In Search of Meaningful Events

Curatorial Algorithms and Malleable Aesthetics

Andy C. Deck

So much of the contemporary experience of the World Wide Web is searching and sampling of the unfinished, insufficient precursors of what one may expect tomorrow. In the climate of "cool" that prevails, most are occupied with looking, searching, and waiting rather than with dialogue, synthesis, derivation, and playful production. In fact the trend is encouraged: the more pages of "search results" we scan, the more advertising we see. The importance of the software that shapes the public's use of the Internet extends beyond aesthetics. Yet the concentration of authority around the mass market Internet portals engages discursive formations that have emerged in recent years about art and the public sphere, institutional critique, and freedom of speech. Artists are posing questions about the limits and function of the arts in the sphere of the Internet. Insofar as public institutions treat their Web sites as publicity tools­­glorified pamphlets­­one must turn toward the margins of creative production for clues concerning latent alternatives. Fortunately the margins and centers of the Internet continue to mingle, especially in a technical sense.1 So one can still entertain some marginally central questions: If Internet art overturns the paradigm of broadcast and allows reciprocal communication between transmitter and receiver, what becomes of curation? Is the curator then a moderator? An artist? A programmer? Malleable aesthetics provoke such questions.

The cybernetic systems I have in mind may prove incompatible with typical museum practices (selection and display of objects that are made prior to their presentation). The malleable aesthetic transforms hypertext and hypermedia: its essence is profound reconfigurability in response to feedback from interested participants. Whereas the beauty of most existing hypertext and hypermedia art is supposed to reside in the masterful interplay of prospective narratives, wired in by the author; the allure of malleable aesthetics is the potential digression toward and development of almost any direction.

Some of the most intriguing microcosms of malleable aesthetics to date have been textual interactive systems that privilege the familiar signification of natural language as writing. The call and response of textual dialogue, whether at the pace of conversation or in a slower, gradual accumulation of related ideas ("bulletin boards"), offers a suggestive model for the articulation of meaning online. There are dozens of varieties of dialogic text systems that embody different approaches to the control of events and interface. One bland form, chat, imposes a potential anonymity, sometimes resists the use of one's preferred language, and favors people who can type. But usually there is little or no censorship. While this sort of communication may foster subtle, shared experiences and ideas, usually there is no lasting record, except in the memories of the readers and writers.

comic chat
Microsoft's Comic Chat.
Drawings by Jim Woodring.

Variations of the chat interface involve the categorization of discussion topics, and the ability to whisper to (or exclude) some readers. Due perhaps to the transience of this type of speech, "freedom of chat" remains controversial only among the most domineering hosts. Microsoft and AOL use the popularity of chat to leverage their power as marketing venues. Despite differences in their revenue models, both of these "providers" are developing a base of users for online consumer colonies. That users have no profound, programmatic control of the underlying software greatly limits the buildable potential of such systems. Public feedback used in this way becomes a disposable byproduct in the pursuit of market share. It seems clear enough that developments which do not serve this pursuit will not flourish in regulated commercial software spaces, where feedback is reduced to what Hans Magnus Enzensberger has termed the "lowest level compatible with the system."2

There is little cause to think of communicative reciprocity as it was envisioned­­every receiver a transmitter. The phrase seems quite dated, appearing to have more to do with ham radio than with the cagey mediation of software. To dwell on the possibility of two-way communication through computer networks, while dismissing the industrialization of coercive software, is to hold intellectual freedom in very low regard. Likewise it is suspect to assume that the appearance of persistence, in the form of the database, is a panacea for the airborne transience of broadcast. The proliferation of databases is certainly an interesting development, but one that poses new problems, too. For in the superabundance of choices, there emerges another regime of alienation. The acquisition of search engine databases is proving instrumental in the centralization of influence and authority around a few mediocre, distraction-filled Internet portals.

Given these conditions, how can the public meaningfully and intelligently affect the purposes and qualities of reciprocal Internet systems? As an artist using the Internet, the question of how to involve people in meaningful events is paramount. Inspiring participation in something useful or fun, or enlightening is okay. But better still is orchestrating contributions to something good that lasts longer than the event itself, adding an historical dimension. And yet, because computer systems obsolesce so quickly, the desire to create permanence is forcefully resisted. Instead, evolutionary mechanisms hold sway: one can aspire to affect characteristics of the next generation of network use. Ultimately it is through revision and adoption of alternatives that cybernetic systems are transformed­­qualities emerge or reappear in the successive iterations of code. To what extent will the public be included in this process of becoming? The diffusion of control over public interactive systems leads to malleable aesthetics.3

Currently interfaces that give people opportunities to contribute in a cumulative or editorial way are usually absent from the fortified Web sites of corporate media (and museums). Embracing open participation lets loose a Pandora's Box of expressive liberties to an anonymous public. Most institutions are reluctant to pursue Web ventures that cannot insure a degree of decorum. Fearing what might happen, they front Web sites that inexpensively advertise their stuff. In so doing, they contribute to the mutation of the Internet into an elaborate yellow pages catalog. Here, public institutions follow the lead of the marketing sector, rather than standing with artists and intellectuals who would advance a different culture than the familiar authoritarianism of commercial television.

Clearly this is not to advocate institutional support for puerile uses of the new media. The amazing growth of the Web is evidence of the will of people to leave a mark--aesthetical, rhetorical, or otherwise­­and it is with this in mind that one must ask whether museums, and other public institutions, will endorse or reject that will. Insofar as the museum on the Web will become more than a repository of support materials referring to a permanent collection, it must address the issue of feedback. Is it wanted? With what limits? How might it be used?

Situating the museum and curator in the network of public collaboration, it may be said that control or moderation of "input" can take the form of preemptive interface design and programming, concurrent moderation by a living participant, or retroactive exclusion, censorship, and editing. Concurrent moderation is the unlikely union of broadcast and the chaperone. It is difficult to imagine this form of network activity becoming a social problem. As a means of controlling people it does not, as the buzz phrase goes, scale well. More disturbing are automated forms of restriction. They prevent and obscure critical judgment. That there exists a technical solution able to resolve the political difficulties of real public participation is a myth of instrumental reason. If an interface stops the cacophony of chat and the banality of "my first homepage" it might also prevent the ideal contribution, whatever that may be. Software should not be used like a Kafkaesque bureaucracy to hide the trail of responsibility and decision making. In the excited stir of innovation, political designs, masquerading as bureaucratic or technical necessity, mould the parameters of online participation and expression. As these reductive formulations of human ingenuity proceed apace, institutions need to be criticized for their complicity. Instead of resisting the response of the public, and serving as apologists for mediocre software, and instead of abstractly ascribing responsibility to "technology," institutions need to become more accountable. Due to the manipulative capacity of interactive systems, designs should be open to revision and debate. In moving away from the stuffy, cosmetic Web sites that have sprung up all over, the need for more transparent control policies is evident.

Paradigmatic of this openness are the multi-user, object-oriented "MOO" projects that are derived from role playing multi-user dungeons (MUDs).4 Typically, the people who use the MOOs most become the policy architects and chief content providers. It is not unusual for a leadership committee to consist of a geographically dispersed group of interested users. As with chat, the transience of utterances tends to make censorship a rarity, but in those situations where someone's behavior becomes really disruptive, generally, a joint decision is made concerning the best response. This process, which ranges from democratic to oligarchic in practice, invokes a governmental, consensual protocol for the harmonization of online discourse. This kind of governance has predominated on the MOO systems over the last decade, and in practice it allows a wide latitude for expressiveness.

Regardless of whether the MOO participants consider themselves artists, conflicts of authority arising from the regulation of collaborative, Internet art can be arbitrated in this way. Rather than preemptively or automatically stifling behavior, it encourages participants to resolve their own differences, as tends to happen outside of digital channels. Although not ideal, this model can also provide a flexible mechanism for institutions to defray controversy over online content: the responsibility for settling disagreements is delegated.

MOOs differ significantly from chat for several reasons. For one thing, persistent character and environment descriptions give continuity to repeated visits. Although anonymity is an option, the metaphorical space of a MOO becomes familiar, as do the various personalities. Regular visitors can contribute many types of "objects" to the database­­descriptive, programmatic, automatic, responsive. Users confront one another at the boundary of natural language and programming, but both novices and "wizards" can meet and converse in this border zone. People with modest programming skills add features (scripting, "rooms") to the MOO for others to use. Nor is it unusual for the underlying source code, which can be obtained freely, to undergo changes. The resulting synergy, by which the space is gradually reinvented, opens onto unanticipated paths. The term "malleable aesthetics" as I mean it refers to this ability to accumulate not only statements, or data, but also the structural changes wrought by users of the system. Incompatible with forced enclosure, the purest forms of this category of production are licensed to assure that programming code remains in the public domain.

Hybrid interfaces have appeared that attempt to give a more visual form to MOO personas and places. This has been tried in numerous ways without widespread success. Although networked, user-directed constructions seem possible with multimedia (the World Wide Web itself?), creating graphical avatars and imagery for visual MOOs demands some uncommon skills­­the public is generally unfamiliar with programming languages and authoring tools. Networked action games like Quake, while in some respects graphically responsive, impose a certain thematic inevitablility. Whereas MOO thematics span from quest games to postmodern culture, the graphical hybrids remain mired in layers of graphical preconception. Even if, through simulation, one is able to traverse a game "world," this remains the labyrinthine navigation of non-malleable hypermedia. In the arena of the reality engine, extension of default boundaries founders at the counterintuitive interfaces for level editing, image processing, solids modeling, and motion control. As a result gimmicks that offer superficial choices abound. But whatever virtual wallpaper one selects, a shoot 'em up game is still a shoot 'em up game. The challenge latent in all the hybridized multi-user systems is the retrieval of narrative and visual thematics from engineers, programmers, and the competing demands of commodification in the marketplace of software and services.

Probably it will not be long before microphones and digital video cameras explode the dominance of the mouse as a device for contributing to spontaneous, networked audio-visual events. One can imagine low budget, "on the scene" reporting being produced by independent documentarists equipped with little more than a video camera, microphone, PC, and Internet connection. Some indication of this pattern is already evident in the global spread of talk radio and independent music through streaming digital audio. Such technologies promise to bring an intuitiveness to the creative use of computers that will make the boundary between player/consumer and producer/artist more permeable.

But at the corporate portal, and within the intranets, a concurrent involution threatens this promise with cookie-cutter software that reduces creativity to one of several options and equates expression with consumption. While firewalls, filtration, transmission asymmetry, and security mechanisms encroach insidiously on the technical feasibility of new forms of reciprocal communication, nonetheless experimentation continues. Given the tools that are widely available, the impediments to creative research and playful production are not altogether firm. The Java language, originally conceived of as a control language for interactive television, has become a popular means to orchestrate so called "client-server" communications. Using it to implement a collaborative drawing system, I have become fascinated by the possibilities of networked interactivity. In spite of the limitations of the mouse, the programming language, and the browser context, I've become involved in a relationship with a public imagination that interests me. The things that have been made­­things that have happened­­in this unusual space keep me focused on overcoming the looming exclusion and insipidity that haunt tomorrow's Internet. Many times I have corresponded with people in strange and surprising non-verbal dialogues.

Because the saved drawings can be quickly "played back," in the same sequence of strokes and marks originally used, the products of this program resemble time-lapse studies. While I'm away from the drawing "place," some of the visiting artists leave their experimental animations, whereas others labor away on single images. Soon I plan to give people of an editorial disposition the means to script meta-narratives using the material that has already been made.

open studio

In building the program, I have been struck by the many ways that censorship intertwines itself with design. It may seem that censorship is too strong a term, but the result of a network of software constraints will be censorship. What is more, the limitations are not always apparent. Orwell writes in Homage to Catalonia,

there was a new rule that censored portions of a newspaper must not be left blank but filled up with other matter; as a result it was often impossible to tell when something had been cut out.5

And so it is with the new media: in the plethora of gizmos and widgets, few people will notice the parts that are cut out­­particularly as the censorship may affect procedures that are hidden from view as conditional structures of software. Programming and design decisions determine, broadly, what may be produced with software. Criteria for selection of "good" art can be coded. With this in mind the term "curatorial algorithm" comes to make sense. In practice I rarely delete or suppress what people have added to the archive of drawings. Certainly there are shout outs, graffiti, and attempts at self promotion; but I¹m not keen to tell people how to behave or what to write. Nevertheless, it is difficult to avoid playing a curatorial role in designing public interactive systems. Design decisions, which may stem from programming logistics or various biases, can amount to a form of feedback selection. The people who are put off by the design ultimately will not be represented in the collective art that is produced. It seems to follow that algorithms and interfaces interpellate (call forth) subjects.

The characteristics of this network drawing software have been years in the making. It began as a means to visualize the transformation of drawings in progress. Soon it became a means of distribution, too, and a vehicle for questioning the idea of progress itself. As one of many destinations on the Web, it resembles entertainment, but I conceive of it rather as a bridge between entertainment and a creative process that includes the visitor. Who knows but that this attempt at inclusion, and the relative openness of the system, will cause people to participate instead of passively observing the work? In imagining the next step, I recognize that my creative process now involves individuals of many demeanors. In some respects our interplay resembles theater, and so it is not unreasonable to relate what Andreas Huyssen writes concerning Heiner Müller's "learning plays":

Don¹t learn from them (object lesson, theater as finished product, uncritical acceptance of thesis), but, by actively reproducing them, learn through them (example lesson, re- production of a process, Bei-Spiel, critical trying out of behavior).6

Like theater, software has the potential to stage the contradictory processes of dramatic presentations, and to allow people to learn by the acting through of situations. But it also has the potential to forestall such imaginative investment, and interpretation, leading users through a succession of preselected hierarchical hoops. In order to effectively avoid the latter scenario, which is to say, the avoidance of the cool curation of Silicon Valley, Internet arts will need to be more ambitious about the development of alternatives. As individuals, small groups, or modestly funded museums, the pace of content production set by advertising sponsored, venture capital infused, Web sites appears unattainable. And probably pace is not a virtue anyway. Instead of emulating it, systems can be developed that encourage the creative participation of visitors. Such sites may be popular and at the same time conducive to the critical trying out of behavior. In this sense, it is possible to compete with the emerging Web entertainment industry; and, what is more, this competition need not follow in lock step with the genres and values of the Gee Whiz industry. Rather, the opportunity exists to close the gap between passive pass-times and nuanced artistic and intellectual engagement.






Back [1] [2] [3] [4] [5] [6]

1. Currently, Internet packets are routed without regard to what they're carrying, and "major" sites like have just one Internet address ( just as has one ( This situation is changing, however. The forthcoming IPV6 ("[Internet Protocol version 6] has a built in charge back mechanisms to bill for traffic", Network World, March 1, 1999), Virtual Private Networks, Service Level Agreements (ensuring guaranteed service levels), and MPLS (Multi-protocol Label Switching, which will enable pay-services [eg. movie delivery] to get higher priority in routing of data over the Internet), all combine to paint a very different looking future for what was potentially an unhierarchical network.
2. Enzensberger, Hans M. "Constituents of a Theory of the Media". Critical Essays. Continuum, 1982.
3. The Internet directory is an example of this distributed control.
4. MOO = MUD Object Oriented. MOOs employ a text interface using commands like "go North" and "say Hello". Traditionally, a common Internet tool named "telnet" has been used to connect to the MOO server software. Participants with access to the Internet connect to a computer that hosts the MOO server software and database. The interface allows one to read paragraphs of text, and to enter lines using commands like,

@who (gives listing of connected user/characters)
say something that other users may wish to read
emote some phrase (another way of speaking).

5. Orwell, George. Homage to Catalonia. Harcourt Brace, 1952.
6. Huyssen, Andreas. After the Great Divide. Indiana University Press, 1986. p. 83.



© Andy Deck 1998
Note: This is a revised version of an essay presented at the Museums and the Web conference, New Orleans, 1999. It also appears in a fall 1999 edition of Millennium Film Journal.