Thinking Strategically about Electronic Publishing

James J. O'Donnell

for a conference on scholarly communication for the next century, held in Vancouver, 5-8 March 1997, under the sponsorship of Simon Fraser University

I have today essentially one point to make: that we know far less about where we are going with electronic scholarly communications than we would like, and that we must match our ambitions and our plans to our ignorance as well as to our optimism.

My contribution will arise from two distinct lines of my experience. On the one hand, I have since 1990 been electronic editor of Bryn Mawr Classical Review (since 1993: also Bryn Mawr Medieval Review). As such, I realize with some embarrassment that I am both a pioneer and a grizzled veteran of the community that seeks to create rigorous new bodies of electronic scholarly publication by outright imitation of existing models. BMCR/BMMR are new titles and though one has a paper avatar, they exist largely as e-creations; but they are distinctly pat of the community that also includes now many hundreds of direct transfers: the same journal appearing in paper and now in electronic form.

But on the other hand, I have also earned a place among the early adopters of the WWW as a place for less formal publication related to research and teaching. Like many faculty around the world, I have found the WWW a place to go to create and shape resources for my students and colleagues on a growing network of my own web pages. Preprints, postprints, primary texts, electronic versions of old classroom "handouts", always linked to similar resources elsewhere: people like me have almost literally been falling over ourselves the last three years to "publish" in this way.

To look to the next millennium, I argue, requires us to make sense of a world in which both of the impulses I represent will be strong forces. It might seem we would be better off choosing, and choosing the former of my personae.

For we live inside a system of scholarly communication that we have built up carefully over decades, even centuries. Scholars and librarians alike have built up skills as hunters and gatherers in the forest of learning that depend heavily on our knowledge of the complex and ramified system of publication. We know our way around our libraries the way Mark Twain and other pilots learned their way up and down the Mississippi, even to the point of having the same ability to track a thousand vital small changes in what looks to the unpracticed eye to be a stable and unchanging environment.

And so people like me know what we want out of a new electronic environment: the same thing we've always had, but we'd like some more, please, and we'd like it to work faster, be searchable, and, oh yes, could it please have more pictures? And if those requests were met with a set of resources that were recognizable on-line versions of existing journals, or on-line journals that faithfully followed the model of print journals as far as possible, we might be very happy indeed. If we could expect that modest tranformation, we could move rapidly to assessing the specific costs and benefits of each transformation, perhaps culling some titles in the process. And then we could settle back, as quickly as possible, to our old ways of doing things.

I think you can already detect from my tone that this is a future I don't believe in, and perhaps even surmise my reasons. But I linger on this option because there is an important caution to offer. Many of the specific projects that we now see a-borning to carry this kind of conservative, cost-benefit-analyzed option forwards seem to me to err by assuming implicitly that it is possible to move straightforwardly from one medium to another: that it is possible to assume, or even impose, a familiar and defined system on the new media of communication within which to content yourself.

The deficiency of this approach is that such innovators, often quite innocently, are creating sub-systems which, to work properly, would require the whole rest of the world to use their system for optimum efficiency. What in fact happens in that situation is that the end-user finds a new set of commands to learn and needs to remember, when searching for the Journal of Medieval Palaeobotany, who the publisher is, what system of coding they have used, and how to navigate that site. These idiosyncrasies, which are imposed to add functionality to the individual resource, have the effect of making the totality of such resources less interoperable than our existing print literature is. Unless we imagine some impending war of the publishers to establish the hegemony of one scheme over all others, this kind of multiple system-generation will lead us to a bright future through a dark and noisy intermediate stage of indefinite duration. The theoretical error lies in assuming that the totality of a system exists, and that individual choices may be made in that context.

By contrast, the world of the WWW is very different. Amateurs, underfunded, have no comparable powers of specialization. We use HTML, we type our own angle-brackets or get very clunky HTML-assistant programs to do them for us. Our products are often a little gauche and naive, but they are strikingly interoperable. The battle of the browsers has not yet created crippling differentiation, and so we can reasonably assume a certain configuration of hardware and software that virtually all our intended readers will have, and we can cater to them there.

What emerges looks in one way like blooming buzzing confusion. There is no peer review, no quality control, and managing the exploding universe of such information is a continuing struggle. The last year has seen Yahoo and AltaVista become standards. In their present form, they can probably last another year or so before the body of information they organize grows too large to manage. It is all enough to make you want some benevolent agency to fund a project to create the digital library to end all digital libraries.

But even if a Croesus rich enough to fund such a venture could be found -- and I think it is an important fact about the universe that such a patron will not be found -- such an attempt to enclose the world in a single vision of structures is fundamentally flawed. It calls to mind the mad struggle of Ivar Krueger, the "match king", to monopolize his industry world-wide in the days after World War I. I wish to argue that it is the uncontrolled experiments, the ones that do not know what system they are building, that contain the future. They begin not with an idea of an overarching system and without any reassuring idea of a predictable context, and they seek to build useful things. The usefulness of these things may not be long-lived, but the environment is one in which competition brings refreshment. Survival of the fittest is a disconcerting management tool for scholarly communication, and we as scholars will resist it. But my point is that in a time of real change and innovation, we dare not resist it too strongly.

Let me give a case in point. For many years, serious people have been working on the standards of the Text-Encoding Initiative, based on SGML (Standard Generalized Markup Language), and promoting their use for serious publication. Many of the special projects I spoke of above use some form of SGML for their sites. But oddly enough it is precisely in its desire to impose a standard that TEI/SGML fails to achieve a standard and instead becomes yet another closed system in a world of competing closed systems. If I must first choose your standard to benefit from it (and go out and get the right software and learn how to use it), then it cannot yet become a standard. HTML is by comparison to SGML a paltry thing, but it has become a standard by the enthusiastic adoption of a large population.

Wisdom, then, consists in not rejecting a true standard for a more popular, but false one, but in managing to live within the true standard we have and working to push its capacities towards greater functionality. This means in the short run learning to forego some of the bells and whistles that we all want to have when we create a new resource of our own, but in so doing we do a more responsible job of building a common space of discourse and scholarly communication. The challenge is to think strategically about the future without mistaking strategic thinking for a veiled form of imposing our own closed system, our own will, on a larger universe. Thinking in open systems terms is difficult and not instinctive for academics, but at the same time it offers our best hope for building together a system of discourse that is robust, universal, and effective.