The Qualities and Significance of Documentation

Annet Dekker (Amsterdam) and Gabriella Giannachi (Exeter)





The acquisition of performative, time-based, and digital artworks, which are often hybrid, ephemeral, and subject to change, has led to significant changes in practices of documentation, collection, and preservation for museums and other cultural organisations. Funded by the AHRC, and a collaboration between University of Exeter, London South Bank University, LIMA, The Photographers’ Gallery and The Australian National University, The Documenting Digital Art project (2019–23), explored how digital art has induced museums and other cultural organisations to reflect about what, despite these changes, had remained undone, undocumented, un-conservable. Four workshops organised by LIMA, Amsterdam[1], one of the project partners, offered an insight into the work carried out by museums and cultural organisations not only to understand these artworks’ components and technical requirements, as well as the parameters for their (re)activation, but also to imagine how these artworks might transform over time, and consider whether these processes of transformation would affect them beyond recognition. Reflecting about what and when to document, here, we suggest that thinking of these artworks through a series of axioms, namely process, interaction, relationality, pervasiveness, and decay, may generate useful knowledge for understanding and potentially creating their documentation. In what follows we are analysing the artworks based on their documentation qualities. In focussing on how documentation is created and used for rehearsal and reactivation, we noticed that the significance of documentation shifts from a past-facing documentary bearing witness about an ‘original’ artwork, to the use of documentation as (part of) the artwork, indicating that documentation is becoming a generative method to enable further iterations of existing works as well as the development of new artworks.



Documentary documentation

Many performative, time-based and digital artworks are conceived of and developed in relation to societal factors which both shape and, on occasion, respond to the artwork. This means that, over time, it is possible to become unaware of specific relational qualities of an artwork, struggling to remember the ‘original’ context for its production, exhibition, and conservation. Here documentation can bring together the different contextual elements of an artwork. In the workshops this was exemplified in two specific ways: as documentary instructions, images, or videos that showed the functioning of the artwork, and as an (online) publication that was established retrospectively to further document the documentation.



(1 + 2) C. Dullaart, (2011), screenshots


Constant Dullaart presented his (2011), a website that displayed a Google start-page which was partly obscured by different filters. Visitors to the site can use the search function, but each time the page is reloaded, a new filter appears, erasing some of the information previously displayed by distinctive brushstrokes. Originally the artist used iframe, an HTML element that can be used to embed a website frame into another website, but Google later blocked the iFrame, so Dullaart used a dedicated proxi. The artwork was blacklisted by Google for the potential phishing threat it posed.


The server-side of, i.e. the domain name, files, and source code, is stored in a system of virtual servers called ArtHost at LIMA, and, in collaboration with Rhizome, the team wants to document the client-side of the artwork (the operations that are performed in the browser). Furthermore, inspired by the Dullaart-Sakrowski method [Spreeuwenberg et al. 2012], LIMA used a split-screen recording to show the artwork’s interactivity via a screen recording and an off-shoulder video shot that shows the artist interacting with the online artwork while he speaks about the specifics of the artwork (based on a set of questions that was shared with the artist beforehand). The documentation brings together different aspects of the artwork: showing at the same time the experience of the artist interacting with the website and his explanation of the process and context, as well as the different relational aspects of the artwork. As such, it is close to the conventional documentary methods in which the video shows the artwork in a factual and informative manner.


Naked on Pluto



(3) Dave Griffiths, Aymeric Mansoux and Marloes de Valk, Naked on Pluto (2010-2013), Screenshot


Such a documentary approach is not unusual and is also used in another case study that was discussed: Naked on Pluto (2010–13) by Dave Griffiths, Aymeric Mansoux and Marloes de Valk. Here the single channel video shows elements of the gameplay while it is interspersed with quotes and descriptions from the artists who highlight the context of an artwork. However, Naked on Pluto’s documentation approaches were manifold. As part of LIMA’s more general research into the preservation of digital art, Mansoux, together with Julie Boschat Thorez (artist and researcher) and Dušan Barok (artist and researcher), tried to come up with a documentation strategy to connect these elements of the documentation.


Naked on Pluto is a research-based, open source, multi-player, participatory online game that consists of a game in Facebook, essays, blogs, a series of interviews, workshops, installations, and interviews with players conducted by a bot. The artwork, which mixed traces from various profiles, and saw players team up to try to crash the system, caricatured the proliferation of virtual agents harvesting personal data through Facebook. When the artwork was exhibited, the artists explored how to transform the artwork into a physical experience, and they decided to use one of the main objectives in the game: the Plutonian Library. As part of the artwork and to contextualise the topic, the artists also interviewed twelve experts to map their perspectives on social networks and data privacy. After several exhibitions and winning the international VIDA 13.2. prize, the game stopped functioning after a couple of years due to the depreciation of Facebook’s Application Programme Interface (API) – a piece of software that pulls data from one place to another. Yet, the core of Naked on Pluto (an open source game engine) and the accumulated documentation about the artwork continued to serve as an impulse for a wider examination of privacy in the age of social media through exhibitions, workshops, lectures, interviews, books and websites [Dekker 2018].

In an attempt to gather the loosely connected documentation that was scattered across the web, the team wanted to integrate and connect different parts of it in one space. During the workshop, Barok, Boschat Thorez and Mansoux suggested that when approaching Naked on Pluto as a discursive artwork its documentation happened along four axes which consisted of: 1. A dispersed archival repository that included the digital and physical material related to the artwork, i.e. articles, files, data, etc.; 2. A series of public manifestations that ranged from public lectures to workshops and exhibitions; 3. A narrative that favoured the presentation and experience of the game over its preservation; and 4. A description of the context in which the artwork was created and presented to situate the artwork in its time and space, to provide a possible future key to its comprehension. These offered information about the process, how people interacted with the work, and the work’s relationality.

Barok, Boschat Thorez and Mansoux used a wiki as a publishing platform to present these different axes on separate pages. They used the notion of “preservation by publication” to emphasize that preservation of the artwork was not based on its medium but rather on the conceptual impetus, which in this case was a problem-driven artwork, also referred to as artistic research. Although the artwork clearly focused on technical development and social media platforms, and hence could be seen as digital art or net art, choosing one medium as the primary focus for the preservation strategy would merely address one specific aspect of the artwork. This would neglect the main intention of the artwork: its discursive and research-based approach in which different aspects, elements and forms of interaction came together. When considering the preservation strategy for all these disparate elements, a wiki turned out to be the best format. The wiki allowed Barok, Boschat Thorez and Mansoux to see the various elements of the artwork not as represented in documentation forms, but as articles, whose templates could be adapted to suit the particular needs of artworks, contributors may rearrange sections, change headings, embed media files and so on [Barok et al. 2019]. Currently, they divided the wiki into three main parts: the first page presents the narrative presentation, the second introduces the different parts of the artwork in a structured way, and the third one provides the index, an inventory of the artwork’s archive. Next to these there is an introduction page that explains the rationale of the set up.

Some still argue the relevance of video to document an artwork. However, the shifting of the emphasis from merely a technical or functional overview to include comments from the artists and potentially others involved in the artwork, such as those who experienced it, is a strategy to better understand the context in which the artwork thrived. Such a form follows the conventions in documentaries to document primarily for the purposes of instruction, education, or maintaining a historical record. Naked on Pluto also followed this approach, but the team extended it by bringing together the dispersed archive that was left after the artwork ended. The intention was not to embed or copy the actual material that was hosted on numerous websites, rather they followed a post-custodial practice, in which archiving objects moves beyond those in custody of an institution, thereby shifting the balance of power in preservation. Linking to and contextualizing the ‘original’ sources instead of archiving them on a new server has the benefit that the material remains in its original environment, at least as long as those websites are operational. Publishing as preservation is therefore closely connected to the documentary format, albeit in a new medium. Moreover, here the resulting inventory was also set up to give an extensive overview that could be used as a basis from which new artworks or narratives could be developed.



Documentation (as part) of the artwork

The allographic nature of performative artworks, which often do not entail a script, or set of instructions, means that there can be some degree of variation between different performances of the same piece, especially when different performers are involved. Often, the most valuable knowledge about the ‘identity’ of an artwork, to use Pip Laurenson’s term [Laurenson 2006], is obtained not so much through the documentation of any one individual performance but through the documentation of the rehearsal process leading to the performance, which not only captures the artist’s intent but also shows how the artist negotiates the realization of the artwork in relation to the capabilities of the performers.


Your Face is / is not enough

The case study that best illustrated the value of the documentation of the rehearsal process was Kevin Beasley’s Your Face is / is not enough (2016) which was commissioned for the Between the Ticks of the Watch exhibition at the Renaissance Society in Chicago in 2016. The case study was presented by Tate, Ana Ribeira, Time-based media Conservator at Tate, and Louise Lawson, Conservation Manager at Tate, which had acquired the artwork between 2016 and 2018. Your Face is / is not enough is a performance-based artwork consisting of twelve gas masks with megaphones sculpted by the artist. The masks, akin to those used by the military and the police, are heavily altered by the artist and represent a take on an object usually worn for protection but modified here as a way to comment on contemporary society. The performers are ideally local to the performance site and are meant to come from different racial and ethnic backgrounds, without prior experience of performance. Prior to the performance, several stands are installed in the museum space. When the performers enter the gallery, they attach the hand-held voice receiver to the megaphone and the mask and carry out a series of deep breaths which have a strong sound effect akin to that of a siren. The sequence is repeated several times over twenty-five minutes. At the end, the performers look at the audience, immobile, and then take their masks off before leaving the space. The masks are then left installed in the gallery.

At the point of acquisition, Tate received a video documentation provided by the artist gathered before the exhibition at the Liverpool Biennial in 2018 and during the post-activation, comprising documentation collected during the exhibition, including instructions on how to care for the masks and engage with the performers. The masks are not replicable, which means that when they have deteriorated, the performance can no longer take place. Additionally, Tate documents the artwork each time it is exhibited. The performative nature of the artwork prompted Tate to place an effort in gathering the documentation of the rehearsal process in which the artist was describing the artwork in ways that had not been captured elsewhere, evidencing the importance, in this case, of capturing the process but also decay of the artwork.

Tate started to acquire performance art in 2004. Since then, it developed the Performance Specification Tool which is built around 7 core themes that include flexible questions within each of them focused on what is needed for the artwork to persist over time. Rather than reviewing or analysing what the performance was, this approach included considerations for the possibility that the artwork might change. The documentation tool Activation Report is a record of stakeholders involved in the activation, considering possible changes to be implemented for the activation of the artwork. Information from the Activation Report can feed back into the Performance Specification. In the case of Your Face is / is not enough, the documentation of the process captured knowledge beyond any one performance that was to be used to transmit key instructions to future carers of the artwork. This contextual knowledge is crucial for the activation of performance-based artworks.[Lawson, Finbow and Marçal 2019]


Agent Ruby


(4) Lynn Hershman Leeson, Agent Ruby, 1999-2002, screenshot. Commissioned by SFMOMA; Collection SFMOMA, gift of bitforms gallery, Paule Anglim Gallery, and the artist; © Lynn Hershman Leeson


Interestingly, what could not be captured was what would happen to the artwork once the masks could no longer be used. This, paradoxically, and in a reversal of usual dynamics in performance practice, created a situation in which the ephemeral part of the artwork was not the performance, but the prop used within it, as it was in fact the prop that defined whether the performance could be activated. This increasingly more ambiguous role of the objects and documents that are part of a performance also became clear in the next example of Lynn Hershman Leeson’s Agent Ruby (1999–2002) in which audience documentation became part of and could be said to even substitute, or become, the artwork.

Digital artworks are often experienced very subjectively, which means that to capture artworks that are interactive it is important to document how people have been interacting with and relating to them. Ideally, it is important to preserve user-generated documentation, as they often offer additional insight into an artwork. This means that interactive artworks consist not only of the artwork itself but also of their history of engagement, which in turn can become (part of) the artwork. The value of the documentation of the user experience and its importance for the artwork presentation became crucial in the case of Lynn Hershman Leeson’s Agent Ruby, which was presented at the LIMA workshop by Layna White (Director of Collections SFMOMA), Rudolf Frieling (Curator of Media Arts SFMOMA), Mark Hellar (Technology Consultant, Hellar Studio) and Grace Weiss (Assistant Registrar, Media Arts, SFMOMA).

Agent Ruby consists of an artificial intelligence web agent whose identity was shaped by its interactions with users. The artwork’s latest iteration consists not only of the custom-made code of artificial intelligence embodied by Ruby, the online bot that chats with users that ‘seek’ it, but also the history of Ruby’s interactions with users over the years. The artwork originated from Hershman Leeson’s film Teknolust (2002) in which a bio-geneticist models three self-replicating automatons after herself. Ruby is one of the automatons. With Agent Ruby, Hershman Leeson created a new kind of experience, one in which the artwork, Agent Ruby, grows to reflect its interactions with others.


Agent Ruby is among the first web-based artworks commissioned by SFMOMA and was originally presented by the museum in 2002 on, the museum’s online exhibition platform (2002-2008). SFMOMA subsequently acquired the artwork in 2008, taking on its long-term maintenance, display, and preservation. The SFMOMA staff at the workshop highlighted how the museum’s care of the artwork includes the website, code, design prototypes, and the expanding chat log of Ruby’s conversations with users (2002–present). The documentation of Agent Ruby’s history of engagement is therefore not just for future consultation regarding the artwork’s preservation, but has literally become part of the artwork and was installed along the artwork when it was last exhibited at the museum as The Agent Ruby Files in 2013.

While the artwork originally may have consisted of the code and bot, over time its identity changed, emerging through the conversations with the people who sought answers from her, so much so that the records of these conversations now form part of how the museum thinks about the artwork itself. The artwork still exists online, but it can no longer grow, so in a sense, to seek what it was, users now need to go back to the documentation of its past interactions with the public. This shifts the status of documentation which not only bears witness but also becomes part of the artwork. Agent Ruby and The Agent Ruby Files are therefore at once the same work and different iterations of it, illustrating the role documentation plays in defining the identity of an artwork.



Generative documentation

Digital artworks often evolve pervasively over prolonged periods of time. An artwork that started as a digital artwork, might evolve to include not only user interactions but also different formats and manifestations, such as performances, talks, and even research papers. These can be defined as convergent, sometimes even pervasive, by which we mean that they add up to more than the individual elements that form them, over a prolonged period of time, and they can often, quite deliberately so, not be grasped in their totality.



(5+6) Martine Neddam, (1996–), screenshots


The case study that best illustrates the value of the documentation of the convergence of pervasive artworks is Martine Neddam’s (1996–) presented at the LIMA workshop by Mila van der Weide, Assistant Conservation and Documentation, LIMA and Patricia Black, Research Intern, LIMA. is an interactive website evoking the diary of a 13-year-old girl, Mouchette, who expresses her thoughts on death, desire, and suicide. The artwork consists of several image and text compositions which can lead to different paths of interaction, including web questionnaires and fora. The artwork’s participatory nature presents documentation challenges compounded by the fact that the artwork included live performances and installations. Hence the artwork has an expansive and pervasive character, and its documentation is based on people’s memory of events, as well as documents of the site itself. In this sense, Mouchette is both a performative site and its own archive.


It can be said to have an ‘identity’ that manifests itself across multiple forms, times, and spaces. Hence, the Stedelijk Museum (Amsterdam) developed a documentation strategy that consists of the iterative mapping of various convergent dimensions of the artwork, including a warning system when links become broken. Interestingly, the Stedelijk Museum acquired a time-stamped version, which emphasises that the artwork does not so much have an original, as it constantly grows, and changes literally across artforms. Capturing the convergence of these components over time required a pro-active type of conservation. Neddam herself refers to the concept of ‘generative preservation’. This term also questions the ambiguity, or perhaps emphasises the inaccuracy, of notions that are commonly used when referring to an artwork, such as the ‘original’ and ‘unique’. In generative preservation the distinction between the original and its subsequent iterations is hard to make because earlier parts of the artwork and its documentation can be re-used as part of a new iteration of an artwork.


Colored Sculpture

Something similar happens in cases where performative, time-based and digital artworks are built so that their decay forms part of the life of the artwork. This suggests that the artwork is not only meant to be ephemeral, but also that part of its life cycle is that the artwork could be allowed to self-destroy. This is particularly visible in Jordan Wolfson’s Colored Sculpture (2016), which was presented at the LIMA workshop by Patricia Falcao, Time-based media conservator, Tate. Colored Sculpture is a large-scale animatronic installation in which a two-metre hight puppet is dragged across the gallery floor and then smashed against it in a loud and violent self-destructive act against the sound backdrop of Percy Sledge’s ‘When a Man Loves a Woman’. The chains holding the cartoon-like puppet are driven by motors usually used for boat anchors and the puppet’s eyes are custom-made video screens following visitors as they move around the space. Colored Sculpture is continuously damaged by its activation, raising major questions as to its future exhibition plans as well as its durability in time once its components decay and become obsolete.


At the time of acquisition, Colored Sculpture was heavily documented both in terms of the conditions for its production and its conservation. The acquisition included a very precise set of instructions created by the artist, including hardware and software manuals, video materials, an installation plan, and schedule. These were taken over by Tate in their regular conservation practices, but they also included the different stakeholders and suppliers, and the possible changes induced due to obsolescence and decay over time. At the workshop, Falcao pointed out that as the artwork is very theatrical, from a conservation point of view, it was crucial not only to document its components and operating system (including the possibilities for its emulation) but also the visitor experience, for which video was used. While this has become a more common practice, as we noted above, Tate also used its collection management system, The Museum System (TMS), to record the artwork’s documentation, as well as a conservation wiki created by Duncan Harvey, to document processes adopted in the day-to-day care of the artwork. The different steps were structured in such a way that they could easily be traced when new staff members would need to conserve the artwork. This shows how the museum has taken steps to identify the ‘network of care’ [Dekker 2018: 71-98] for this work, offering insight for the long-term care of the artwork once its components have decayed beyond use.


When analysing the challenges of collecting performance art for the museum Laurenson and Van Saaze concluded that: “It is not the problem of non-materiality that currently represents the greatest challenge for museums in collecting performance but of maintaining – conceived of as a process of active engagement – the networks which support the work” [Laurenson and Van Saaze 2014: 29]. Similarly, for digital art preservation this could mean that next to organizing the management within the organization it requires the management of the extended network around the artwork. This could lead to what Neddam described as generative preservation: a system of smart contracts, in which the different parts of the evolving artwork could be tracked via the block chain mechanism. Such a network of interconnected documents can easily track the different elements and stakeholders of, yet it could also contain the building blocks of the future of Importantly, Neddam stresses that this is not about re-interpretation, which she regards as inherent in working with the web [Neddam 2016]. Instead, and in the case of, what generative preservation aims for is to preserve the database while keeping all the projects and its evolving interactions and documentation active, thus, acknowledging “the human input (in creation and maintenance)” stressing “notions of growth and expansion instead of repetition or replication” [Dekker 2018: 93]. With this, documentation becomes a propagating, generative mechanism, taking artworks into the future.



Performative, time-based and digital artworks tend to change over time, sometimes decaying beyond recognition. Often they are defined by the interactions carried out by users. On occasion, their iterations become more pervasive. These artworks are interesting not only as objects but also as processes, including what happened before and after the life of the artwork in the public domain. Museums and cultural organisations have developed increasingly complex documentation methods, tools and platforms, which intend to capture the artwork’s identity but also instruct museum staff about how to deal with its iterations over time. In this sense, as we suggested in Documentation as Art (2022), documentation is both past and future facing. Many of these artworks can be described through a series of axioms to do with process, interaction, relationality, pervasiveness, and decay. These axioms have led to the production of three types of documentation, i.e. documentary documentation; documentation as part of the artwork; and generative documentation which bring out the past, present, and future of the artwork. Finally, such an expanded notion of documentation challenges the conventional object-oriented approach of art. Artworks evolve over time, and so do their documentations, bringing up the question whether the terms ‘artwork’ and ‘documentation’ are still appropriate or if it would be better to refer to an ‘art environment’ to emphasise the ongoing evolvement of both that is bigger than the sum of its parts.




Dekker, Annet. Collecting and Conserving Net Art. Moving Beyond Conventional Methods. London 2018.
Dekker, Annet and Gabriella Giannachi (eds.). Documentation as Art. Expanded Digital Practices. London 2022 (forthcoming).
Laurenson, Pip. “Authenticity, Change and Loss in the Conservation of Time-Based Media Installations.” Tate Papers. No. 6, 2006. 26.05.2022.
Laurenson, Pip and Vivian Van Saaze. “Collecting Performance-Based Art: New Challenges and Shifting Perspectives.” In Performativity in the Gallery. Staging Interactive Encounters, edited by Outi Remers, Laura MacCulloch and Marika Leino. Bern 2014: 27–42.
Lawson, Louise, Acatia Finbow & Helia Marçal. “Developing a strategy for the conservation of performance-based artworks at Tate”. In: Journal of the Institute of Conservation. 2019.
Neddam, Martine. 2016. “Generative Preservation.” Notes from conference presentation at LIMA. 2016. Online: 26.05.2022..
Spreeuwenberg, Kimberley et al. 2012. “Documenting Internet-based Art: The Dullaart-Sakrowski Method”. 26.05 .2022.


Download PDF