated by low level 3. It should be noted that for systems with standardized
control as the work progressed, so that the architect cladding panel types, the engineering time spent on
did not have to be concerned with every adjustable anchor design approaches the range of 35-
imperfection in construction (Brooks, 2004). As 45%.
4. This premise can be compared to Cyril Smith’s
the assemblies become more exact and complex,
the potential of low level control is diminished, assertion, “The craftsman can compensate for
so the architect must predict potential problems differences in the qualities of his material, for he
that continually arise on chaotic building sites. can adjust the precise strength and pattern of
This refers to the ‘top-down’ approach of current application of his tools to the material’s local
B.I.M methodology. The more exact the design vagaries. Conversely, the constant motion of a
is defined, the more it must be controlled. machine requires constant materials.” (Smith,
1992)
Much can be learned from the craft of the sculptor.
It might be said that the sculptor, like the architect,
never realizes the true conceptual design, and thus
the ambiguity of ‘design intent’ steps forward. For
the sculptor, it is adequate to begin with a sketch
and perhaps some rough dimensions to determine
the size of the stone from which the design will
be carved. After all, the final product is ultimately
a qualitative judgement by a lone craftsman and
there are (typically) no precision drawings or
computer models that must be followed with nano
precision. Her entire process is analogue in both
the haptic and relational sense. First, she begins
with a stone and an idea. As she begins sculpting,
she is always working in proportion to existing
carvings, in analogue, without referencing
abstracted numbers. It is a constant iterative
process that in some ways represents the natural
optimization of biological structures. Just like a
stem that eventually grows stiffer in strong winds,
a sculptor can constantly adapt to variations in the
stone and the tool; she can strike harder if the chisel
becomes dull. 4 Craft is marked by the mind
relating the purpose of the work to the motions of
the hand. As such, it currently defies codification.
We cannot yet instruct a robot to provide the depth,
originality, and richness of an artisan, but CAD/

References
Akinci, B., & Boukamp, F.(2004) Towards

Automated Defect Detection: Object-Oriented
Modeling of Construction Specifications. Xth
International Conference on Computing in Civil
and Building Engineering, Weimar: Bauhaus-
Universität
Ballast, David Kent. (1994). Handbook of

Construction Tolerances. New York, NY:
McGraw-Hill.
Brooks, Alan & Poole, Dominique. (2004).

Innovation in Architecture, London: Spon Press
Gordon, C., Boukamp, F., Huber, D., Latimer, E., Park,

K., & Akinci, B.(2003) Combining Reality Capture
Technologies for Construction Defect Detection: A Case
Study. E-Activities and Intelligent Support in Design
and the Built Environment, 9th EuropIA International
Conference (pp. 99-108), Istanbul, Turkey.
Lindfors, C., Chang, P. & Stone, W. (1999). Survey of

Construction Metrology Options for AEC Industry,
Journal of Aerospace Engineering, Volume 12, Reston,
VA: ASCE. (pp. 58-64)
National Association of Architectural Metal

Manufacturers. (1960). Metal Curtain Wall
Manual, Chicago: NAAMM.
Shih, Naai-Jung, Wu, Ming-Chang & Kunz, John.

(2004). The Inspections of As-built Construction
Records by 3d Point Clouds, CIFE Working Paper #90
Stanford, CA: Stanford University.
Witzgall, Christoph & Cheok, Geraldine. (2001).

Registering 3D Point Clouds: An Experimental
Evaluation, Gaithersburg, MD: National Institute
of Standards and Technology, Building and Fire
Research Laboratory.
Smith, Cyril Stanley. (1992) Matter Versus

Materials: A Historical View in A Search for
Structure (Cambridge, MA: MIT Press. (pp.115)

M. Wierzbicki-Neagu / Unfolding Architecture
Unfolding Architecture – Study, Development and

Application of New Kinetic Structure Topologies
Madalina Wierzbicki-Neagu1
1
University of British Columbia
Abstract
Advances in design tools and material engineering open new possibilities for architectural structures that may
respond better to the demands of the increasing density of development, better space management and lesser
environmental impact. Folding structures that provide adjustable on demand configurations can be effectively
conceptualized if appropriate interdisciplinary expertise is brought together. Kinematic chain geometries
borrowed from traditional mechanics can be developed into a variety of topologies suitable for architectural
structures. Rectilinear deformable grids can provide the functionality of expanding and collapsing as well as
the ability to be infinitely arrayed. Converging grids allow for circular arrays and fan like folding. The challenge
is to translate a two-dimensional chain concept into a three-dimensional array of interleaved frames that form
a stable structure and can bear the necessary loads. In order to complement the folding structure with the
corresponding foldable shell, the algebra of rigid folds can be adapted to develop viable geometrical concepts.
The demands of the design process needed to develop kinetic structures will expand the traditional architectural
workflow to include parametric modeling tools that are common in mechanical engineering. Folding
architectural structures require, besides traditional architectural layout development, parametric assembly
capabilities and motion analysis typical for mechanical design. Potential application development, marketing,
building code changes and effective multidisciplinary collaboration must take place for kinetic structures to
enter the architectural mainstream.

Kinetics and Architecture Kinetic Structures
The idea of configurable kinetic elements in Traditional kinematic chains used in mechanics
architecture is as old as architecture itself. Doors, may provide suitable concepts for architectural
windows, blinds, gates, draw bridges and other, structures. The prerequisite is the ability to be
similar, simple in concept devices were known for assembled into large and repeatable grids.
ages. Recent decades brought large retractable Rectilinear deformable grids can provide the
roofs over stadiums and stages, elaborate partition functionality of expanding and collapsing as well
systems in conference centers, configurable floor as ability to be infinitely arrayed. Converging grids
installations in concert halls, anti sway devices in allow for circular arrays and fan like folding, but
high rise towers and automatically adjustable they require the added challenge of developing a
columns that compensate for the ground settling. kinematic chain that is scalable within iso-angular
However, these kinetic elements have mostly a module. A simple kinematic chain, the pantograph,
limited, localized function of controlling egress can be developed into both rectilinear and
and adjusting various aspects of the exterior shell. converging versions. A variety of arraying options
The concept of a ‘foldable’ building or a further increases the range of possible solutions
reconfigurable on demand layout is not here quite as illustrated in Figures 1 and 2.
yet. The progress in design tools and material
technologies offer a potential that has not been In mechanics, kinematic chains are, most of the
fully exploited yet. This paper focuses on time, a part of much larger assembly that provides
exploring the possibilities of new design methods, anchors, supports and constraints therefore they
tools and recently available materials that can be can be easily modeled as two-dimensional devices.
applicable for the architectural workflow of Since architectural structures bear a combination
tomorrow. of static loads of construction materials and
Figure 1. Arraying of rectilinear kinematic chains
Figure 2. Arraying of converging kinematic chains

dynamic loads induced by use and elements, the Foldable structures do not reach the conceptual
two-dimensional concept chains need to be extremes of the flat pattern or the flat fold. The
transformed into a three-dimensional system of physical dimensions of components as well as
interleaved frames and off-set joints to assure load assembly offsets limit the range of transformation.
bearing capacity and static integrity. This process The purpose of the folded state for a foldable
is more difficult for converging grids as the structure is to achieve a certain practical level of
architectural structures do not scale as easily as compactness. The deployed state forms the
simplified mechanical chains. The challenge of functional configuration that satisfies the design
developing two-dimensional concepts into viable requirements. Geometrical constraints required to
folding structures is a part of developing the satisfy the full range of folding from the flat pattern
methodology of kinetic architecture. to the flat fold are most restrictive. Since they do
not apply to practical folding structures, significant
Foldable Shells freedom of possible configurations is available to
designers. The only requirement that a foldable
Kinematic chains provide good concepts for the mesh must satisfy is the condition of rigid folding
design of articulated skeletal structures. However, what means that mesh facets can not be bent or
a different method is needed to develop foldable twisted. Figure 3 illustrates the basic stages of
shelling elements that can complement kinetic folding of flat sheets and 3d shells.
structures and form a finished, functional
architectural solution. Size and shape change can
be achieved by either sliding over or folding the
selected facets of the exterior. Folding is a more
desirable method since it results in more rigid and
easier to seal assemblies.
Figure 4. Controlled folding of a flat sheet
3d meshes provide seemingly unlimited design

possibilities what may increase the difficulty of
initial conceptualizing. Therefore, basic
geometrical configuration like the eight crease
sheet fold provides a good starting point for the
design development. Figure 4 illustrates the
process of controlled folding of a flat sheet. A
controlling grid comprised of two pairs of pivoting
Figure 3. Folding of flat sheets and foldable shells points is imposed over a pattern of creases. If the
sheet is folded by bringing in the corners to
Commonly, transformations of folding are coincide with the points of the controlling grid,
associated with flat sheets and play the key role in the result is a three dimensional, fully constrained
sheet metal fabrication, packaging industry and and rigid shell. The shape of the shell can be varied
art of origami. A more generalized approach is to by adjusting the parameters of the controlling grid
consider folding as degrees of freedom available as shown in Figure 5. In a practical application
in an under constrained mesh. This mesh may be, only angle ‘a’ of the pivoting grid provides feasible
in a special case, a tessellated flat sheet but, in means of adjustment.
general, any 3d mesh can be subjected to folding.
A flat pattern and a flat fold define the extreme Non-flat folding meshes can be easily developed
states of a foldable mesh. Only a flat sheet can by constructing individual facets directly over the
form the flat pattern while the state of the flat fold controlling geometry as illustrated in Figure 6. The
can be applied also to some non-flat meshes. controlling grid is transformed into a network of

Figure 5. Adjustment of fold parameters
Figure 6. Constructing a 3d mesh
Figure 7. Range of transformation
To gain more freedom in forming a foldable mesh,

3d controlling edges by means of offsetting a
selected corner points of the controlling grid can
neutral point N off the grid’s plane to define a
be offset off the construction plane. Figure 8
construction vertex. Four sectors are formed in this
illustrates the development of offset points over a
case. They can be independently populated with
converging controlling grid. The asymmetrically
facets that follow locally relevant geometrical
lifted edges allow for shell overlapping in arrayed
confines. The stretched sector requires that the sum
configurations. The constructed shell is coupled
of the vertex angles of the facets is larger than the
to a folding frame that employs the same
angle between the controlling edges in order to
converging controlling geometry. As a result, a
avoid the singularity of a fully stretched state.
modular segment of a foldable structure is formed.
The requirement for all the facets to form a flat
The design decisions made while constructing the
pattern is irrelevant since the assumed controlling
folding mesh define its functional folding range.
geometry as well as the practical motion range
Figure 9 illustrates the deployed and folded states
were not intended to reach either fully stretched
of the example module. The compactness of the
or fully folded state. In this example, the total of
folded form can be further optimized if needed by
the vertex angles is more than 360 degrees and
adjusting the controlling grid and refining the mesh
the facets form a saddle type compound surface.
facets. More compact shape may be beneficial for
The pivoting angle ‘a’ is the parameter that
the ease of transportation if a mobile solution is
controls the folding of the finished shell. Figure 7
required.
indicates the available range of transformation.

Figure 8. Development of a foldable module
Figure 9. Details of a foldable module
Design workflow
The scope of the traditional architectural workflow assembly level and employing then algorithmic
is already undergoing expansion as motion and capacity of forward kinematics modeling. This is
actuation elements migrate gradually into building a straightforward bottom-up design strategy. The
structure designs. The design of kinetic development of a foldable mesh requires a
architectural structures will expose architects to different approach. The kinetic complexity of
the unique demands of modeling of the translation under constrained meshes makes algorithmic
geometry and simulation of the motion range. The descriptions tedious to define. However, a loosely
design tools that perform similar tasks are widely improvised set of facets that follows locally
available in mechanical engineering. In particular, applicable geometric restrictions can be easily
parametric assembly environments offer the ease appended to an already defined controlling grid
and flexibility of both forward and inverse and quickly verified by means of inverse
kinematics to model complex kinetic designs. The kinematics.
increasing importance of detailing of lightweight
and reliably performing fittings will benefit from The initial draft can be further refined through a
stress analysis tools that are readily available in sequence of design iterations. The efficiency of
these software packages. this process relies on the seamless integration of
solid parametric modeler with parametric
Figure 10 illustrates the logical phases of the assembly modeler. Once the components can be
kinetic structure design development and the modified directly within the environment of
benefits of the integrated parametric CAD kinematic simulation, the development process
environments that are readily available today. The focuses on optimization of the form without the
key CAD feature that helps with tackling the cumbersome burden of switching between
complexities of foldable structures is the flexibility different design environments. Overall, the
of switching between bottom-up and top-down concept undergoes a series of design iterations as
methods while developing the design. The it progresses through development stages. An
assumed initial concept of the controlling grid and effectively integrated design environment makes
the folding structure can easily be put together these iterations feel more like intuitive adjustments
while starting with components, progressing to the rather than a frustration of starting all over again.

Figure 10. Design workflow
As most of parametric modelers have the finite increased load capacity to weight ratio are
element analysis capacity fully integrated, the particularly suitable for kinetic structures. High
verification of the tweaked in configuration in strength fibers can be effectively integrated into
terms of safety and load performance can be final components where properties like weight,
performed within the same design environment. mechanical performance, transparency and
The design data can be shared between various aesthetic finish can be easily designed in thanks
development workflows, architectural for to advances in manufacturing processes of pull-
instance, in a variety of commonly supported 2d truding, molding, injection molding and
and 3d formats. lamination. The success of developing of this new
breed of architectural structures will depend on
Material engineering is another area that will effective interdisciplinary collaboration to bring
further expand its role in the architectural design together an unprecedented range of expert
workflow. Composite materials of significantly knowledge.

Figure 11. Adjustable shading system
Figure 12. Mobile system
Applications of Foldable Structures scenarios, Figure 11. It also offers dramatic

changes in the appearance of the space.
An adjustable shading system offers beneficial
features for outdoor exhibitions as the space can The development of quickly deployable, mobile
be easily adapted to accommodate different shelters may be useful for emergency situations

Figure 13. Applications
and harsh climatic conditions. Figure 12 References

demonstrates a foldable system that can be
deployed off dedicated, motorized platform. Calvert, J.B. (2003). Mechanism. Retrieved from
www.du.edu/~jcalvert/ on March 24, 2005.
Figure 13 illustrates mobile systems deployed in
the desert and the interior of an exhibition space. Hull, T. (1994). On the Mathematics of Flat
Origamis. Congressus Numerantium (100), 215-
Closing Remarks 224.
The development of feasible applications, Zuk, W., Clark R.H. (1970). Kinetic Architecture.
assessment of benefits and marketability must New York, NY: Van Nostrand Reinhold.
adequately support the design effort. Safety and
performance tests need to be conducted as well as
appropriate changes to building codes need to take
place. The unique features of foldable structures
may change the traditional building maintenance
and lifecycle models and offer reconfiguration as
an option to demolition. In general, configurable
structures will provide lesser environmental
impact than traditional technologies as they are
better suited for re-using, modifications and re-
location. Inherently modular, they facilitate
assembling of infinite variations from a limited
set of prefabricated components. Design of
configurable habitable spaces in highly populated
areas is another area that kinetic architecture may
resolve better than traditional solutions.

B. Sheil, C. Leung / Kielder Probes
‘Kielder Probes’ – Bespoke Tools for an

Indeterminate Design Process
Bob Sheil1, Chris Leung2
1
The Bartlett School of Architecture UCL
2
The Bartlett School of Architecture UCL
Abstract
Sixteen*(makers) are a group of practicing architects, academics, designers and makers who assemble when
key questions surrounding design, fabrication, use and adaptability in architecture emerge. Initially, the group
was formed out of a motivation to engage as designers with the physical and tactile aspects of production
without a dependency upon drawing. Now, in the post digital age, the age of digital fabrication, boundaries
between drawing and making, between the designer and the maker, have dissolved. Consequently
sixteen*(makers) work is now engaged with questions of knowledge transfer, expertise, and innovation where
modes of investigation are equally embedded within in the analogue and the digital world. This article relates
to our latest ongoing work which is due for completion in 2005/06. The work has been developed as a specific
response to the award of an architectural residency by the Art and Architecture Partnership at Kielder Park,
Northumbria, England. From the outset, it has not been a requirement of the residency that an outcome is
identified early on. In fact, as I write, the outcome remains open.
Presented with an extraordinary site and coinciding with a time of rapid change the work has begun by
exploring a design process that is adaptable, indeterminate, and informed by site conditions. In October 2003,
sixteen*(makers)1 were awarded an architecture residency by The Art and Architecture Programme at Kielder
(AAPK) of Northumbria, UK. This organization is well known for commissioning works such as the ‘Belvedere’
by Softroom and the ‘Skyspace’ by James Turrell. Coordinated by Peter Sharp, AAPK consists of a number of
large public bodies, including The Forestry Commission, Northumbrian Water and Tyndale District Council.
Together they manage a land area of 62,000 ha’s centred on the UK’s largest reservoir and surrounded on all
sides by one of Europe’s largest managed forests.
1Sixteen*(makers) were formed in 1995 by Nick Callicott and Bob Sheil, the practice now includes Phil
Ayres and Chris Leung.

Responsive Architecture and ‘Paraforms’ experience of the work, the audience’s reading is
one of equal or even greater importance than that
Allied to the research of automated making of the maker. The role of the artist or architect
processes, sixteen*(makers) are also concerned therefore has a shifted emphasis from one who
with the integration of constructs, the prescribes to one who provides, and as many
environments in which they reside, and the related artists, and nearly all architects, do not make their
audience. Such integration has the capacity to work, such works also address the emergence of
evolve buildings and building elements not only the designer/maker in future works in the public
into customised passive constructs, but such realm.
constructs within active and responsive
customised environments we call ‘Paraforms’. The Blusher Project
Paraforms are responsive composites with highly ‘Blusher’1 was composed of two systems, one a
adaptable behaviours designed to recognise set of structural components for varied
audience needs, patterns of use, and modes of configuration of the form: The Skin, the other an
communication. Paraforms can respond to these integrated system of sensors and responsive
elements by altering form and behaviour according elements: The Feathers. The first consideration for
to the acquired and specified knowledge of this experiment arose through ideas of behaviour.
embedded microprocessors. Those that encounter
Paraforms are met with something strangely ‘Blusher’ was named in recognition of the
familiar, an interactive architecture that invites common human response to an unexpected
users to shape and alter their environments. moment of contact, be it either one of pleasure or
Paraforms may demonstrate interactivity through disgust. Such a response is particularly active in
associated responsive behaviour, for example spaces where behaviour is, by tradition or habit,
changes of physical property such as light coded; such as a gallery, especially the pristine
emission, sound or form, or changes of content and white variety. Galleries, of one sort or another,
such as text or image. The making of interactive were the elected venues for the national tour of
works is focused upon two principal aspects: the ‘Making Buildings’ and so, during the initial phase
production and design of its responsiveness, and of design ‘a gallery induced blush’ became the
the conditions upon which the work is generic character to induce behaviour in the
encountered. paraform.
Sixteen*(makers) present exploration of ‘Blusher’ would glow and ruffle its feathers if
interactivity is informed by the broad manner in anyone came close, and intensify this behaviour
which audiences typically encounter physical if they came even closer. Its skin was sturdy
works of art and architecture. In particular, it is enough to invite the uninhibited to climb or walk
critical of the ‘secondary’ experience of upon it (although this was only ever attempted by
reproduction in which the majority of audiences the very young). Such irreverence to the ‘look but
acquire or supplement their immediate knowledge. don’t touch attitude’, was generated by our
Even the ‘actual’ experience of encountering concern that ‘Blusher’ should be understood as a
works is commonly paralleled by the availability work of investigative Architecture rather than Art
of elaborate catalogues, posters, and mass- or Craft. Furthermore, it was intended that
produced copies of the material on view. However, ‘Blusher’ should be considered as a work of
interactivity in art and architecture has the capacity potential for the making of buildings rather than a
to dispel the distinction between an actual and a statement of absolute intent, a question rather than
secondary experience and open up the potential an answer.
for multiple ‘original’ experiences of no
hierarchical value to both a local and remote ‘Blusher’s’ second consideration was one of
audience. adaptability to site specific conditions. Prior
information on each exhibition space was either
Interactivity in physical constructs offers the work too minimal or too slow in arriving and time
as ‘a system of parts’ for audiences to reassemble constraints did not allow a cross country
and manipulate in accordance with other events, reconnaissance mission. ‘Blusher’s’ structural skin
contexts and needs. By validating their personal was therefore developed for a range of

approximate conditions; tall spaces, narrow manufacture and use of artefacts in the built
spaces, low spaces, convoluted spaces, indoor environment?
spaces, and outdoor spaces. Decisions on how to 2. How can architectural practice adapt to non-
assemble and install the work were taken on site linear design strategies?
in a few hours a day or so before each venue
opened. This resulted in a unique configuration at Kielder Probes
each venue, either forming a singular construct,
or by fragmentation a series of scattered pieces in The first built output of the project will be a series
accordance with the volumetric, institutional and of ‘smart’ surveying instruments or ‘probes’. The
organisational character of its host space. ‘Blusher’ design of the probes is driven by two principal
would have a local and a generic state, in some factors of the site. Firstly, the qualities of the site
way a demonstration of the process of making that interest us are in a perpetual state of change
buildings. that requires surveying tools to measure difference
rather than the static characteristics of any given
The Kielder Residency instant. Secondly, the change that takes place
varies in time-scale and speed, thus the surveying
At the core of architectural practice is the gathering tools must be embedded on the site for a
of information from which spatial and formal sufficiently long period of time to capture those
interventions are proposed. In itself this is the first changes. For long periods of time they will be on
act of design. their own in which case they must be remotely
accessible both to collect their data and possibly
The territory of the forest possesses a variety of to reprogram them. These criteria necessitate the
spaces with particular architectonic qualities. design and fabrication of new instruments,
These qualities are continually redefined through instruments that seek out rather than simply
changing edge, canopy, ground conditions, and measure. Their data-acquisition strategy is a
varying degrees of density. These dynamic balance of opportunity, possibility and time with
qualities are the result but also affect the strategic which to observe, measure, log, store and transmit
moves in the management of the territory as a site-specific data as information that can usefully
whole. In turn, the management of such forests is inform a future architectural strategy. Exactly how
emerging as a highly sophisticated and the balance is weighted will only become apparent
manipulated process where the consequences of once the probes are settled on the site.
mass harvesting are considered for their impact
on the visual nature of the landscape as a whole. Kielder is an environment of exposure and
isolation that demands some self-sufficiency for
The residency project will attempt to decipher and those that live and work there. To achieve their
respond to these qualities. In this sense we are objective and cope with such an environment they
interested in exposing such issues as key concerns must be robust and occasionally exercise
in architectural design practice. In doing so, it is autonomy, to this end their operating system will
anticipated that the work will question how a site, be delegated a degree of decision-making needed
a sequence of sites, and components that define a to balance work with ‘survival’.
site, may respond to configuration, patterns of
occupancy, use and environmental cycles. In In addition, for the broad purpose the residency,
Kielder, of course, these cycles are both managed the installation of the probes is also planned as an
and unmanaged. architectural action. Their presence as customised
objects designed and made for specific
Aims, Objectives and Questions surroundings will present the first evidence of an
initial design process. Connections with local
To develop a design methodology in which built conditions such as ground, vegetation, exposure
artefacts have the ability to become increasingly and so on, will recognise uniqueness. It will be
specific to location and purpose. To explore the apparent they are made for Kielder and the dense
inherent potential of manufacturing techniques to repertoire that forms its identity and sense of place.
produce variety and difference in built artefacts.
1. Can notions of ’growth’ and ’adaptation’ Sites to position the probes will also be chosen
become processes that actuate the design, for varied character, aspect, proximity to a location

of interest, and so on. So not only will they will be unpredictable availability and timing of
function as survey instruments, they will also the energy source(s) that drive the thermo-
perform a purpose of physical interventions. They hydraulic pistons as conditioned by the weather
will enclose a territory and alter its status from a and site microclimate, in addition the probe’s
seemingly random plot of topography into a place operating system will be programmed to have
that anticipates future action. The probes are different responses (for and against) to harness the
therefore being designed to elicit change by availability sun-light.
altering their own physical character; they will
unfold, they will move, they will leave a trace, These are compelling reasons to expect each probe
and after time they will be replaced by a new to exhibit unique behaviour at any given time on
generation of architectural elements. a specific site. We anticipate this will add a quality
of uncertainty and delight for a visitor who comes
The time-based measurement of the behaviour of across one of the probes while walking amongst
the probes on each site captures metadata the forests of Kielder.
(information about information) that is mapped
onto a spatial-temporal representation of the site
and the probe’s location within it. It is this
representation that drives a virtual simulation
environment to model the behaviour of the probe 1
The Blusher Project was built for the ‘Making
in the computer. It also provides information in Buildings’ Crafts Council UK touring exhibition
real-time for the probe’s operating system to of 2001. The work was published in the following
control its internal behaviour in the face of the locations: 2001 Sheil, R, Callicott, N, and Till, J.
changing conditions of its microclimate. The
Kielder Forest and Reservoir present a territory ‘sixteen*(makers)’ in ‘Young Blood‘ profile no.
of such scale that potential sites of interest can 149 Architectural Design (AD) v71 n.1 (Feb
each have their own distinct microclimate. To 2001). Ed Spiller, N. p22-28.
capture the characteristics of each site with
individual spatial-temporal maps is the focus of 2001 Callicott, N. ‘Computer-Aided Manufacture
the probe’s design and will be the rationale for in Architecture -The Pursuit of Novelty’
how each probe is sited. The study of the data they Architectural Press (June 2001) p7-9, 79-97.
collect and the results of the virtual simulation will
potentially inform their adjustment in-situ, re- 2002 Jones, W. ‘Responsive Behaviour’ in Frame
location within their vicinity or replacement with No.26. pp108-103.
a revised or entirely new intervention.
2001 Melhuish, C. ‘High-tech and Customisation’,
Unlike a conventional architectural survey that a profile of sixteen* (makers) in ‘Making
will produce a representation of a site with optical Buildings’, A Crafts Council publication (Feb
instruments such as a theodolite, the probes will 2001) p25-29.
embody a formal response to their respective site
articulated by passive thermo-hydraulic pistons 2004 Sheil, R. ‘Design through Making’ in
that move the probe’s body. The piston’s actuation ‘Evolving Tool, Evolving Ideas-The Place for
is proportional to temperature within a specific Digital fabrication in Architectural Education’ Ed
range. There are two arrangements of the piston, Temkin, A p17-18. 2004 Sheil, R, Callicott, N,
one responds primarily to ambient (dry-bulb air) Ayres, P. ‘Shorting the Automation Circuit’ in
temperature and will be similar wherever the ‘Digital Fabricators’
probe(s) are sited at a particular point in time, in ,
some ways this will be deterministic because of Ed Stacey, M.pp80-81.
the average climatic temperatures experienced in
the region. The second is solar heated using a
parabolic reflector which is more unpredictable
as it is at the mercy of the site-specific
characteristics of shade or exposure in the forest,
minute-by-minute changes in cloud cover and self-
shadowing by elements of the probe itself. There

London June 2005
Figure 1. Kielder, Northumbria, England, April 2005
Figure 2. The Blusher, Venue 7: The Bartlett School of Figure 4. The Blusher, Venue 5: Aberstwyth Art Centre
Architecture, UCL
Figure 3. The Blusher, Venue 7: The Bartlett School of Figure 5. The Blusher, Venue 4: Turnpike Gallery Leith
Architecture, UCL

Figure 6. The Blusher Venue 1:

The New Art Gallery Walsall
Figure 7. The Blusher Venue 1: Figure 9. Kielder probe: simulation sequence of

The New Art Gallery Walsall thermohydraulic piston actuation.
Figure 8. Kielder probe: ambient air temperature trace from

thermochron®.

2005 - Acadia05 - Book Smart Homes For Rural

Uploaded by

Copyright:

Available Formats

2005 - Acadia05 - Book Smart Homes For Rural

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

2005 - Acadia05 - Book Smart Homes For Rural

Uploaded by

Copyright:

Available Formats

ACADIA05: Smart Architecture 1

Copyright © 2005 by The Association for Computer-Aided

All rights reserved by the individual paper authors who are

Cover design by Osman Ataman.

No part of this work covered by the copyright hereon my be

Library of Congress Catalog Card Number: 2005934032

2 ACADIA05: Smart Architecture

President, Kevin Klinger

ACADIA’05 Technical Chair, Osman Ataman

ACADIA05: Smart Architecture 3

Anders, Peter Johnson, Brian

Ataman, Osman Johnson, Scott

Bermudez, Julio Kim, Micheal

Castillo, Tim Klinger, Kevin

Chan, Chiu-Shui Kolarevic, Branko

Clayton, Mark Kvan, Thomas

Elvin, George Luhan, Gregory

Fowler, Tom Martens, Bob

Guzowski, Mary McCullough, Malcolm

Harfmann, Anton Muller, Voelker

4 ACADIA05: Smart Architecture

ACADIA05: Smart Architecture 5

ACADIA’05 Officers and Anne Marie Due Schmidt

ACADIA’05 Tecnical Review Committee 4 Carmina Sanchez-del-Valle

Sachin Anshuman Jaewook Lee, Yehuda E. Kalay

Mahesh Senagala Daniel Barker, Andy Dong

Michael Silver Thomas Modeen, Christine Pasquire,

Peter Anders, Werner Lonsing

6 ACADIA05: Smart Architecture

Nimish Biloria, Kas Oosterhuis, Cas Aalbers

Bob Sheil, Chris Leung

ACADIA05: Smart Architecture 7

8 ACADIA05: Smart Architecture

ACADIA05: Smart Architecture 9

10 ACADIA05: Smart Architecture

José Daveiga, Paulo Ferreira

Michael Fox, Catherine Hu

Peter Anders, Werner Lonsing

Anne Marie Due Schmidt

ACADIA05: Smart Architecture 11

Responsiveness and Social Expression; Seeking

12 ACADIA05: Smart Architecture

Introduction Augmenting such space would require

ACADIA05: Smart Architecture 13

14 ACADIA05: Smart Architecture

ACADIA05: Smart Architecture 15

Figure 2. Hyposurface is pneumatically driven animate membrane developed by dECOi

16 ACADIA05: Smart Architecture

ACADIA05: Smart Architecture 17

Figure 4. Primary model of contemporary building services’ architecture

18 ACADIA05: Smart Architecture

responsiveness beyond mechanical automation, functionally and socially, require an outlook

ACADIA05: Smart Architecture 19

Table 1: User interaction and climatic input in intelligent façades

20 ACADIA05: Smart Architecture

Table 2: Existing Intelligent Building Standards and Attributes

ACADIA05: Smart Architecture 21

22 ACADIA05: Smart Architecture