Tactile Paper Prototyping with Blind Subjects

Mei Miao1 , Wiebke Köhlmann2 , Maria Schiewe3 , and Gerhard Weber1

1
Technische Universität Dresden, Institut für Angewandte Informatik
Nöthnitzer Straße 46, 01187 Dresden, Germany
{mei.miao,gerhard.weber}@tu-dresden.de
2
Universität Potsdam, Institut für Informatik
August-Bebel-Straße 89, 14482 Potsdam, Germany
koehlmann@cs.uni-potsdam.de
3
F. H. Papenmeier GmbH & Co. KG
Talweg 2, 58239 Schwerte, Germany
maria.schiewe@papenmeier.de

Abstract. With tactile paper prototyping user interfaces can be evalu-

ated with blind users in an early design stage. First, we describe two exist-
ing paper prototyping methods, visual and haptic paper prototyping, and
indicate their limitations for blind users. Subsequently, we present our
experiences while preparing, conducting and analysing tests performed
using tactile paper prototyping. Based on our experiences, we provide
recommendations for this new usability evaluation method.

Keywords: tactile paper prototyping, low-fidelity prototyping, usability

evaluation method, visually impaired, tactile interaction, design method-
ologies, usability, user-centred design.

1 Introduction
Paper prototyping is a widely used method in user-centred design (see ISO 13407)
to develop software that meets users’ expectations and needs. Testing concepts
with prototypes before implementation allows for inexpensive changes as paper
mock-ups, e. g. hand-sketched drawings, can be adapted quickly according to
users’ comments.
Henry [1] describes common procedures for testing accessible software. Be-
sides the conformance to accessibility standards, evaluation expertise and the
experience of people with disabilities is needed to evaluate applications. Speci-
fied methods such as heuristic evaluation, walkthroughs and screening techniques
can be conducted with design team members or users. Though, Henry does not
describe any concrete methods for testing with subjects with disabilities in the
early development stage.
Visually impaired access digital information using assistive technology such as
screen readers and Braille displays. Within the project HyperBraille1 , a tactile
two-dimensional display with the size of 120×60 pins, the BrailleDis 9000, is
1
HyperBraille project website: http://www.hyperbraille.com/

M.E. Altinsoy, U. Jekosch, and S. Brewster (Eds.): HAID 2009, LNCS 5763, pp. 81–90, 2009.

c Springer-Verlag Berlin Heidelberg 2009
82 M. Miao et al.

being developed which can be used to display multiple lines of text and graphical
information [2]. Furthermore, interaction is possible through its touch-sensitive
surface. In addition to the hardware, a software system for presenting content
of conventional applications (e. g. Microsoft Office, Internet Explorer) is being
developed which considers the special needs of blind users. The adaptation of
detailed and coloured GUIs to a lower binary tactile resolution with adjusted
interaction techniques required an elaborate conceptual design accompanied by
ongoing formative evaluation2 .
The first usability test of our concept’s tactile user interface was conducted
in an early development stage. For this evaluation we applied tactile paper pro-
totyping (see Section 3) in combination with audio confrontation [4]. The focus
of this paper is not to give a summary of our evaluation, but to present our
observations and recommendations for conducting tactile paper prototyping.
The paper is structured as follows. First, an overview of types of paper proto-
typing is given. After a brief description of the evaluation, our observations while
preparing, conducting and evaluating our tests are discussed. In the following,
recommendations for conducting tests using tactile paper prototyping with blind
subjects are given. The paper closes with a conclusion and an outlook.

2 Paper Prototyping

In user-centred design, paper prototyping is a widely used low-fidelity usability

inspection method to evaluate drafts in an early stage of product design. Proto-
types similar to the final product are called high-fidelity while those less similar
are called low-fidelity [5]. In this paper we focus on prototypes for user interfaces.
Subjects evaluate products or applications via mock-ups that provide low
functionality and can consist of different materials. Alternatively, prototypes,
usually computer applications, provide more functionality, but are normally cre-
ated later in the development process when basic concepts have already been
approved [6]. Paper prototyping does not only serve to evaluate existing con-
cepts and to identify weaknesses, it also offers the possibility to the subjects
to make suggestions for improvement. This technique is very inexpensive and
effective as it allows for testing a product before implementation.
Mock-ups are prepared in advance by mock-up designers who are not neces-
sarily identical to the product designers. Conducting paper prototyping normally
requires four responsibilities: greeter, facilitator, computer and observer [7]. A
greeter is responsible for welcoming subjects, the facilitator conducts the session.
Usually developers play the role of computers, manipulating the interface pieces
according to the subject’s actions. Observers take notes and are responsible for
recordings.
Depending on modes of perception, we can classify current paper prototyping
methods in visual (2.1) and haptic (2.2) paper prototyping. According to this
naming convention we call our new approach tactile (2.3) paper prototyping.
2
For more information on formative evaluation see [3].
 Tactile Paper Prototyping with Blind Subjects 83

2.1 Visual Paper Prototyping

Among the three methods visual paper prototyping is the most widely used. Its
mock-ups consist of drawn interfaces on one sheet of paper or of several movable
and interchangeable individual interface elements which simulate interaction on
a static background interface.
According to Snyder [8] most paper mock-ups do not need straight lines or
typed text and consistent sizing of components. A complete and neat looking
design rather encourages unwanted pedantic feedback, e. g. concerning alignment
and sizing. Additionally, Snyder points out that paper prototyping encourages
creativity as the handwritten mock-up looks unfinished.
The nature of visual paper prototyping assumes that subjects are sighted and
can evaluate the designs with the help of visual information. Thus, this method
excludes visually impaired and blind users.

2.2 Haptic Paper Prototyping

Haptic paper prototyping is a special form of haptic low-fidelity prototyping [9].
It serves to simulate and evaluate haptic interaction with a haptic application
in an early development stage. A common material for mock-ups is cardboard.
In contrast to visual paper prototyping, haptic paper prototyping can find
limited use with visually impaired and blind subjects under the condition that
pure haptic interaction is concerned and visual perception is unnecessary. For this
reason this method is seldomly applicable with blind subjects when evaluating
GUIs. One of the few examples is a media set (see Fig. 1) for teaching graphical
user interfaces to blind students, developed by the project EBSGO [10].

Fig. 1. GUI-Taktil of the EBSGO project showing a search dialog box

Moreover, Tanhua-Piironinen and Raisamo [11] used two types of haptic

mock-ups consisting of cardboard models and plastic artefacts for tests with
visually impaired children. They pointed out that a possible drawback of this
method was the abstract model which does not allow for a full conception of the
application as a whole.
84 M. Miao et al.

2.3 Tactile Paper Prototyping

To be able to perform tests with blind subjects, we adapted visual paper proto-
typing according to our requirements. In principle, the concept (see Section 2.1)
is also applicable to test user interfaces with blind subjects, but the special needs
of this user group have to be considered.
Accordingly, our evaluation method allows for evaluating user interfaces re-
specting the subjects’ needs. Indeed, speech output of screen readers can be
presented neither by paper nor by tactile mock-ups. Thus, to create a realistic
work environment, speech output needs to be presented by the conductors in
the role of computers.
The integration of haptic prototyping techniques in our method is conceivable,
as interface elements with a structured surface can be used to indicate certain
details, e. g. focused elements, as a compensation for highlighting on visual mock-
ups. In our special case, evaluating an interface for a device whose pins can only
be set or not (on/off), additional haptic elements were not needed.

Fig. 2. Subjects exploring our tactile mock-ups

Test material for visual paper prototyping usually consists of individual inter-
face elements which can be produced, arranged and changed quickly according
to the simulated action or the suggestions of the subject. As more preparation
is needed for tactile mock-ups when compared to the pen-and-paper approach,
changes are unlikely while conducting the evaluation. Anyhow, it is possible
to generate new mock-up elements with users during the evaluation e. g. using
paper and heat-pen or Braille paper.

3 Evaluation
The evaluation is only briefly overviewed as the focus of this paper lies in the
evaluation technique of tactile paper prototyping. A comprehensive description
of the evaluation can be found in [4].
In our evaluation, we used paper prototyping in a vertical3 and low-fidelity
manner with a total of 11 blind subjects in groups of two to five. Low-fidelity
3
Vertical prototyping tests the exact functionality of few elements of a GUI. In con-
trast, horizontal prototyping tests a broad spectrum of GUI elements with a low
level of functionality.
 Tactile Paper Prototyping with Blind Subjects 85

prototyping was appropriate as it was the first test of the designed interface. We
used vertical prototyping, as we only focused on some features such as layout
and application concept, while neglecting navigation and interaction.
Our test comprised several scenarios with one or two pages each which rep-
resented the adapted GUIs for our two-dimensional tactile device. Embossed
printings, matching the resolution of the target output device, served as test
material. Within our project we relied on two self-developed programmes called
HBGraphicsExchange and HBBrailleExchange [12] which allowed our blind and
sighted designers to create mock-ups in the appropriate resolution and size,
printable with any embosser. To allow for comfortable turning and to sustain
the order, the sheets were assembled in binders (see Fig. 2). The hands of the
subjects on the mock-ups were videotaped and the discussions were recorded.

4 Observations

In the following, we only report on observations concerning blind subjects. Of

course, most general aspects for preparing, conducting and analysing a test with
sighted subjects [3] apply as well. Additionally, general issues [1][13] regarding
blind subjects have to be considered. Subjects might need escort and transporta-
tion to and from the facility and are likely to bring along a guide dog. Thus,
setting up the test takes more time and during its conduction additional breaks
might be needed. Furthermore, extra room for service animals or assistants is
needed and obstacles should be cleared out of the way.

4.1 Preparing

For tactile paper prototyping thorough preparation is important. To be able to

present several solutions, it is advisable to produce multiple mock-ups for each
scenario in advance, as new ideas of subjects are difficult to cover during a test.
Embossers. Creating tactile paper mock-ups requires special hardware. In the
following we only discuss Braille printers, called embossers, because their prints
have similar physical properties as Braille displays and most planar tactile de-
vices. In general, the material used should resemble the final product as much
as possible to ensure realistic results.
Before designing tactile mock-ups, essential facts of the embosser are needed.
These comprise the embosser’s resolution, equidistant or non-equidistant output,
and the format and type of paper needed. Printable file formats of the chosen
embosser affect the choice of the software for designing the mock-ups.
Software. Alternatively to the programmes for creating mock-ups used in our
evaluation, Word documents can be interpreted by most printer drivers. While
this works well for text, graphics are interpolated, resulting in tilted lines of
varying thickness. Thus, it is best to create graphics matching the embosser’s
resolution. Graphics software can also be used, if its file formats are directly
supported or if translation software is available.
86 M. Miao et al.

Material. In contrast to sighted users who perceive representations as a whole

and focus on details later on, blind users first explore details to construct a
complete mental model. Thus, one representation of our test comprising multi-
ple widgets on one sheet of paper caused difficulties in locating the described
element. Therefore, it is advisable to display only one representation per sheet.
To avoid orientation difficulties due to sketchy representations it proved es-
sential to map geometrical shapes as precise as possible, and to maintain the
proportions and scale of the original output device, especially as Braille requires
a fixed resolution. Thus, the special needs of the target group impede confor-
mance to Snyders demand for handwritten visual mock-ups (see Section 2.1).
In our test, subjects were irritated by missing elements as some regions of
our representation were not as detailed as the planned implementation. In such
cases missing elements need to be indicated on the mock-ups or mentioned by
the facilitator to avoid confusion. As it is more difficult to refer to and to dis-
cuss certain elements or positions on the mock-ups than with sighted subjects,
a coordinate system can be helpful to improve the subjects’ orientation, e. g.
dividing the representation like a chess board and referring to the squares.
Proofreading. Before producing copies for the subjects, mock-ups should be
proofread by Braille literates. This is necessary as spacing and spelling mis-
takes frequently occur when transcribing from print to Braille. In a non-visual
context, spelling mistakes irritate even more than in visual mock-ups as the two-
dimensional representation is unfamiliar to the users and an overall overview is
missing. Additionally, the facilitator, not necessarily Braille literate, cannot con-
trol and correct mistakes as easily as in print.
Unless sighted designers are experts in Braille and working techniques of the
blind, it is, of course, best to include blind mock-up designers in the prepara-
tion process. A couple of our mock-ups were dismissed by the subjects because
the design strongly resembled visual concepts. These mock-ups had not been
reviewed by our blind designers and had inevitably failed during the test.
Assembling. Supporting a smooth work flow, the order of the mock-ups used has
to be consistent with the tasks. It proved useful to label the sheets with numbers
or letters in Braille and print for the subjects and the facilitator.
When movable individual interface elements are used, they have to cling to
the main representation (with help of magnets, sticker, felt etc.) to avoid shifting
during exploration. As we conducted our tests with groups, individual interface
elements would have been impracticable. Thus, changing screens were simulated
by different prints, available by turning the page.

4.2 Conducting
After preparing the test material, five main aspects have to be taken into con-
sideration. These comprise setup, recording and team structure. Furthermore,
timing and explaining are crucial for tactile paper prototyping.
Setup. Our mock-ups were in landscape format; therefore the opened binders
faced the subjects with their narrow side. In one location, the width of the table
 Tactile Paper Prototyping with Blind Subjects 87

was not sufficient, thus the two facing binders touched each other and impeded
the turning of pages. It is favourable to choose a table which is large enough to
arrange all test materials comfortably and to allow the subjects to move their
hands and arms freely to avoid collisions while exploring a paper mock-up.
Recording. As a repetition of tests with small user groups is difficult, it is ad-
visable to prepare and record the sessions carefully. In advance, the positions of
microphones and cameras need to be considered and tested depending on the
seating arrangements and the area of interest to be recorded. We positioned a
camera on the table to record the subjects’ hands and the mock-ups.

Fig. 3. Subject with hands under mock-ups

Unfortunately, with some subjects the recording was useless, as they explored
the mock-ups with their hands underneath the previous sheet of paper, i. e. they
did not turn the page (see Fig. 3). In such cases the facilitator must ask the
subjects to turn the pages completely.
While the video shows, where the subjects’ hands are positioned, it does not
show satisfactorily if there was contact with the mock-up at all and which parts
of the hand touched the mock-up, and how much pressure was applied.
Team. When using movable individual interface elements, the number of people
needed for conducting tactile paper prototyping increases with the number of
subjects. For a test with several subjects, the facilitator cannot demonstrate
tactile representations to all group members simultaneously. Therefore, subjects
must be provided with their own copy of a mock-up as it can only be explored by
one person at a time. The subjects must be able to handle these copies on their
own or, depending on complexity, one assistant (in the role of the computer) for
one or two subjects is necessary.
Timing. Before dealing with a new mock-up, the facilitator has to make sure
through announcing the mock-up’s label that each subject has the correct mate-
rial in front of him. Furthermore, the facilitator has to consider different reading
speeds and exploration styles (one- or two-handed) of the subjects. Without
allowing for sufficient exploration time, subjects might be overwhelmed by the
amount of information, try to keep up at the expense of exploring details, ask
unnecessary questions or be more likely to abort the test.
88 M. Miao et al.

Presenting mock-ups in binders can be problematic, when subjects are too

curious. Some of our subjects were distracted by successive scenarios because
they turned pages. An alternative could be to hand out individual sheets of
paper. When doing so, the computer has to ensure that all mock-ups lie in
the correct orientation in front of the subjects to avoid confusion. Therefore
piles of mock-ups should not be circled around but rather be placed in front of
each subject individually. Though, this procedure can be disruptive and time
consuming. The facilitator should thus, when using binders, indicate during and
in the beginning of the test when it is appropriate to explore which mock-up or
when e. g. listening is desired.
Explaining. As blind subjects seldom have experience and knowledge about
GUIs, it is essential to explain the design, content and purpose of the general
design concept and the current mock-up from their perspective. This task can
become a challenge, because difficulties in understanding might occur, even if
the facilitator and his assistants are familiar with blind work techniques. It is
advisable to take help from blind experts in advance and prepare explanations
with them beforehand.
Additional illustration is also needed for specific tasks in applications that
blind users are not familiar with, in order to allow them to comment on the
implementation proposal. In our test, prior explanation of the mock-ups proved
helpful because not all subjects possessed sufficient background knowledge.

4.3 Analysing
Shortly after the test, the results and impressions should be documented and
analysed to avoid forgetting important details. The evaluation can comprise
reading and completing the minutes, transcribing the audio to collect comments
and evaluating the video. When video recordings are used, it is advisable to
have ink printing superposed with embossing on one sheet. One could first print
and then emboss or use a special printer which can do both simultaneously.
In our test the sole embossed printings were hard to perceive on the recording
when analysing the video, so that the elements explored in scenes could only be
guessed by the positions of the subjects’ hands.
After extracting impressions and comments of the subjects, it should be de-
cided whether it is a problem of concept, material, explanation or a personal
preference. When testing multiple mock-ups, e. g. for different scenarios, it is ad-
visable to compare the comments concerning the different mock-ups to extract
aspects which can be applied to the entire concept or design.

5 Recommendations
The main goal of an evaluation is to identify existing problems and to find po-
tential for improvements. To achieve this aim, the evaluation must be prepared,
conducted and analysed carefully and adequately. In the following, we present
recommendations for tactile paper prototyping which have been condensed from
our observations (see Section 4):
 Tactile Paper Prototyping with Blind Subjects 89

– Consider general issues when hosting blind users.

– Provide adequate facilities according to the special needs.
– Check for special hardware and software required.
– Design mock-ups not for sighted but for blind users.
– Even better, have blind people design the mock-ups.
– Proofread the mock-ups before conducting the test.
– Check the recording before and during the test, as repetition is expensive.
– Provide a sufficient number of assistants.
– Make sure that mock-ups are provided synchronously to each subject.
– Allow for sufficient time to explore the mock-ups.
– Explain from the blind users’ perspective.

Different usability evaluation methods have different aspects which should be

considered. Nevertheless, many of these recommendations do not only apply to
tactile paper prototyping but are also applicable for other evaluation techniques
involving blind subjects.

6 Conclusion and Outlook

Tactile paper prototyping is a new approach to design haptic user interfaces. It

allows to bridge between the visual and haptic modality while ensuring multi-
modality when using assistive technology to gain access to graphical user inter-
faces. Tactile paper prototyping applies to user centred design.
We developed 16 tactile paper mock-ups to design a user interface which
includes Braille and tactile graphics using a planar tactile display. Due to limi-
tations of resolution and size of such a tactile display re-design of visual concepts
is required. Like paper mock-ups, tactile mock-ups allow for verification of de-
sign concepts before implementation and involvement of end users. Therefore,
we found it essential that blind people contribute to mock-up production and
that mock-ups are evaluated by prospective blind users.
To aim at validation of our approach a follow-up study using a wizard-of-oz
approach has been conducted. It involved three separate users who confirmed the
suitability of a selection of haptic designs on the actual planar tactile device [14].
The mock-ups that were rejected during the tactile prototyping had deliberately
not been included in this evaluation.
Future work will have to extend our approach to the design of audio-haptic
interfaces possibly supporting also Braille-illiterate users. The suitability of tac-
tile paper prototyping must be tested for other application areas such as maps,
games, or collaborative software. The analysis of hand contact will have to be
considered in more detail in order to understand failures and mismatches in the
design more easily.

Acknowledgements. We thank all blind subjects who participated in our eval-

uations. We also thank Oliver Nadig and Ursula Weber for organising the tests
and Christiane Taras for providing us with the software to create the mock-ups.
90 M. Miao et al.

The HyperBraille project is sponsored by the Bundesministerium für Wirt-

schaft und Technologie (German Ministry of Economy and Technology) under
the grant number 01MT07003 for Universität Potsdam and 01MT07004 for Tech-
nische Universität Dresden. Only the authors of this paper are responsible for
its content.

References
1. Henry, L.S.: Just Ask: Integrating Accessibility Throughout Design (2007)
2. Völkel, T., Weber, G., Baumann, U.: Tactile Graphics Revised: The Novel
BrailleDis 9000 Pin-Matrix Device with Multitouch Input. In: Miesenberger,
K., Klaus, J., Zagler, W.L., Karshmer, A.I. (eds.) ICCHP 2008. LNCS, vol. 5105,
pp. 835–842. Springer, Heidelberg (2008)
3. Rubin, J., Chisnell, D.: Handbook of Usability Testing: How to Plan, Design, and
Conduct Effective Tests, 2nd edn. Wiley, Chichester (2008)
4. Schiewe, M., Köhlmann, W., Nadig, O., Weber, G.: What You Feel is What You
Get: Mapping GUIs on Planar Tactile Displays. In: HCI International 2009 (2009)
5. Walker, M., Takayama, L., Landsey, J.: High-fidelity or Low-fidelity, Paper or Com-
puter? In: Proceedings of the Human Factors and Ergonomics Society, pp. 661–665
(2002)
6. Snyder, C.: Paper Prototyping: The Fast and Easy Way to Design and Refine User
Interfaces. Morgan Kaufmann, San Francisco (2003)
7. Rettig, M.: Prototyping for tiny fingers. Commun. ACM 37(4), 21–27 (1994)
8. Snyder, C.: Paper prototyping. IBM developerWorks (2001)
9. Magnusson, C., Brewster, S. (eds.): Guidelines for Haptic Lo-Fi Prototyping (2008)
10. Schwede, K.J., Klose, U.: EBSGO-Lehrerhandbuch: Grundlagen für den blinden-
und sehbehindertenspezifischen Unterricht mit grafischen Oberflächen [EBSGO
Teacher’s Guide: Basics of Education for the Blind and Visually Impaired with
GUIs]. Tectum-Verlag, Marburg (2001)
11. Tanhua-Piiroinen, E., Raisamo, R.: Tangible Models in Prototyping and Testing
of Haptic Interfaces with Visually Impaired Children. In: NordiCHI 2008 (2008)
12. Taras, C., Ertl, T.: Interaction with Colored Graphical Representations on Braille
Devices. In: HCI International 2009 (2009)
13. Dumas, J.S., Loring, B.A.: Moderating Usability Tests: Principles & Practices for
Interacting, pp. 133–155. Morgan Kaufmann, San Francisco (2008)
14. Prescher, D.: Ein taktiles Fenstersystem mit Multitouch-Bedienung [A Tactile Win-
dow System with Multitouch Operation]. Diploma thesis, Dept. of Computer Sci-
ence, Technical University of Dresden, p. 122 (2009)