In the twenty or so minutes I have for this presentation, I'd like to give you some examples of what the current network infrastructure will support, what it won't support, and persuade you that it is important to have humanists involved in bridging that gap. With one exception, everything I show you will be from the World-Wide Web server at the Institute for Advanced Technology in the Humanities at the University of Virginia, a humanities computing research center set up three years ago with a major gift from IBM and with a major commitment of technical support, space, and operating expenses from the University.
In the examples I will consider, the particular needs of researchers in the humanities have produced some very interesting adaptations of the Web, and they point the way to more. It is my contention that we need to support humanities research in this medium for the very reason that the needs of the humanities scholars are, in certain definable respects, ahead of what scholars in other fields require, and therefore can drive some significant developments in tools and techniques--developments that other audiences, both scholarly and general, will have a use for, though they would not be the first to ask for them.
In a project entitled "Valley of the Shadow," Ed Ayers has, for the last two years, been compiling a detailed multimedia historical database on the towns of Staunton, VA and Chambersburg, PA. These two towns, located on opposite sides of the Mason-Dixon line and at either end of the Great Valley in which many of the pivotal campaigns of the Civil War were fought, are the basis of an in-depth investigation of the impact of the War on the lives of ordinary citizens.
The database consists of several basic types of primary resources, centrally accessible from the project's Archive page. From this page, one can search newspapers from the two towns, military records, census data, diaries, maps, and photographs.
In cooperation with the Institute for Advanced Technology in the Humanities and the Richmond Renaissance, Professor Ayers is currently planning to develop a museum installation based on this project, an installation which we expect to become the core of a national Civil War Archive, with geographically distributed but centrally maintained facilities for access, and remote stations for scanning in new materials provided by the public.
We are also negotiating with the American Art branch of the Smithsonian Museum for a online, curated exhibit of artworks from the period, which would be accessible from the Archive page. The Library of Congress's Matthew Brady photographs, already available on the Web, are another obvious point of connection. In fact, if we go back to the Institute's Archive and search the Military Rosters, selecting the appropriate rank and company, we will find Pvt. John J. Rhodes, Company K, 5th Virginia Regiment, C.S.A. (Stonewall Brigade), pictured in the Brady Collection. The entire content of the Roster entry for Pvt. Rhodes is here. If we search the Census, we can find records pertaining to people who might well have been related to Pvt. Rhodes, and if we search the WAIS index of the Civil War Archive for Rhodes, we will find that possible relatives appear not only in the Census, but also in the diary of William Heyser.
We see this project as a compelling prototype for networked humanities publishing, one which demonstrates the potential of interconnecting networked databases, as well as the appeal to a wider audience of humanities scholarship conducted in this medium.
Published three times a year since September of 1990, and co-published with Oxford University Press since 1992, Postmodern Culture is the oldest peer-reviewed electronic journal in the Humanities. Its World-Wide Web distribution site is the Institute for Advanced Technology in the Humanities, where it has been publishing multimedia essays since January of this year.
When PMC first began publishing, by Listserv, I predicted that electronic journals would pass directly from the state of being able to do less than print journals to the state of being able to do more than print journals, without ever passing through a phase of equivalence. Though our first WWW issues did not quite bear out this claim, having only still images as illustrations, our most recent issue has demonstrated the accuracy of that prediction. Robert Kolker's essay "The Moving Image Reclaimed", a comparative study of Hitchcock and Scorsese's camera techniques, is illustrated extensively with film clips from the works of these two directors. In one instance, Kolker has actually annotated one of the clips to indicate the actors' lines of sight. Elsewhere in this issue, Three Poems by Charles Bernstein can be read in textual form, to the accompaniment of recordings of the author's own reading.
I should note that, the further one moves in the direction of true multimedia publishing on the network, the more tenuous are the standards, and the more primitive the authoring tools. The MPEG standard does represent a step in the right direction, but it so far lacks sound, and therefore there are still good arguments for providing a Quicktime version of film clips, which we did. This was no extra work, as the author's clips began in the Windows .avi format, and we had to convert them to Quicktime on the Mac in order to make MPEGs out of them, given the software tools available to us. The annotated clip, in particular, with its black-and-white frames overlaid with colored lines of annotation, were particularly difficult to convert. The sound files were a little easier, but in this realm the problem is more with standards than with tools: there are the de facto standards of .au, .wav, and .aiff, and we were fortunate that a good utility (SoX) existed for converting from one format to another. Nonentheless, the Windows Mosaic client seems (still) unable to transfer the .wav clips, and though it is entirely possible that the problem is in the configuration of our client, I can testify that, in general, true networked multimedia is far from plug-and-play at this point.
In this project, an international group of artists, writers, and filmmakers has used an unconventional server based on a MOO (a real-time, text-based virtual reality program) to track state information about users throughout a WWW session. MOOs, originally developed at Xerox PARC, have been used for conferencing, teaching, and other academic pursuits, but their original heritage is in Multi-User Dungeons, an interactive networked version of Dungeons and Dragons. MOOs, like MUDs, provide for unique login ids and password verification, and players in MOOs belong to one or more player classes, with permissions (for everything from reading and writing to copying, moving around, executing verbs) derived from the classes to which they belong. What the Wax people did was to use the MOO itself to store the objects which would be read off as HTML pages, and they used the state-tracking capability of their MOO server to establish which browsers, or players, had permission to alter the text of a particular page, based on who owned the object that was that page. Using this, and fill-out forms, they developed a Web-based authoring tool for collaborative creative writing. And using that tool, they successfully constructed, or perhaps retrofitted, a full-length movie into networked hypermedia.
I've run several MOOs, including PMC-MOO--the second largest real-time virtual community on the internet--for the journal, Postmodern Culture. When I looked at what the Wax group had done, with the journal in mind, it was immediately obvious that there were applications for this tool in the context of peer review, where one might want work-groups assigned to review a particular article (but without permission to alter it or see one another's comments), authors capable of altering the item under review and seeing comments but not altering them, and editors functioning as superusers of these workgroups, permitted to see all the comments and to edit the essay. In its potential, this is an excellent example of generalizable benefits from humanities research on the network, and though it remains to be seen what will come of the collaboration, we have started to work together, and to seek outside support for the further articulation of special-purpose MOO-based WWW servers for a variety of applications where hierarchical read/write permissions are needed.
Another example of a generalizable tool that has begun to emerge directly from the humanities research taking place at the Institute is our image annotation tool. This piece of software, originally written in InterViews and now re-written, and about to be released in beta, as a Motif-based application, was designed in response to the desire expressed by a number of our fellows, but especially Jerome McGann, to be able to annotate images as one would annotate text. In the first release of this tool, the researcher can draw on an overlay and can attach those details to textual annotations. In addition, it is possible to resize the image independently of the frame, a feature which was developed specifically in response to Ed Ayers' need to work with full-size newspaper pages on a computer monitor. The image annotation tool is currently being used in the museum preview installation of the the Civil War archive, as the default image viewer, without any of its annotation features, simply for its convenient resizing capability.
The image annotation tool archives its overlays and the associated annotations in Hierarchical Data Format, and in future versions we hope to provide for direct image-to-image annotation (linking a study of a head, for instance, to the head in the finished painting), multiple overlays (different annotations of the same image, for different research or instructional contexts), and the on-demand conversion of image-plus-detail into a new graphic with an associated ismap instruction file. We hope these features, and the rewrite in Motif, will make this a widely used authoring tool for the Web, as well as a convenient private image viewer.
This painting is part of McGann's Rossetti Archive, and it depicts Dante on the first anniversary of the death of Beatrice. Supposedly, the three people standing behind Dante are visitors watching him at his work, but I can't help seeing another scene overlaid on the original, in which McGann sits as his Xterminal working on the Rossetti archive while three highly paid technical support staff stand at his elbow, scared to death of his next idea.
Technical support of an expert and fairly specialized nature is an absolute requirement for each of the projects that has come to the Institute, and I think the success of our efforts, such as it is, has depended largely on the collaboration between computer experts and humanities scholars. That collaboration has produced what may well be a unique hybrid atmosphere at the Institute, where the members of two disparate cultures learn to speak one another's language and even establish a common vocabulary. The exercise of developing technical solutions to humanistic research problems--the need to conceptualize the problems of one discipline in the terms of another--extends the territory of both.
My final example is an example of this, as realized in the Rossetti Archive's DTDs. A DTD is a Document Type Description, the set of rules that governs tagging in the Standard Generalized Markup Language, for a particular document or set of documents. HTML is a DTD of SGML--the most widely used DTD. But HTML is a general-purpose DTD, and a very simple one, and therefore it is not especially well suited to capturing highly specialized information. With the exception of the tags for citations and addresses, few elements of the HTML DTD have much to say about document content or structure--and yet SGML, as a grammar, is primarily intended to express information of that sort. For instance, here is information about the painting of Dante that could, in principle, be captured in appropriate fields of a DTD (and therefore could be searchable by field--show me all the water colors by Rossetti depicting Dante), but which can only be broken out as information on the page in HTML.
Before McGann was even aware of the existence of HTML, he and the technical staff and the Institute had begun to work out a DTD designed specifically to capture information about paintings, poems, and printed editions. The process of articulating the fields in this DTD became, for all involved, an intellectual excercise as well as a technical one, because it involved declaring the relation of elements within a work and from work to work. For example, mythological reference could occur in either a painting or a poem, and since Rossetti frequently created works in different media that dealt with the same subject matter, it became necessary to decide whether the mythological reference tag should occur in the section of the DTD that dealt with the literary work, which would be appropriate if one considered the literary work to be primary, or in the DTD for paintings, or both. Ultimately, and for reasons both practical and theoretical, this sort of discussion led to a rigorous conception of the "work" as something transcending any individual text of the work, and that conception was expressed in highly codified form in the Rossetti Archive's tripartite DTD.
My point is not that the Web is incapable of functioning as a browser for complex information: far from it. With the necessary filtering programs, information stored in the Rossetti Archive DTDs can be searched by field and the search results can be turned into browsable HTML on the fly. The point I want to make is rather that the shortcomings of HTML would not be evident, were it not for high-maintenance, highly specialized users. Pushing our infrastructure to meet the needs of those users, even though they do not represent the mainstream, will produce tools and techniques that will benefit mainstream users. Whether it is the English scholar who wants to construct complex, layered electronic editions, or the Religion scholar who wants to present texts in English, Hebrew, Greek, and Arabic on the same page, or the Archaeologist who wants a virtual walk-through of a reconstructed ruin, we would do well to see that their needs are met. There's an old saying: one fool can ask more questions than a thousand wise men can answer--to which I should add, the humanist on the internet is not your average fool.