Reading MediaBASE: Design as Debate in the Digital Humanities

The web presentation of the MediaBASE software concludes with these words:

‘The concept palette, intended to provide a conceptual framework for a given user group, is input ahead of usage time, and updated by an administrator. Thus, while each object is tagged with “objective” metadata during the importation process, it also accretes cumulative “subjective” metadata during usage by virtue of its associations with concepts, discursive text and other objects in compositions. As older compositions are sampled and repurposed for new ones, the media objects contained within them develop deep histories especially meaningful to the user community. Over a period of time, MediaBASE could provide a valuable ethnographic tool for exposing a cross-sectional view of media artifacts and the social structures and networks that surround them.’

It’s apparent here that the MediaBASE design project attempts to provide a basis for forging user communities around media composition activities, providing concept-based analysis of digital texts, with the idea that these concepts, and descriptive data associated with them, can then be used as meta-data tags, or used to determine additional species of tags, that would allow the recording of how digital authors put together, use, exchange, and re-use multimedia compositions.

However, most of the demonstration prototype is dedicated towards presenting relatively generic composition and search functions (browsing, loading, composing, and presentation of multimedia content), without regard to the larger questions of connectivity, context, community, and so forth. In this design prototype, such high-level notions as “network”, “user,” “community,” “concept,” “history” (whether of user actions, or of a composition, or of concepts or meta-data tags) should have some representation somewhere, but they are not presented specifically in the software (even in a non-functional way, for after all, MediaBASE is presented as a conceptual prototype). So a major problem for me is that I can’t see how these high-level constructs are represented (or even mapped, or concretely proposed) by the software designers. Generally, then, while the software description says that MediaBASE allows conceptual searching (that is, a semantics) of media content, use, and pattern of use, I found the treatment of what constitutes a concept (and therefore, a meta-data tag) largely not solved, or even clearly posed as a problem; and this observation applies to the key notion of “concept” but also to the other high-level constructs which the project implies as part of the rationale for its development (community, network, etc.)

This ambiguity in MediaBASE (its apparent lack of solutions in representation of high-level constructs) leads me to conclude that the prototype doesn’t really speak to the stated goals of the project in even a non-functional way. The project does claim to provide sophisticated tagging of media content in what it calls “objective” and “subjective” modes (this opposition seems to be the only definition we get of a “user” and it is an implicit definition: “subjectivity” ...?). This opposition between “subjective” and “objective” organizing of data tags presumes that the way media objects are composed (if this opposition between objective and subjective tags could be upheld) would enable the meaning of their composition to be seen in relation to the history of their use. Presumably, such a conceptual mapping would allow a gauge of the “meaning” of media compositions that would be impossible with only an automated analysis (whether textual, visual, sonic, or otherwise). In this way, then, MediaBASE aims at the classic advantages of “mark-up languages” over computational description/prediction processes that aim to produce “machine knowledges” somehow equivalent to “human knowledges” (in a word, AI; in a different mode and with a different object, A-Life.).

That’s to say, that while MediaBASE aims to skirt the historical failures of AI or other automated analysis/production of media content by deploying meta-data tagging mechanisms to support semantic analysis (and thereby enable advanced browsing, composition, and other user actions), it seems to do this by setting up a simple opposition between “objective” and “subjective” tagging. While it would make sense that “data” as such might be seen as “objective,” (file size, date of use, pattern or incidence of use, attribution of user ID, cost, type, protocol, etc.) it appears that MediaBASE means “objective” in a more extensive way (although, again, we cannot really tell). And of concern here is the fact that the meta-data tags projected to be determined in “administrative” functions seems to obscure the question of how tags are generated, what concepts are, and how these relate to the processes of authoring generally. Because the presentation of the software promises that the tagging of media material matters ultimately in relation to the social networks in which these are used, we can only say that this “administrative” function in relation to meta-data is vague at best.

The larger problem, it seems to me, is that with this sorting out of administered and user-generated tags (as objective and subjective modes), the larger work of determining what the specification of higher-level constructs would be has gone by the wayside (community, connectivity, composition, meaning of use or exchange, etc.; apparently, these would eventually become features of the MediaBASE software, on the basis of the composition, use, and exchange of digital texts).

In attempting to use meta-data tags to generate a description of compositions and their histories, MediaBASE sees the value of gathering data about use from the user’s actions (like social networking sites, which generate “communities” on the basis of those actions). But because what a “concept” or “composition” or “community” or “network” remains unrepresented in meaningful ways, it is hard to see how the data gleaned from the tags and their traces, or the traces of user actions, would actually improve the process of composing multimedia objects in a networked context. (Of particular interest to me was the concept tag “propaganda.” What basis as a meta-data tag would the determination of “propaganda” have, in as much as “propaganda” indicates a semantically sophisticated content analysis based on the determination of subjectivity, instrumentality, and media power? Here, the MediaBASE software might appear to aim at automating a classical theory of 20th century communications theory, itself based on and enabled by the administrative organizational and analytical frameworks of, amongst others, the Rockefeller Foundation in collaboration with Princeton University, CBS, and the U.S. government in response to the growing geopolitical reach and cultural importance of “new media” like radio in the 1930s; see for example, the various emanations of those attempts, whether in Lazarsfeld, Adorno, or others.) Broadly, that which has tagging value in MediaBASE, but which cannot be attributed to specific user actions or intents, will be determined by “an administrator”—but what does that mean? What does the administrator’s interface look like? Is this administrator a person, or a program, or a process? Can this work be outsourced, and would English language facility be necessary to handle it? Etc.

Are the meta-data tags being provided or generated here by an administrator relevant to the social networks producing media objects, and why would they be? And how exactly does this database or these concept groups relate to the internet/web? Can MediaBASE represent web sociality as such, in some meaningful form?

These are important concerns, because they speak to the usability of this kind of application if we are assuming that it might intersect with the cyber-social spaces administered by sites such as MySpace.com. So, ultimately, what is MediaBASE for? Is it for social networking, facilitating the description of social networking as the analysis of media artifacts, as the presentation claims? Does it then assume that media composition as such determines social networking? Or that, in effect, the “social network” is not different from a multimedia composition? That social networks, too, are media compositions? If not, why not?

I have to admit I’ve been impressed with the capacity of sites like MySpace.com to catalyze what have long been deeply problematizing aspects of specifically administrative intersections of technical, social, and media networks, especially in the way that affective articulations of connectivity on the level of one network might map, or not map, to other forms of filiation in other levels. A close friend of mine (in our shared Los Angeles space) exists in a MySpace network as a friend of Hollywood star Vin Diesel. But that doesn’t get me onto the red carpet, or into the pages of the weeklies. But I’m not complaining (and neither is my friend). When it comes to the projection of personal identity as spectacle, we can very easily get into the pages of MySpace.

Some larger questions, then, would have to be answered in order to compare MediaBASE with social networking sites like MySpace.com, which make much fewer claims about “organizing” data semantically, but do successfully emphasize descriptions of social networks in highly specific (though very limited) ways (“my friends,” “my favorite friends,” “my music,” “my new messages,” etc.). These extremely limited but undeniably powerful descriptions have made social networking sites incredibly productive of networked sociality evidenced precisely as the composition of digital texts (here, it would be useful to summarize the kinds of descriptions of social activities that MySpace provides its users, and the kinds of variegations these descriptions allow in the projection of “myself”).

Briefly considering these concerns lead me to conclude that MediaBASE has not targeted the kinds of problems it would need to solve in order to meet its development goals (media composition; ethnographic analysis of media communities). What would an “ethnographic tool” of the sort imagined by the designers mean, anyway? Do web users compose multimedia, only for these objects to be given over to ethnographic analysis?

Here, the work of history is being reversed: MediaBASE assumes that we make things for others not involved in the process of exchange to analyze them, rather than for the object itself to be used, re-used, exchanged, etc. This reversal suggests the larger problem here: MediaBASE is functioning both as a theory of multimedia, and as a design prototype, without any differentiation of “multimedia composition” as a historical category, as a social formation, and as “software application.” The result is that, methodologically, history collapses into software, which can then appear as (hypothetically) the accumulated traces of what the software produces.

I’m relatively sure that enterprises like MySpace.com don’t claim to be analyzing the social networks emerging around their site with the same tools users compose their pages with. In fact, their very status as enterprises as such has to indicate that they must have a much more powerful set of analytic tools to do that work; it would be surprising indeed if business plan, marketing strategies, concept and design development, application implementation, and site maintenance, development, and revision were all conceived within the terms, options, meanings, and devices provided to users, though the claim of precisely that concept of a “flat” driving of technological and social change as radically user-centric may well be part of the appeal of the site. Largely, though, such a “flat” projection of socio-technical organization finds a more fully fleshed out mode of operation in the confluence of open software development, peer networking, and data sharing, and almost always tends to exist as unevenly developing “alternatives.” That alternatives exist does not mean that such alternative systems, development projects, orientations, or “communities” do not provide all manner of commercial opportunities, options, resources, or other material, social, or technical benefits to commercial enterprises; open software development or peer sharing communities and their varying practices have been, over the last 15 years, important parts of innovation and commercialization strategies for industrial development projects, from the hardware and software industries to the various content-oriented industries, whether Hollywood cinema, video game, music, publishing, and so forth; and at the same time works within the horizons of technological goals partly established by administerial mandates, regardless of the particular legality or illegality of web-based file sharing, for instance, or of non-normative or of spectacular forms of political speech. If Castells’ comments about the social character of the internet are considered (that the internet is essentially a product of four interacting social groups: techno-meritocratic research elites, hackers, social communitarians, and entrepreneurs), we would wonder about the ways web-centric internet development generally might reflect or diverge from the broader historical forces shaping the internet (whether or not we are satisfied with Castells’ profile of the social character of the internet, and there are numerous questions I would want to raise in that regard, too).

In any case, because it concentrates on user identities projected largely as a personal composition practice of administrated digital texts, and without consideration of the other concerns I’ve touched on here, MediaBASE is making a very large leap in suggesting that the tool of composition can also function as the tool of analysis of and for communities emerging around the artifacts composed. This is a very significant claim, precisely because it is based on the composition of digital texts, and so, reflects a rather different set of different models (I may be granting the design more coherence than it actually reflects): the claim here centers around a form of literacy, a literacy based on practices of composition enabled by computational assistants (or, as I indicated above, some undefined “administrative” processes) wherein analytical concepts such as “propaganda” or “nation” become part and parcel of the historical process of authoring, whether that authoring takes place within networks of users, or as networks of texts.

But I don’t see this claim of “literacy in and as the digital text” supported, nor am I sure that aiming at this goal, even implicitly, is a reasonable idea. Arguably, on the one hand, it’s the ideal of literacy as software: readers become authors, and the concepts relating reading to authoring are defined as computational functions and as social functions.

But on the other hand, in the way the compositional and administrative designs are presented in MediaBASE, it’s also the idea of “total software.” What is embedded in MediaBASE, actually, is not so much a complex negotiation between, say, the, technological, juridical, economic, social, institutional, or aesthetic projects currently marking out the horizons of digital speech (whether as digital text or of digital readers and authors), but rather, a debate between critical theory and cultural studies, though this debate is one that is, quite interestingly, articulated as a design approach: can the theorization of the “techno-social apparatus” also serve as the description of the effects of that apparatus? Can history be theorized as the output of a social “instrument”? Or, again, shouldn’t any productivist framework be tempered with an account of the unpredictabilities of reception?

(And in this regard, I think that as a design approach exemplifying a critical perspective on negotiating the problematics of semantics, form, structure, articulation, and technique in and for contemporary digital speech, MySpace succeeds because it doesn’t attempt to make those two frameworks (production; reception) adequate to one another; it merely assumes that if you can match the articulation of social networks (without thinking too much about either the syntax or semantics thereof) to network contexts and technologies, new usage will proliferate, according to the developmental tendencies of networks themselves.)

The importance, though, of MediaBASE’s theoretical presumptions, is that I (but who am I? Again, a user, or an ethnologizing “administrator”?) would be able to move seamlessly from the “network of multimedia compositions” back to the “social networks” wherein the compositions were made: giving an ethnography of the history of media composition, produced as MediaBASE artifacts. Such a possibility would mean that the entire work of social interaction would somehow be captured in the software (and not only the multimedia object itself).

MySpace is successful ultimately because it does not attempt to “move back” to an “original” social network generically. (Any “ethnography” — or any other “administrative” function—would be a separate activity requiring a different “interface”.) It assumes that it will encourage the construction of certain kinds of social networks, and limits the kinds of activities done within social networks, in order to maximize the expansion of the social networks themselves. Media composition is clearly less important here. And if MediaBASE is conceived as potentially extending what network sites like MySpace are doing, it will have to bring to bear on the process of media composition some bigger problems of representation than simply “meta-data” for digital texts and their potential ethnographic consumption: concept, connectivity, affect, use, articulation, history, technology, person, project, life.