Because the print industry remains a powerful, institutional force, existing academic archives like MUSE and CIAO will continue to emerge as more publishers of esoteric scholarly publications realize the benefits of the online subscription model for multiple journals. Rather than witnessing a continued decline in individual journal subscriptions by academic libraries, publishing houses should expect to regain subscriptions to digital archives. As more journals go online and researchers come to expect online access to these resources from home, libraries will begin to make decisions as to which journal packages they will pay for on a subscription basis.

In the long run, such expectations will likely lead to new classes of intellectual property managers who will serve as third party distributors, organizers, and coordinators of new disciplinary knowledge. Malcolm Getz, Associate Professor of Economics, suggests that out-sourcing information management to third party "electronic agents" is not all that different from current models for typesetting and printing. The model for such electronic agents exists already in forms like Silver Platter's Psychology, ERIC, Dissertation abstracts and in Online Clinical Trials, which "offers publishers the opportunity to sell electronic access to [more than 250] journals by both subscription and pay-per-look." (We discuss these agents at length in the fourth model that we describe.) But before the market is developed by electronic agents, the publishing industry itself will provide an important source of the value added filtering, linking, and, integrating (disciplinary knowledge) that researchers will increasingly expect.

Rather than selling rights for storage and management of published texts to third parties, publishers are likely to continue holding on to copyrights for books or journals, materials which already exist in printed form. Ownership of intellectual property is so fundamental to the current industry that large-scale changes are unlikely to happen quickly or uniformly. More likely, multiple new models will emerge simultaneously and compete for market share. Chief among them will be different versions of MUSE, CIAO, and even JSTOR archives that keep long-standing institutional structures intact.

Hence, in the short run, publishers of academic journals are more likely to build and distribute their own website archives to paying customers and manage them as new materials are added and new technologies overlayed. These archives will be delineated, "demarcated" as professional working spaces, and serve specific functions in relation to the publishing houses they reflect. They may also align themselves with professional organizations like MLA, APA, and others or create their own affiliation of linked websites as they attempt to turn their archives into "living spaces" for research and intellectual exchange. One tremendous advantage of such a system is a generous fair use policy; subscribers and their constituents can download and copy journal articles for archiving and educational (teaching) purposes.

In order to remain competitive and warrant high development and management costs, these webbed archives will need to provide greater services than their print counterparts could offer. Publishers might hire web-editors to identify pertinent conference proceedings, working papers, annotations, and evaluations of texts. For instance, these web-editors might be hired to annotate texts in the archives by virtue of their usefulness to various scholars, researchers, and teachers; they could also create hypertextual links throughout the archive. Some texts will be more useful to generalists, some to specialists, and others to scholars outside the discipline (Varian). Such a filtering function could validate important contributions to professional communities, thereby providing documentation for promotion and tenure committees.

Web-editors, according to Hal Varian in "The Future of Electronic Journals," could also substantially reduce the labor costs and turn-around time necessary for high-end print production. If web-editors reviewed texts, copy-edited them, and ranked them, then production costs could be lower. Furthermore, the cost and time involved in print communications (sending manuscripts through the mail, waiting for reviews) limit the number of projects a publisher can undertake. And too often, years transpire before desperately needed information is actually distributed. If more electronic journals were included in archives like CAIO, then valuable information would be accessible more widely and more immediately.

With all of the advantages of integrated bodies of knowledge streaming from individual publishers, archives are still bound by disciplinary and institutional constraints. The cultural capital of such islands of webbed knowledge can be greatly enhanced when they become interlinked to a wider net than even the JSTOR model (the library initiative of an interdisciplinary set of journals). If archival spaces always "presuppose a system of opening and closing that both isolates them and makes them penetrable," then why not develop searching mechanisms among larger numbers of related (or potentially related) archives? Such meta-hypertexts are possible even with current technology.