Our research comes at a time when an interest in the theoretical and critical concerns around archives has begun to pervade discussions of designing for the archive. This tendency reflects themes in archive-oriented humanities scholarship, since the early 1990’s in particular [Derrida 1995], about the role of archives as repositories of culture and archivists and institutions as producers of versions of the past [Hardiman 2009]. As archival practice has become increasingly mediated by digital technologies, these themes have become of key interest to scholars positioning their work in terms of the Digital Humanities e.g. [Drucker 2013] [Kirschenbaum 2008]. Thus a converging set of theoretical concerns, disciplines and epistemologies center on the technologies and practices of archiving, and their relationship to older critical perspectives from the humanities. The overlap of Human Computer Interaction (HCI) and Interaction Design research with concerns from the Digital Humanities represents a compelling opportunity for inter-disciplinary research.

Our research was conducted in the context of an eighteen-month-long research project. At the outset our university had recently acquired the archive of a publisher of contemporary poetry, Bloodaxe Books, consisting of around 60,000 items (see Figure 1), mostly edited manuscripts of poems.

We were engaged to produce experimental interface design research as part of an inter-disciplinary team together with other researchers from the creative arts and humanities. Around thirty project participants, mostly poets themselves, were engaged in personal, creative-practice research in the archive. Importantly, the archive itself was in a completely un-catalogued state at the start of the project. Research was consequently being conducted in parallel to the cataloguing and digitization process and we will contend that this contemporaneity afforded a range of creative and technical opportunities that helped us explore, describe and use the concept of archival liveness.

Liveness has been productively recognized as a concept in the field of HCI and has been investigated across areas including live electronic music [Hook, Schofield, and Taylor 2012] [Hook 2013], interactive art [Taylor 2011] [Reeves, Benford, and Malley 2005], and TV production [Bartindale and Oliver 2009] sometimes looking to theatre and performance art for comparison [Auslander 2008]. Liveness has been shown to have significant implications for user experience, particularly of live events. Some authors have suggested that work remains in considering liveness in interaction more broadly. During the early phases of our research we were frequently confronted with a number of technical and experiential ways that the archive could be considered as subject to live processes (such as the ongoing activities of digitization and cataloguing) and indeed, could be considered itself as “live”. These considerations lead us to consider notions of archival liveness and its possible application to the study of the user experience of archives, collections and their interfaces.

Previous accounts of liveness [Hook, Schofield, and Taylor 2012] [Hook 2013] [Taylor 2011] [Reeves, Benford, and Malley 2005] [Bartindale and Oliver 2009] describe how a sense of liveness can be achieved and supported, and emphasise the power of a feeling of liveness to affect experiences of technologically mediated events. Definitions of liveness as “…the properties of intimacy and immediacy experienced by both spectators and performers”  [Hook, Schofield, and Taylor 2012] are calibrated to evoke interpersonal, affective encounters. However the more recent use of the term in the context of data visualization [Jacobs 2013] provokes a reconsideration of its features of interest away from performance. An expanded sense of liveness has the potential to inform interaction design more broadly [Hook, Schofield, and Taylor 2012]. The study of liveness has, in summary, hitherto focused primarily on the experience of liveness, through which performances, art or entertainment are lived.

In the context of data-driven technological experiences (such as viewing or using visualization) however, an account of how the formal characteristics of the supporting infrastructure actually contribute to the phenomenon of liveness is so far under-developed. As part of our second contribution, we develop this formal description and describe the role it played in affording opportunities for design. In this sense we find commonality with research in live electronic music, in particular Freeth et al. [Freeth, Bowers, and Hogg 2014] who describe how the experience of a live, networked, collaborative musical composition is shaped by the technical constraints of the system. Others [Tanaka 2000] have suggested that such constraints should be seen as features of particular material interest and be viewed as aesthetically integral to the piece of music, rather than being viewed as problems. In the study presented herein we follow prior work e.g. [Tanaka 2000] [Freeth, Bowers, and Hogg 2014] by seeing such technological materiality as part of the landscape of what makes the archive a particular kind of live.

Concerns about the power of archives to produce particular versions of our cultural past [Hardiman 2009] have motivated some researchers to adopt participatory design (PD) or similar approaches to involving future users of archives in the design process. Doing so raises a series of complex questions regarding the stakeholders in such archives (whose archive is it?) and their relationship to archival standards (by whose terms should the items be described?). Featherstone asks “should the focus be on received traditions and the canon, or on local knowledge and diversity?”  [Featherstone 2006], and a number of research projects have opted emphatically for the latter. Povinelli [Povinelli 2011] describes the production of an archive of cultural information with and about indigenous Australians. The interface to this archive attempted to recreate patterns of knowledge sharing (and exclusion) particular to the group in question. Huvila [Huvila 2008] highlights the importance of knowledge held by future users of an archive of local history and describes its integration into the design of a records management system. These explicitly participatory projects come with a considerable commitment in time and resources. Although these results are valuable, we suggest that there is scope for approaches which are responsive to users but without the sometimes prohibitive costs involved in pure PD processes. One such process is described by Feinberg [Feinberg 2012] who suggests a notion of “dialogue” between collections interface designer (as author) and user (as audience).

With the integration of archival record keeping and networked technologies (such as document-storing servers, online record aggregators [JISC], and front-end collection interfaces) [Hedstrom 2002, 21] Hedstrom notes that as a “critical node” in the relationship between archive, archivist and user, archival interfaces are implicated in the interconnections between power, institution, memory and artefact described above. As well as the decisions taken by archivists as to inclusion, preservation, access and description, a set of further considerations inform the presentation and curation of archives. Such considerations encompass interface design, data management strategy and access policy and are a site of potential interdisciplinary collaboration and user engagement. The context of a research project afforded us an opportunity to pursue creative exploratory work in this design space.

We have seen first that liveness has been of interest in design communities in providing clues about the experience of live performance and interaction. Significantly, we have noted that we can develop the concept of liveness in the archive by supporting the discussion of experience with a description of some of the formal features of the archival process which contribute to the phenomenon of liveness. The development of designs exploring archival liveness is proposed as a useful contribution to the rich but complicated design space of archival interfaces.

We adopted a Research Through Design methodology as exemplified by Frayling, Gaver, Bowers and others [Frayling 1994] [Gaver 2012] [Bowers 2012], in that artefacts produced “allow a range of topical, procedural, pragmatic and conceptual insights to be articulated”  [Gaver 2012]. Research Through Design has been demonstrated as compatible with a critical [Bardzell 2013] orientation which is consistent with our aims to consider how archival liveness might support the production of a more open, experimental or creative engagement with archival materials. Our approach is critical in that we propose archival liveness as a concept through which to query existing design practices in archive interfaces and to provide pathways towards new kinds of design.

In the following we describe four distinct design activities: Archive Inserts, Box Log Connections, The Marginalia Machine and a Twitter bot. These activities were conducted on an overlapping timeline; here we present them in a roughly chronological sequence that shows their formative role in developing our understanding of archival liveness.

Inspired by previous Research Through Design work [Gaver 2007] [Gaver 2004] we conceived of a “cultural probes” activity to gain insight into the way the archive was being used by our project participants and to uncover some of the things they found interesting about the materials themselves. As poets with substantial domain knowledge they could be considered as “lead users”  [Von Hippel 1986] with the potential to give insight informing future designs for a broader group.

With our project partners we also held a series of four, hour-long discussion meetings with groups of around ten participants at a time. These meetings were audio recorded and transcribed. In the following section we draw on the probing activity and the results of the discussion together. We emphasize that these activities were undertaken not as information gathering or requirements analysis activities per se, but to sensitize ourselves to the material and its prospective users. That is to say our discussions were not directed towards user requirements specifically but instead tried to get some flavor of their experience in the archive.

Towards this end we also spent time physically situated in the archive ourselves. We observed during this period that the library silence imposed by the archive’s location in the university “Special Collections” room meant that participants who were potentially conducting related research were unable to communicate. Indeed a participant commented during a session that,

“Both times that I have been to the archive, I have requested a box and I was emailed back that somebody else had requested a box the same as me. So I got this strange feeling that there was this other me, who was interested in the same things that I am interested in. I think the fact that we were actually going to the same session, and wanting the same box for it, I started to feel a bit weird about it.”

Partly in response to this scenario, we created a probing activity, “Archive Inserts” (see Figure 2) which used a bookmark-like insert to act as a conversation “backchannel” for participants. A trial of this activity revealed a number of interesting features of the archive which focused our interests and informed future designs.

We scanned the completed inserts and conducted a basic thematic analysis, sorting results into the following categories; “materiality”, “typographical”, “literary criticism”, “questions about the cataloguing process”, “poems” (several participants chose to leave poems for others to find), “links between items”, “illustrations” (several participants drew pictures of what they found) and “declarations” (comments which didn’t appear to solicit a particular response).

There were two principal outcomes from our analysis. First: it was clear that temporality was a significant factor in the participants’ experience of the archive. Second: virtually all of the aspects identified as interesting by the participants were not intended to be described in metadata in the forthcoming catalogue, and thus would not be represented in future interfaces to the archive. This latter observation is fundamental to the value we identify for a Research Through Design process in this context. These outcomes are described further below.

Our work with participants had revealed a number of features of interest in the archive that informed our designs. First our sense of the vital importance of the temporality of archive materials was clearly shared by participants. Second, the aspects of materiality valued by the participants presented us with some clear ideas about what kinds of material to work with. Lastly the comments about the relationship between materials and the ongoing cataloguing supported our earlier impression that presenting the archive during the cataloguing process might engage users at this critical and contested stage.

Having spent time in the archive and with project participants we developed two main design proposals both of which adopted the archival marginalia as a starting point for a creative design intervention. A further design was produced early in this process and is described separately. Although neither of the former two was in itself an interface design, both were produced to “lead” to the catalogue and future interfaces to it in different ways. The first design intervention was a live Twitter bot, connected to the digitization process in the library (see Figure 4). The server running the bot uses machine learning techniques to separate handwritten notes and illustrations before tweeting the resulting image with a link back to a catalogue page with basic catalogue information. The detection of this marginalia relies on the results of optical character recognition (OCR) as performed by the Tesseract OCR engine [Smith 2007].

OCR detects typewritten symbols, associating a level of detection confidence with each symbol. This allows estimation of the expected colour and dimensions of typewritten symbols for a given document. Regions of content either ignored by OCR or whose colour and/or dimensions lie outside the expected range for typewritten symbols are then deemed to be marginalia.

The second intervention was an artwork, the Marginalia Machine (see Figure 5). The Marginalia Machine is an “X/Y” or “Cartesian” plotter which draws details [1] of the marginalia on a continuous scroll of paper. It was displayed during “Poetry International” an internationally recognized festival of poetry at the Southbank Centre, London, UK.

The Marginalia Machine is based on a kit of parts sold by Makeblock [Makeblock 2014] which was significantly modified and improved. We added stabilizing belts and pulleys and designed a paper roller mechanism that would advance after each document was drawn. The Marginalia Machine was controlled by an Arduino microcontroller in serial communication with a Java application which we wrote specifically for the project. This application sent a series of point coordinates combined with “pen up” or “pen down” messages. To produce these coordinates we wrote a Python script which vectorised the marginalia images (produced as described above) and produced an XML document of coordinate data. A supporting documentation video, shown in the gallery space, described the technological processes behind the machine and its context in our design research.

The design process afforded a number of insights that support our contribution of a definition of archival liveness. As we developed our designs, a set of three “sites of liveness” emerged, each with its own, idiosyncratic sense of liveness. These were: the special collections area of the library itself, the server infrastructure that would host aspects of our design interventions, and the manifestations of our designs on Twitter and in the gallery.

In the previous section we described three distinct temporal facets to our work in the archive: the working rhythms of the archivist and digital assistant, the execution and scheduling times of the technologies, and the frequency with which we updated the Twitter feed and Marginalia Machine. Our motivation for these designs was to manifest the liveness of the process of archiving within our designs but a tight binding to this process was in conflict with the technical and experiential concerns we outlined. However rather than seeing this as a compromise, we suggest that archival liveness is embodied equally in these three facets and that weight should be given to each within the design process. Following Freeth and Tanaka [Freeth, Bowers, and Hogg 2014] [Tanaka 2000] we prefer to see these different temporal qualities as features of material interest. We observe that the adoption of and experimentation with such features of materials is entirely characteristic of Research Through Design processes and indeed one aspect that differentiates them from other design practices.

By distinguishing between the working rhythm of the staff in the archive and the various constraints and affordances of the server infrastructure we reflect an obvious historical tension between the working patterns of people and technologies design to support them. Rather than representing this distinction as a binary one however, our approach based in archival liveness nuanced our understanding of the particular kinds of temporality in different parts of what might be thought of as an ecosystem of work and technology. The pursuit of archival liveness, as we have said, brought us into closer contact with the archivist and digital assistant as well as the library technical services team. With this close contact we gained unusual access to the way their work was organized across a diverse technical infrastructure. We observed a disjuncture in the temporal patterns of the archival process and the way that the archive was accessed by our participant researchers. Where the archivist and digital assistant worked to long time frames with a tendency to batch processing and repetitive activity, the participants were inclined to “dip in and out” of the archive irregularly. Our intervention through Twitter in some senses mimics this comparison. After running the bot for several weeks, we noticed that in periods when digitization was taking place, tweets would commonly appear around 5pm; seemingly the digital assistant placed scanned PDFs in the “complete” folder as one of her last tasks of the day. The appearance of the tweeted marginalia image on followers’ Twitter timelines however will be interspersed with other material. What was a semi-regular event for the digital assistant manifested as sporadic for the followers of the bot.

As well as presenting design opportunities, we note that our particular engagement also created a kind of “gearing in” [Durrant 2011] [Gurwitsch 1979] to the archiving process and the technologies around it. Durrant et al. [Durrant 2011] referencing Gurwitsch [Gurwitsch 1979] describe a scenario wherein visitors adjust to the infrastructure of a theme park as it aims to present a particular mode of experience. The process by which the visitors begin to socially organise their visit around this infrastructure is one of “gearing in”. In the context of the archive we can see the progression of our design interventions from the cultural probes to box log connections, to the Twitter bot and Marginalia Machine as ways of learning to “mesh” our designs with the work of the archive. Our treatment of archival liveness was in many ways a process of orientating ourselves to the various rhythms of the archiving process and the materials available at each stage. One practical advantage of this “gearing in” was that we were able to produce engaging, timely designs at stages of the project earlier than might have been otherwise possible. More significantly, this gearing in gave us access to a rich set of materials and concepts, prompting designs that offer new modes of engagement and reflection on the archive.

In this project we were focused on the liveness of the archiving activity principally because of our desire to engage users with the archiving work. However by highlighting the archiving process we were also creating a public performance of the Bloodaxe Archive as live. Liveness was being presented as part of the public identity of the archive. In this presentation, we find direct accord with other digital heritage initiatives. As archive catalogues are aggregated into web services and interfaces, a number of interesting points arise regarding the permanence and stability of archives and their expression through catalogues and interfaces. Application Programming Interfaces (APIs) such as [National Library of Australia] now provide aggregated access to the collections of thousands of institutions. The constantly changing nature of the collections defined by the scale of such aggregations suggests a particular sense of the live archive (as growing, more available, more diverse) in contrast with the “dead” one (as complete, finished and closed). The National Archives of Australia recently launched “SODA” [National Archives of Australia] a service which publishes the most recently digitized records from across their aggregated resources. The developers make the point that the constantly changing state of the catalogue makes it difficult to know which new records are available. By facilitating easy access and sharing the Archive’s aim to convey a sense of the rate at which the collection is growing [National Library of Australia] they hope to engage potential users in the new content. As in our case, liveness is implicitly part of the way the public identity of the archive is performed.

Similarly, Circus Oz maintains a “Living Archive”  [Circus Oz] of video and photographs of its performances since its inception in 1978. Through their “stories” tab, they emphasize the connection of the archive to the lives of the performers and audiences. By doing so they suggest that the archive is also live in the sense that is engaged with the daily lives of the people represented in it.