what's this ?
what's this ?
excerpts here
excerpts out
peer review
Click on a tag above to see relevant excerpts from this site.
Click on a tag above to see relevant excerpts from other articles in the mesh.
Search this article for any word:

The innocuous-looking wall label--featuring a single artist, title, date, medium, dimension, and collection--represents a cultural paradigm based on singularity and stasis and rather than multiplicity and movement. The most dynamic art of the past half-century will die if this paradigm isn't overturned.

The gravest threat to the cultural survival of new media art may very well be its wall label. Few manacles on creativity have been as ubiquitous. Employed by curators everywhere, the wall label, along with the catalogue caption, has been joined in the past few decades by a younger generation of digital descriptors, the collection management record and online citation. Together this typographical dynasty has conspired to reduce every artwork, from the street happening to the stick spiral, to a single artist, date, medium, dimension, and collection.

While the reductionism of the wall label enfeebles conceptual and single-performance art, it threatens to obliterate digital culture completely. For new media art can survive only by multiplying and mutating. From computer-based installations to video multicasts, digital collaborations are the rule rather than the exception, and a work often undergoes changes in personnel, equipment, and scale as it diffuses across new media festivals, exhibitions, and web sites. Like a shark, a new media artwork must keep moving to survive.

Unfortunately, many de facto custodians of culture -- museum curators and conservators -- are ill equipped to maximize an artwork's adaptability, because their job usually seems to require that the artwork remain static. To safeguard the rich legacy of artistic media born of the digital and Internet revolutions requires something more than storing an artist's web site as a data file on a Windows-formatted CD-ROM. Within twenty years, the browser to read the data will have become obsolete; within thirty, the only CD-ROM drive may be in a vitrine in a computer museum; within forty, Windows will be dead media, and within fifty, the CD itself will have delaminated. For digital culture, fixity equals death.

Yet fixity is what wall labels impose on artworks in any form of new media. Consider the de facto standard for these labels as applied to a representative work of media art:

Nam June Paik
TV Garden, 1974
Nineteen 36-inch SONY video monitors, fifteen 21-inch SONY video monitors, and twelve 5-inch Magnavox video monitors; five pairs of speakers; DVD player; three video distribution amplifiers; cables; DVD; wood, soil, and 187 live potted plants of various types
Installed in an arc 1261.2 cm long, 639.8 cm wide, and 122.6 cm high
Solomon R. Guggenheim Museum
Purchased with funds contributed by the International Director's Council and Executive Committee Members: Ann Ames, Edythe Broad, Henry Buhl, Elaine Terner Cooper, Dimitris Daskalopoulos, Harry David, Gail May Engelberg, Ronnie Heyman, Dakis Joannou, Cindy Johnson, Barbara Lane, Linda Macklowe, Peter Norton, Willem Peppler, Denise Rich, Simonetta Seragnoli, David Teiger, Ginny Williams and Elliot K. Wolk, 2001

While the absurdity of this label may be obvious to anyone who has actually created or installed a work of new media, to many curators and archivists such labels are a reassuring echo of time-honored conventions for documenting paintings and sculpture. Unfortunately, in adhering to such conventions a registrar squanders time recording evanescent details of the installation while neglecting information about the work critical to its presentation and preservation. The video artist Bill Viola has remarked that museum staff who condition-check his work assiduously note every fingerprint on his video decks but fail to notice when a functional component like a speaker or transformer is missing or obsolete.[1] Eyes trained in traditional conservation are not necessarily prepared to see what matters in new media installations, where adaptability and change are the means, rather than an obstacle, to survival.2

Wall labels are the pins that fix the butterflies of new media to museum walls. We need to pull out those pins if new media works are to thrive. In this essay I explain how, drawing inspiration from the open code movement and two structures currently under development -- the Variable Media kernel and The Pool collaborative environment -- that are designed to keep the butterfly in motion. These structures remove the conceptual blinders that prevent curators from realizing that works can have more than one an author, title, date, medium, dimension, and credit line.

This section looks at three different approaches to recognizing creators of a work: by teams, interpreters, and other models.

If collaborations are the rule rather than the exception for new media art, you would hardly know it from looking most museum labels. They rarely cite an artist's assistants or technicians. To be sure, there is a difference between a carpenter or programmer whose work is perfunctory or fungible and one whose stylistic or technical contribution is inseparable from the aesthetic result.[2] Hidebound cataloguing systems, however, reduce such nuances to a stark "yes or no" choice, and cannot describe how the cast of characters for an evolving work may change over time.The most common concession to the authorial fluidity of new media projects may be the banding of artists together under a single name. This practice is so pervasive that it is hard to find a directory of new media artists that isn't full of group monikers; for example, the artist index for the letter 'e' in Christiane Paul's book Digital Art consists of "Electronic Caf© International," "Electronic Disturbance Theater," "Entropy8Zuper!," and "etoy."[3] In some cases, the anonymity of a name serves a group's strategic purpose; for example, the gorilla masks that diffuse accountability for the activist Guerrilla Girls have online equivalents in the anonymous corporate identities assumed by the tactical media groups etoy, ark, and the Bureau of Inverse Technologies.[4]

Yet in a dozen years of working in and with artistic collaboratives, I have never been comfortable with group names for artistic production. It's easy to forget that they mask the relationship and interaction of the participants, substituting a putative group mindset for a more believable scenario of dissenting or compromising individuals. Yes, some group memberships fluctuate too much to be published as a fixed set of names; the etoy that created the Digital Hijack in 1996 included Hans Bernhard, while the etoy that created DAYCARE in 2001 did not. If our working assumption is that collaborators change over time rather than remain the same, all such contributor lists should be dated. I have seen, in the worst cases, the instigator of supposedly "egoless" productions getting all the credit for the work of his nameless collaborators.[5]

Recently I have been struck by how comfortably a group name can dovetail with the single-genius paradigm beloved by art history and the art market. To say that a work was produced by KOS or Group Z is not all that different from saying it was produced by Tim Rollins or Micha«l Samyn, because even if critics and collectors know only the collective name, they can still imagine that the work issues from a single studio grooving to a single mindset. The irony is that many such groups deliberately chose collective attribution to avoid the art star system. Unfortunately, however, a single appellation obscures a nuanced view of the author stream for a given work, which may be multi-user, nongeographic, and asynchronous.

One documentation system designed to document, and indeed encourage, such fluid author streams is The Pool. A project conceived by Joline Blais and myself at the University of Maine's Still Water program[1], The Pool is a shared resource for online art, code, and texts assembled by and for creators of new media. Surfers who access The Pool's artworks and texts are invited to contribute factual and evaluative information about them. The accrued ratings of a work determine where it appears in the sea of projects floating in The Pool's primary interface, where works of high recognition or approval rise to the top. Contributors can propose a concept for others to implement or respond to invitations to explore, debug, or remix the existing works they have viewed. In place of the single-artist, single-artwork paradigm favored by the majority of documentation systems, The Pool stimulates and documents collaboration in a variety of forms.In one of our meetings to develop a new release of The Pool, three of my collaborators, John Bell, Matt James, and Justin Russell, argued against allowing Pool users to substitute a group nickname as shorthand for a list of collaborating creators. They convinced me that this authorial convenience, commonplace among new media artists, does not suit a vision of distributed creativity. So, artists, when you start a collaborative project by asking your partner, "What are we going to call ourselves?," resist the urge to go with Menudo or %20 or Myst3r10us H4q0rz. The names on your passports -- or your favorite pseudonyms -- will do fine. Forcing everyone else to enumerate those names when they refer to your project will remind them how impoverished the single-artist paradigm is for describing networked culture.[6]

In my work with the Variable Media Network (, a consortium of museums and archives devoted to innovative approaches to new media preservation, one of the most promising strategies we have found for keeping digital art alive is to re-create it on a completely different platform.[7] Somewhat naively, we first hoped this would be a matter of hiring a programmer to write an appropriate emulator to run, say, CP/M PASCAL on Windows XP. However, we quickly found that the aesthetic issues involved in emulating peripherals or pacing are just as daunting as the technical ones; the programmer you hired to emulate the display may end up deciding such critical visual elements as color depth, screen resolution, and tempo. Most new media curators wouldn't be surprised to learn that an artist's studio assistant plays a role in making such decisions, as long as the artist was consulted about the result. But that consultation isn't going to happen when the artist is long dead.

In some cases, the re-creator's role may even exceed that of a studio assistant and become analogous to that of a performer. For example, software artists such as Mark Napier have recommended that future preservationists reinterpret their Java applets by re-programming them in whatever the default interface language may be for, say, the year 2050. Such reinterpretations require a combination of aesthetic fidelity and individual creativity; their closest analogues are probably a pianist playing Bach or a troupe performing Shakespeare in mukluks and miniskirts.

If re-interpreters of some works get too little credit, re-interpreters of others seem to get too much. Take La Monte Young's Composition 1960 #10 (to Bob Morris), a proto-Fluxus work that consists merely of the instruction, "Draw a line and follow it." By far the best-known interpretation of this score is Nam June Paik's 1962 performance, when he dunked his head (or on other occasions, his tie) in a bucket of ink and then dragged it across a paper scroll on the ground. Unfortunately, when catalogue essays or captions cite Paik's performance, they mention the title by which it has come to be known, Zen for Head, but neglect to mention any role La Monte Young played in its inception. Fast-forward to 1992, replace the ink with hair dye and the scroll with the floor of London's Anthony d'Offay Gallery, and you have Janine Antoni's performance Loving Care -- a work whose citation never mentions Paik or Young despite the obvious lineage.

Of course, there are legitimate reasons for giving Paik and Antoni full credit for their interpretations of Young's work. Both artists literally threw themselves into the work, and their cultural preoccupations -- Paik's audacious take on Zen artlessness, Antoni's flair for extrapolating feminine qualities to absurd excess -- came along for the ride. Furthermore, while music critics would cringe if a conductor played a Mozart adagio at a presto tempo, Young's conceptual score was loose enough to permit a wide interpretive license.

Nevertheless, just as it would be misleading to credit a recording of Beethoven's Fifth Symphony by mentioning only the conductor, not the composer, so it is misleading not to give Young credit for Paik's performance and Young and Paik credit for Antoni's. Many critics believe it unlikely that the art world, online or off, will ever jettison its dependence on what Foucault described as the "author function."[8] And they are probably right. But can't we rewrite Foucault's function to accommodate networked creativity? The art world, if left to its own devices, will always fix on a single creator, excluding any antecedents or successors. So what other devices can we suggest for acknowledging them?

Listing the collaborators' names, separated by commas, is not an ideal solution; such lists grow unwieldy for large collaboratives (does anyone really read the names that scroll by on the Photoshop launch screen?), and they fail to capture the cultural context that can be conveyed by a term like Fluxus or RTMark. The new media artist Cory Arcangel has suggested a compromise: "When I make a work that uses the BEIGE aesthetic, I simply grant ownership of the work to myself and the group at the same time:"[9]

Cory Arcangel / BEIGE

Another model is suggested by the variable media strategy of reinterpretation described earlier, where both "composer" and "performer" are listed:[10]

Nam June Paik, Zen for Head (1962). Reinterpretation of La Monte Young, Composition 1960 #10 (to Bob Morris)

Finally, The Pool offers an "expandable" author function, which can be as short as the list of names responsible for an intent or as long as a breakdown of the roles every participant played --conceptual, technical, or perceptual -- in the artifact's many versions. The relationships of these authorial moments are captured in PHP and MySQL on The Pool server -- thus rendering literal Foucault's metaphor of the author "function."[11]

The collaboration between the animator Walt Disney and the conductor Leopold Stokowsky now known as the film Fantasia was originally entitled The Concert Feature.[12] Until the choreographer Martha Graham proposed a new name on opening night, Aaron Copeland's score Appalachian Spring was entitled Ballet for Martha. Artists often change their intent in the process of making their work, so it's no surprise that they might want to change the work's title as well.

In more open collaborations such as the free software movement, project titles can mutate multiple times. As long as it was a commercial web browser, the Netscape brand name varied only slightly from its launch in 1993 to the dismantling of the company in 2003. The open-code projects spun off from Netscape, by contrast, have spawned a slew of names, beginning with the Mozilla browser, which was publicly released in 2002.[13] Within two years, five alternative versions of Mozilla appeared with new names, including the Chimera browser for the Macintosh, renamed in 2003 the Camino browser, and the 2002 Phoenix 0.1 browser, which in version 0.6 (2003) became the Firebird browser, which in turn by version 0.8 (2004) became the Firefox browser.

Although these browser names have sometimes been changed to avoid trademark overlaps, at other times a critical mass of developers decides to fork the original project in a parallel but different path. This freedom is one of the reasons many coders choose to work on open rather than proprietary software development. Not that this freedom makes them any less invested in the results of their labor; on the contrary, the hundreds or thousands of developers at work on open code projects often want control over the name for their project as well as the code that runs it. Emotions run so high about names that Mozilla's developers devoted an entire FAQ web page to "Mozilla Firefox Brand Name," with headings from "What's a Firefox?" to "How was the community consulted?" to "But I hate the new name. It's stupid."

Sometimes the proliferation of names can seem ridiculous even to open-code advocates. To parody the seemingly endless variation of open source browser names, the developer Michael O'Rourke wrote a Firebird/Firefox extension called Firesomething.[14] Firesomething generates a random name for the browser every time it launches, rewriting the title bar with new names like Mozilla Firedog or Mozilla Moonpossum.

Designing and building a collaborative application or artwork is hard enough without the worry of getting the title right on first go, especially when the name has to satisfy a community of collaborators and steer clear of brand confusions or trademark infringements. That is why The Pool's architecture encourages creators to float many ideas and get feedback quickly. Only after collaborators have settled on an idea they would like to implement should they have to brainstorm a clever marketing handle and search for potential clashes of trademark or copyright.

To avoid focusing too early on names and other changeable aspects of a project, The Pool distinguishes between the initial conception of an artwork and its subsequent versions. The Pool's structure is designed to make it easy to track the "wake" left by a contributor's idea, as it gets picked up by new artists or rendered in new mediums or is accessed by different users with different technologies over subsequent years. To encourage adaptable projects rather than those carved in stone, The Pool tracks three phases of any project: intent, approach, and release. A code, art, or text "intent" is a suggestion for a project, typically a verbal description or evocation. An "approach" is an interpretation of an "intent," typically a mockup or proposal in a presentation format like Photoshop or PowerPoint rather than the final medium. Many different approaches can fork off the same initial intent. "Releases" are the subsequent attempts to fulfill that intent through the approach in question. Releases are prototypes or finished projects, implemented in the medium intended for public distribution.

For example, a handful of students at the University of Maine floated an idea for a social networking software in The Pool that carried the descriptive if less than mellifluous title UMaine's People. In the process of designing an approach to go with this intent, the students came up with the new title Team Tooler. How should this change be indicated for students who originally knew the project as UMaine's People? As in the case of multiple authors, a telescoping format may be the most practical solution. In this case, an bibliographic and museological convention could accommodate the change: a secondary (in this case, original) title in parentheses after the primary title:

Jeremy Knope, Shaun Leeper, Timothy Oliver, and David Phillips, Team Tooler (UMaine's People).

As indicated in the previous section, The Pool offers an expandable view of a "project stream"; the more detailed format would reveal, for example, that Tim Oliver devised the intent of UMaine's People but Jeremy, Shaun, and David joined him to produce the approach and release.

Team Tooler by Jeremy Knope, Shaun Leeper, Timothy Oliver, and David Phillips
Based on UMaine's People by Timothy Oliver

One of the biggest dilemmas curators of conceptual, performative, and media art face is determining which date to write on the wall label. Some artists insist, perhaps on the advice of their dealers, on the year of the original work -- or even of its conception. As misleading as it may seem to date a plywood box back to 1961 if it was hammered together yesterday, it is equally misleading to cite only the year of a refabrication or new variant without reference to its history.

Again, open-code collaboration -- whose unofficial motto is "Release early and often"[15] -- offers a different approach. If the art market favors unique objects whose historical patina dates them to a particular moment in time, the community of software users and developers yearns for ever-more shiny new releases. The only way to keep track of constantly morphing code projects is to index new iterations by date (a practice commonly used for commercial software such as Microsoft Windows 95/98/2000) or number (the practice preferred for open software such as Mozilla version 1.0/1.1/1.2).

A good test of such a scheme for new media artworks -- as opposed to the monolithic dates common to traditional labeling -- would be a work that has changed repeatedly since its inception. Fortunately, works for which a single year is a misleading reduction are everywhere on the new media landscape. One such work is Apartment, software art by Marek Walczak and Martin Wattenberg with Jonathan Feinberg. When visitors to the Apartment web site type in sentences of their own, the program draws an architectural floor plan -- based on the connotations of the individual words -- on the screen. In its first three years, Apartment has already seen dozens of revisions in programming and design, some behind the scenes (as when Wattenberg rewrote the algorithm to run more efficiently on older computers) and some visible to the user (as when the team built a physical interface on the occasion of an installation at the Whitney Museum of American Art). Furthermore, Apartment archives such user-created apartments into aggregate "cities," which means that even if the artists cease to modify the project, daily contributions from its users will continue to modify the work long into the future.

Such mutations in code and appearance are a necessary consequence of adapting to the new media landscape, just as mutations in an organism's genotype and phenotype are a natural and necessary consequence of adapting to a changing ecosystem. The difference between nature and new media is that the evolution for the former happens in the course of millennia, while new media change in a matter of months or weeks. New media artists and technicians are used to this ferocious pace of media turnover, but unfortunately, the curators and archivists charged with capturing an artwork's vital statistics are not. As a result, screenshots of Apartment a year or two apart may look substantially different, but a standard catalogue caption will treat them identically:

Marek Walczak and Martin Wattenberg with Jonathan Feinberg
Apartment, 2001

One way of treating this problem is to adapt a more nuanced numbering system derived from software development.[16] For example, the first public variant of Apartment, launched on on 12 February 2001, might be described as "variant 1.1" -- that is, the first public release of the work. The artists tweaked this variant at least eight times in the two weeks that followed the launch, mostly to change the appearance of rooms or cities. By 25 February, they had released variant 1.8, which would remain essentially unchanged except for user input until the following June. In the meantime, however, the artists were invited to install Apartment in the Whitney Museum's Data Dynamics exhibition; to do so they had to imagine and build a physical interface for Apartment that could sit in a gallery. This resulted in variant 2.1 -- the change in the initial number indicating a significant change in the user's experience of the work. The Whitney variant was eventually reinstalled a year later in Eindhoven, The Netherlands: variant 2.2. To date there has been one more significant variation on the original work, a two-person interface installed at the Ars Electronica Festival in Linz, Austria: variant 3.1.

Unlike numbering versions of software, in which version 2 supersedes version 1, and 3 supersedes 2, the existence of Apartment 3.x doesn't make Apartment 1.x obsolete. Instead, there are three parallel streams for the work. Lower numbers in a series of new media variants will most likely become technologically obsolete before higher numbers -- but that will not make them aesthetically obsolete. In fact, the main difference between numbering versions of software and of new media works is that the latter gives user experience priority over software and hardware implementation.[17]

Yet even variant numbers can't keep track of the endless mutations of a work like Apartment, which may be altered daily or even hourly by new contributions from online or gallery visitors. Database-driven projects like Apartment that vary continuously cannot be assigned release numbers; one way to account for this mutability in a screenshot or photo caption would be to add the city or web domain and day of the reproduction:

Apartment v1.1 (, 12 Feb. 2001)

Apartment v2.1 (New York, 22 Mar. 2001)

Apartment v3.1 (Linz, 1 Sep. 2001)

In the appendix to this essay I list twenty-two different variants of Apartment numbered and dated according to this scheme. Most have been public releases, but the list includes a few development stages (variants 0.1-0.3) that never saw the light of day.

To keep captions and other "tombstone" data as simple as possible yet acknowledge that new media works evolve differently than static artifacts is a difficult tradeoff. But only a nuanced versioning can counteract the misapprehension that new media works like Apartment spring into life fully formed like Athena from Zeus's thigh. Revealing how a work grows and develops in collaboration with new media venues over time reinforces the importance of such incubators for new media projects, and reminds us that the circulation and exhibition of new media works contributes in no small part to their maturation and evolution.

The variable-media approach, as its name suggests, invites creators to imagine how a work might be translated into a new medium in the future once its current medium expires. To capture these artistic intents for future reference, the Guggenheim's variable-media task force conducts interviews and workshops with artists based on a questionnaire that asks which aspects of a work may change and which may not. We have begun to input the results of these questionnaires into a multi-institution database, so that artists' differing perspectives on the long-term maintenance of their work can be shared and compared.

When we first conceived of the questionnaire, we tried to work within familiar art-historical categories such as photography, film, and video. We quickly realized, however, that medium-specific pigeonholes were as transient as medium-specific artworks; as soon as video became obsolete, so would a video-based prescription for re-creating an artwork. Furthermore, as soon as another medium came along -- which happens every ten minutes, it seems, in the age of the Internet -- we would have to add a new category. Finally, categories based on mutually exclusive media wouldn't accommodate hybrid works such as Ken Jacob's Bitemporal Vision: The Sea (1994), which merges film and performance. To circumvent this problem, we decided to explore medium-independent, mutually compatible descriptions of each artwork, which we call "behaviors."

Some artworks, for example, must be "installed" -- not in the ordinary sense of requiring a nail hammered in the wall or a pedestal lugged into a corner, but in the special sense of changing every time there is an installation. For example, Nam June Paik's video installation TV Garden (1974) has been shown in rectangular galleries, on a curving ramp, and in indoor swimming pools. For TV Garden, the Variable Media Questionnaire asks about such preferences as the ideal installation space ("fine art or museum gallery"), lighting requirements ("as dark as code allows"), and distribution of elements ("mass of television and plants should be in a 1:4 ratio").

Other works must be "performed". Most of the questions for this behavior -- whether the props are disposable, where the audience sits -- assume that the work has a theatrical or musical setting. According to the variable media paradigm, however, the term "performed" can apply whenever the re-creators have to reenact original instructions in a new context. For example, to construct Meg Webster's Stick Spiral (1986), the artist asks museum staff to find recently fallen branches from the local environment; in addition, the artist's ecological ethics require that the branches must have been pruned for some reason other than the exhibition. Stick Spiral's fabricators spend more hours in a pickup truck exploring back roads than in a gallery stacking brushwood. This extra dimension means that Webster's spiral is both installed and performed.

The questionnaire also requests information on artworks that are reproduced, duplicated, interactive, encoded, or networked. When the variable media task force first established these behaviors, we were tempted to divide artworks according to analog versus digital media; that distinction, however, was too imprecise to account for the variety of formats present in contemporary art. Instead, we chose the term "reproduced" for any medium that loses quality when copied, including analog prints, photographs, film, audio, and video. For these works the essential questions pertain to who owns the master, whether it is an etched copper plate, a silver-gelatin negative, or a first-generation U-matic videotape.

In contrast, we reserved the word "duplicable" for media that can be cloned, such as the Java applets and Web browser required to view a work such as Mark Napier's net.flag (2002). At the same time, we realized that nondigital works such as Felix Gonzalez-Torres's Untitled (Public Opinion) (1991) can also be duplicated; a museum could produce two indistinguishable versions of this candy spill simply by ordering identical piles of licorice from the manufacturer. A section of the questionnaire asks which forms of distribution are acceptable for duplicable works, whether digital or analog.

Both Public Opinion and net.flag are also meant to be "interactive"; museum visitors can take free candies from the Gonzalez-Torres, and online visitors can modify Napier's flag by adding or subtracting parts of the flags of various nations. Among the important questions for interactive behavior is whether traces of previous visitors should be erased or retained in future exhibitions of the work.

A work is "encoded" if some form of computer programming or annotated score is used in its construction. Some encoded works, including net.flag, are also "networked", distributed across an electronic communications grid such as the Internet, but others, including Grahame Weinbren and Roberta Friedman's interactive video The Erl King (1985), stand alone as sculptures or installations. The Guggenheim partnered with the online arts resource to come up with the essential questions for encoded and networked artworks, including which screen resolution is optimal and whether the programming code is open or closed source.

Partners in the Variable Media Network can choose to extend or multiply these modular behaviors as the need arises. In a recent example, Guggenheim Senior Conservator Carol Stringari raised the point that even paintings and sculptures can provoke prickly questions when some aspect of their construction requires a change. To account for these alterations in otherwise stable media, we added a ""contained" behavior to the questionnaire, asking, for example, whether an oxidized surface should be cleaned or a damaged frame replaced.

It is all well and good to have a medium-independent prescription somewhere in a collection management system, but if the public is to come to terms with this important aspect of variable media works, these prescriptions will have to make it out of the hidden recesses of a museum database and into spotlighted wall labels on the gallery walls. Although writing "variable media" makes the jobs of museum curators and catalogue editors easier, the phrase does not suffice for a complex installation like TV Garden. The goal is not to take the path of least resistance, but to propose an informative standard of description that won't need to be scrapped and rebuilt every few years -- even if the artwork itself has to be.

Fortunately, the medium-independent behaviors of the variable media paradigm offer a middle ground: a vocabulary for writing a medium line so that at least one part of it never changes. One solution, for example, is to concatenate the behaviors that apply to a given work into a readable phrase :

Meg Webster
Stick Spiral
Performed installation with duplicable materials

Nam June Paik
Zen for Head
Performance with duplicable materials

Ken Jacobs
Bitemporal Vision: The Sea
Performance with reproduced film and duplicable hardware

Nam June Paik
TV Garden
Installation with reproduced video and duplicable hardware and materials

Felix Gonzalez-Torres
Untitled (Public Opinion)
Interactive installation with duplicable materials

Grahame Weinbren and Roberta Friedman
The Erl King
Interactive installation with code, reproduced video, and duplicable hardware and materials

Eva Hesse
Expanded Expansion
Installation with contained materials

Mark Napier
Interactive networked code

It is important to convey the behavior of these works, rather than their material, in a wall label or caption; after all, the material should be fairly plain from the piece or illustration accompanying the text. Nevertheless, this shorthand should be able to telescope when more detail is required, with the understanding that the second part of the medium line may vary with the version of the work.

Nam June Paik
TV Garden
Installation with reproduced video and duplicable hardware and materials: Forty-six monitors; five pairs of speakers; three video distribution amplifiers; DVD and DVD player; wood, soil, and approximately 180 live potted plants of various types

Grahame Weinbren and Roberta Friedman
The Erl King
Interactive installation with code, reproduced video, and duplicable hardware and materials: SONY computer ca. 2004, bitmapped video frames, PASCAL source code and Java interpreter, two monitors, two speakers, and wood and cardboard construction

Mark Napier
Interactive networked code: Java applet with server-side text files

I learned a valuable lesson about institutions and intransigence during preparations for "Preserving the Immaterial," the first conference on variable media, which took place at the Guggenheim in March of 2002. The announcement card featured an installation photograph of Meg Webster's 1986 Stick Spiral, for which I wrote a caption that included the line, "Branches, dimensions variable." Webster built her installation from branches recently felled from local trees; as their flowers and leaves slowly withered and dropped to the floor, the shape of the overall spiral changed. My museum colleagues and I had recently established the convention "dimensions vary with installation" for works whose shape changed from one exhibition to another but was static over the course of each exhibition, such as a Richard Long rock installation designed to fill a room; we reserved the expression "dimensions variable" to describe the subset of such works -- including Webster's -- whose shape also changed over the course of the exhibition.

This caption came back from the museum's editorial staff with a revision that I found particularly telling. In place of the phrase "dimensions variable" was a precise specification of the installation's height, width, and depth down to the quarter-inch -- figures no doubt painstakingly noted by a conscientious registrar standing on a ladder, straining to extend a tape measure to the exact position of the topmost leaf. Although those just happened to be the measurements on a random day of the work's previous installation, the edited caption gave the impression that this unruly bramble was always meant to be re-installed to those exact specifications -- a requirement both unattainable and inconsistent with the artist's intentions.

When I pointed out the irony of including such an error in an announcement for the first conference on variable media, the editor told me that there was nothing she could do, for she had found the dimensions in the museum's collection management database, and her staff had instructions to normalize all captions according to that standard.

After much debate, I managed to wrestle back my version the caption by going over the editor's head, but I learned two lessons in the process. First, standards work only if they are worth upholding; and second, the fiercest battlegrounds are often the minutiae of daily life --which are generally presided over by the humblest functionaries of an organization's hierarchy.[19]

During a panel discussion on the preservation of duplicable art at "Preserving the Immaterial," the first conference on variable media in 2001, the artist and Berkeley Art Museum / Pacific Film Archives curator Richard Rinehart pointed to a fundamental conflict between the credit lines prized by museums and the preservation of culture that is supposedly central to their mission:

For museums that are now beginning to accession [duplicable media] into their permanent collections, we have an opportunity to engage a preservation strategy which we have never had before, even with film and video and photography.

And that is a preservation strategy which is commonly used in the world of information technology in relation to digital information, and that is data redundancy. So museums are used to collecting these unique one-of-a-kinds, even if it's a contractual agreement of exclusivity. But as a preservation strategy for digital information, including digital art, a really good strategy is back it up, create multiple copies of the digital information, and then distribute those geographically....

But it brings up two important problems for museums and artists, and that is first of all, there is this competitive edge of museums; we pride ourselves on our unique collections. So how would we as museums address this? How would we brand ourselves, if not on our unique collections?

Secondly, for the artist, it poses the problem of economic models. If you can't write up that contract promising you'll make no more copies of the video, and thus charge 80,000 dollars for a videotape, and you're selling it to fifteen museums, what's the new economic model?[20]

Rinehart's questions about distributed archives and economics were a few years ahead of their time -- museum time, anyway -- but recent initiatives suggest that the answers may be forthcoming. The Museum of Modern Art in New York, the San Francisco Museum of Modern Art, and London's Tate Gallery have been negotiating to build a shared collection of media art. Although it has a precedent in the national collections of certain countries, this move flies in the face of the consumerist logic of museum branding[21]. Nevertheless, in a time of shrinking acquisition budgets for nonprofits and growing reliance on specialized technical expertise for preserving media works, the MoMA-SFMOMA-Tate consortium is poised to turn the duplicability of media art from a liability to an asset. An important question for such consortia is how to ensure that the responsibility for caring for a work in variable media is shared, and that lower profile works do not slip through the cracks because there is no single caretaker charged with maintaining them. Fortunately, with increased square footage and geographic distribution comes increased opportunity for exhibition, and as noted in the discussion about variable dates, exhibitions are frequently the ideal occasions for checking conditions, refurbishing, and / or re-fabricating ephemeral works.

An initiative called the Open Art Network, meanwhile, has been exploring the answer to Rinehart's second problem: new economic and legal frameworks that encourage artists to distribute duplicable works. Many artists want to maintain control over source elements of their work during their lifetime: photographic negatives, video masters, Java source code, or the rights to modify or redistribute online works. Yet it is crucial for those same artists to realize that their legacy will be lost to history if those video masters are lost in a fire or their source code becomes corrupted before being transferred to a public trust. Digital media are the most vulnerable, since crates don't have delete buttons but computers do.

The Open Art Network is exploring with such artists the legal possibility of deferring access to source materials. According to such an agreement, a video artist might deliver to a collector or museum a duplicate master along with the artwork, with the understanding that the artwork's owner cannot access the master until the artist gives permission or dies. A neutral third party could serve as an artistic escrow account, holding artists' source code until the time when a need for open access outweighed their proprietary interest in keeping it secret. There is some precedent for this in the custom software industry, where owners of a software copyright put their source code in escrow with a third party, so that a licensee can access it if the owners go out of business. In the case of an artwork, it may not be a licensee who gets access, but cultural organizations -- online or off -- or the public at large.

If the cramped conventions of current cataloguing systems drain new media art of its essential variability, what snippet of art-historical data could possibly convey the rich micro-history of a new media work? If new media and its culture of mutability ever successfully infiltrate the hidebound customs of museums and archives, I am betting that wall labels and captions will look different from current ones in a couple of respects. As media artworks from the twentieth and twenty-first centuries are increasingly installed without the participation of their original creators, it will become important to credit the re-creators. The misleading use of years alone to signify a chain of events will be replaced by a sequence of versions. And medium and dimension lines will contract or expand as the context permits. If all these predictions come to pass, a page out of a future art history text might look something like this:

TV Garden v1.1 (Kassel, 1974) by Nam June Paik
Based on Global Groove v1.3 (New York, 1973: single channel of reproduced video) by Nam June Paik
Variable installation with one or two channels of reproduced video and duplicable hardware and materials; shown: U-matic videotape and player with color and sound; thirty monitors and three pairs of speakers; wood, soil, and approximately 50 live potted plants
Dimensions variable; shown installed in a rectangle approximately 1.5 meters high, 6 meters long, and 10 meters wide.
Collection of the artist

TV Garden v1.12 (New York, 2000) installed by Nam June Paik, Blair Thurman, and Jon Huffman
Based on TV Garden v1.1 (1974, Kassel) and Global Groove v1.3 (New York, 1973: single channel of reproduced video) by Nam June Paik
Variable installation with one or two channels of reproduced video and duplicable hardware and materials; shown: DVD and DVD player with color and sound; forty-six monitors and five pairs of speakers; three video distribution amplifiers; wood, soil, and approximately 180 live potted plants
Variable dimensions; shown installed in an arc approximately 1.5 meters high, 25 meters long and 6 meters wide.
Solomon R. Guggenheim Museum, New York
Purchased with funds contributed by the International Director's Council and Executive Committee Members: Ann Ames, Edythe Broad, Henry Buhl, Elaine Terner Cooper, Dimitris Daskalopoulos, Harry David, Gail May Engelberg, Ronnie Heyman, Dakis Joannou, Cindy Johnson, Barbara Lane, Linda Macklowe, Peter Norton, Willem Peppler, Denise Rich, Simonetta Seragnoli, David Teiger, Ginny Williams and Elliot K. Wolk, 2001.

TV Garden v3.4 (New York, 2030) installed by Cory Archangel, Jr.
Based on TV Garden v1.1 (1974, Kassel), Global Groove v1.3 (New York, 1973: single channel of reproduced video), and Allan 'n' Allen's Complaint v1.6 (New York, 1982: single channel of reproduced video) by Nam June Paik
Variable installation with one or two channels of reproduced video and duplicable hardware and materials; shown: SONY computer ca. 2030, two channels of bitmapped video frames with color and sound, TurboJava code; Seventy monitors and ten pairs of speakers; wood, soil, and approximately 300 live potted plants
Variable dimensions; shown installed in a circle approximately 2 meters high and 30 meters in diameter
Berkeley Art Museum/Pacific Film Archive, Solomon R. Guggenheim Museum, New York, and ArtBASE

If these hypothetical captions enter strange or complicated territory compared to the predictable standards of painting and sculpture, that's because the works they represent do, too. The first job of wall labels is to educate; it is time for them to start doing their job, by taking a form that reflects the new curatorial realities they describe.

The people at work on preservation instruments such as the variable media questionnaire and production environments such as The Pool are deliberately working to accommodate a more supple paradigm than the wall label's informatic straightjacket of a single author / title / date / medium / dimension / collection. Only by deposing the wall label and its kin can new media art hope to survive into the future.

As nature teaches, in any swiftly changing ecosystem there is safety in numbers (think of spawning fish), in adaptability (think of amphibian DNA, which turns male or female depending on water temperature), and collaboration (think of the clown fish and anemone, or the tickbird and rhino). Media art that is capable of all three means of self-preservation will flourish in the media ecology of the twenty-first century. Media art that fits snugly into a wall label will not.

Variant History of Apartment by Martin Wattenberg, Marek Walczak, and Jonathan Feinberg[22]

Apartment v3.2 (Eindhoven 28 Feb. 2002)

[Second release of third variant, incorporating physical installation with two-user interface, installed at MU in Eindhoven, The Netherlands.]

Apartment v1.10 ( 7 Feb. 2002)

Turbulence was hacked into. Jason and Jonathan put Apartment back together again. They are so great. There are now 5,050 apartments organized in 14 cities.

Apartment v3.1 (Linz 1 Sep. 2001)

[Third variant, incorporating physical installation with two-user interface, installed at Ars Electronica Festival in Linz, Austria.]

Apartment v1.10 ( 19 Aug. 2001)

Archived 10th city. There are now 4057 apartments.

Apartment v1.10 ( 1 Aug. 2001)

For those interested, there are now 3,680 apartment stored on the Turbulence website.

Apartment v1.10 ( 30 July 2001)

Created one archived city in one day. 760 apartments!

Apartment v1.10 ( 19 June 2001)

The level of porn has increased -- I just find it so boring. We now run a script that periodically removes these from the current city. We'd like to make a porn-city for these, but have no time and money. We would be thrilled to do so should someone pay us!

Apartment v1.9 ( 11 June 2001)

We finally split up the city, as it was getting too crowded.

Apartment v2.1 (New York 10 June 2001)

Yesterday was the last day of the Data Dynamics show at the Whitney Museum of Art. For the last few days, the printing and saving functions didn't work. We couldn't figure out why. Today I noticed that the server's keyboard had been pushed under the monitor, so permanently pressing down the 'Escape' key. And that was why....

Apartment v2.1 (New York 22 Mar. 2001)

[Second variant, incorporating physical installation with single-user interface, installed at Whitney Museum of American Art in New York for Data Dynamics exhibition. This is the only variant to date that allowed visitors to print out their apartments.]

Apartment v1.8 ( 25 Feb. 2001)

We adjusted the city, hopefully for the last time.

Apartment v1.7 ( 24 Feb. 2001)

A few kind guinea pigs came over to test the interface. As a result we: made the latest apartment persistently orange in the city-view, put in the days of the week in the city-view, made 'snapping' words in the apartment view not orange (they looked clickable). Martin optimized the code so it should run faster on older machines.

Apartment v1.6 ( 23 Feb. 2001)

We changed the apartments so they consist of rectangular rooms. Its better... especially in 3D - now you get both long vistas and close-ups. So the effect is more spatial.

Apartment v1.5 ( 22 Feb. 2001)

The new radial city is up! We think its far clearer than the previous variant. The city is viewed by time and content.

Apartment v1.4b ( 20 Feb. 2001)

The city view continues to be difficult to understand. Also, we've been wanting to get time into it for ages. So we're going to try a different city map entirely: the weighting of each apartment on various themes will govern distance to the center of the city, with time as the radial locator. Testing this evening.

Apartment v1.4 ( 19 Feb. 2001)

There was a problem with the sounds in the 3d. As they originate from the images themselves in large spaces you don't hear them -- too far away. Jonathan fixed this.

Apartment v1.3 ( 17 Feb. 2001)

One apartment appears to take over half the city, Martin makes a limit to the size an apartment can appear in the city.

Apartment v1.2 ( 13 Feb. 2001)

Apartments are organized in the City depending on the sizes of the various rooms. We are constantly amazed that by simply looking at the words used we can identify content. We decide to be clever and name the identifiers by related neighborhoods of Manhattan. This confuses everyone - though its supposed to make things clear! We get rid of the neighborhoods and mark content instead.

Apartment v1.1 ( 12 Feb. 2001)

Apartment opens. We choose 9 'seed' apartments to place in the city. Maybe after a few apartments are inserted, we can take these out.

Apartment v0.3 ( Fall 2000)

Added the "city" view.

Apartment v0.2 ( Fall 2000)

[This variant generates floor plans that looked more like actual apartments.]

Apartment v0.1 ( Fall 2000)

A variant that took in words and created a floor plan using a map-of-the-market style layout [a rectangle filled with grids of proportional size].[23]


[1] Other contributors include John Bell, Margaretha Haughwout, Matt James, Jerome Knope, Kristen Murphy, Justin Russell, Mike Scott, and Owen Smith.

[2] These examples include artists and titles only in their most abbreviated form; see the final section for citations that demonstrate all the recommendations suggested in this essay.


[1] See accounts of the 2000 SFMOMA symposium "Techarcheology," in the Journal of the American Institute for Conservation, 40, no. 3 (Fall/Winter 2001), p. TK.

[2] As Christiane Paul pointed out in her review of this manuscript, some technicians may be hired to accomplish tasks that the artist cannot do for reasons of time rather than qualification.

[3], accessed June 10, 2004.

[4] Thanks to Christiane Paul for reminding me of this important role group monikers can play for activist artists.

[5] Probably the best-known example of one new media artist getting recognized for other artists' work was the series of Blast "vehicles" organized by Jordan Crandall of the X-Art Foundation. Crandall borrowed the terms Blast from a Vorticist periodical and X-Art from Vittore Baroni's Arte Postale!, though his concept of collective publication owes more to the 1968 SMS portfolio (see, accessed June 10, 2004.) organized by the American Surrealist William Copley, with Lew Syken and Dimitri Petrov. SMS included multiples by brand name artists such as Roy Lichtenstein, Man Ray, and Claes Oldenburg, as well as by many artists who are all but unknown today, such as Bernard Pfreim, George Reavey, and Clovis Trouille. In his twist on SMS, Crandall offered his collaborators a vision of communal publishing in which the absence of labels was supposed to thwart the artworld's preoccupation with authorship; but over time, Crandall's own name became indelibly associated with the X-Art Foundation and its Blast vehicles, and his solo career was inadvertently blessed by this second-hand recognition.

[6] Pertinent here -- but beyond the scope of this essay -- is the fact that merely citing multiple collaborators does not mean the work itself sheds any light on the collaborative process. See Janet Cohen, Keith Frank, and Jon Ippolito, "Sentences on Adversarial Collaborations," 1994, reprinted in Cohen, Frank, Ippolito, The Argument Drawings, exhibition catalogue (New York: Wynn Kramarsky, 1997), mirrored online at, accessed June 10, 2004.

[7] This approach, known as emulation, is one of four strategies accommodated within the variable media paradigm, the others being storage, migration, and re-interpretation. To store a work is to archive it in a crate or on a disk; to migrate is to update its material, equipment, or software to a contemporary version; and to re-interpret the work is to replace it with a functional or metaphoric equivalent with no necessary resemblance to the original. For more information on these strategies, see

[8] For example, Chris Chesher on the Still Water-Eyebeam "Distributed Creativity" list: "One tendency has been to substitute collective authors for individual authorship: groups like Anti-ROM or VNS Matrix. But I suspect the author-function isn't going anywhere fast!", accessed June 10, 2004.

[9], accessed June 10, 2004.

[10] This caption is drawn verbatim from John G. Hanhardt and Jon Ippolito, The Worlds of Nam June Paik [exhibition catalogue] (New York: Guggenheim Museum, 2000), p. TK.

[11] For example, to display the author stream associated with one of its projects, The Pool executes a PHP script that spiders through all of the contributors associated with all of the versions of a project. Depending upon how that script is coded, those authors could be ranked by the quality or quantity of their contributions.

[12] Despite his critical contributions to the film's development, conductor Leopold Stokowski's name rarely appears in any mention of the "Walt Disney's film" -- yet another case of the market's fixation on single authorship. Is it mere coincidence that the project excited Stokowski enough to conduct it for free, while Disney dismissed it as a financial failure but imposed the strictest copyrights over the film anyway?

[13] Dates and versions from and, accessed June 7, 2004.

[14], accessed June 10, 2004.

[15] Eric S. Raymond, "The Cathedral and the Bazaar," v. 3.0,, accessed June 10, 2004.

[16] Or more accurately, configuration management. Jeff Rothenberg, RAND computer scientist and consultant to the Variable Media Network, suggested this protocol as a versioning scheme that doesn't stigmatize iterations with lower numbers.

[17] I'm using the term "variant" in place of the common term "version" to reinforce the presumption that newer releases are better than older ones.


[19] Museums tend to ossify any variable aspect of an artwork's installation, not just its dimensions. See the transcript from the "Preserving the Immaterial" conference for a discussion of the color of Robert Morris's Labyrinth, online at, accessed July 24, 2004.

[20] Remarks at the 2001 conference "Preserving the Immaterial,", accessed July 24, 2004.

[21] For example, France's Fondation national d'art contemporain (Fnac) is an independent collection of international art that can be loaned to individual museums in France or in other countries.

[22] Source: Email exchange between the author and Martin Wattenberg on 19 May 2003 (with additional information from, accessed May 19, 2004.

[23] Martin Wattenberg, Smart Money's Map of the Market,, accessed May 19, 2004.