1995 Annual Report
William G. Bowen
Kevin M. Guthrie
Executive Director, JSTOR
Of the many initiatives sponsored by the Mellon Foundation in recent years, the JSTOR ("journal storage") project has proved to be particularly intriguing--and, we believe, particularly promising. By reconciling the sometimes competing interests of scholarly associations, other publishers of scholarly journals, libraries, and individual users of journal literature, JSTOR offers the exciting prospect of dramatically improved access to scholarly materials for faculty and students, reductions in capital and operating costs for libraries, and greater long-term financial stability for publishers.
Claims for the "revolutionary" promise of new technologies are often wildly exaggerated, but this seems to be one instance in which the well-orchestrated application of technologies that are broadly available can make an enormous difference. In this "update," we hope to provide a context for understanding the current status of the project, a summary of new directions, and likely next steps--including the possibility that a prototype database will be available to charter subscribers by the end of 1996. (Note 1)
Background and current status of the project
Initiated as a demonstration project in the spring of 1994, JSTOR began as an effort to ease the increasing problems faced by libraries seeking to find appropriate stack space for the long runs of back issues of scholarly journals. The basic idea was to convert the back issues of the paper journals into electronic formats that would allow savings in capital costs while simultaneously improving access to the contents of the journals and addressing preservation problems. In August 1994, the Foundation appropriated $700,000 to the University of Michigan to develop software and purchase computer hardware that would allow bitmapped images of journal literature to be accessed over computer networks. In December 1994, the Foundation appropriated an additional $1,500,000 to Michigan to support scanning of pre-1990 issues of ten core journals in history and economics (approximately 750,000 pages in all).
This project has been far more complicated to implement than expected; indeed, as one of us has said on other occasions, if we had known at the outset all that we know now, we might not have marshaled the courage to begin! One experienced observer of both the academic and business worlds said, following his participation in a recent meeting of the 12-person team at Michigan working on the JSTOR "production" cycle, that he had never seen a simple concept that was so complex in its implementation.
To give just one example, we were surprised to learn that a number of long-lived journals lack accurate records of their own runs. Over the years, supplementary issues have been published from time to time, issues have sometimes been skipped, and there have even been errors in numbering (for example, two issues numbered "four" can appear in the same year). Thus, it was difficult to know when we had in fact succeeded in bitmapping the full run of a journal, since there was no known definition of the full run at the outset. An unanticipated contribution of JSTOR has been to provide, for the first time in some instances, a complete publication record for particular journals, with an accurate index of all articles, reviews, and other materials they contain. This laborious process of assembling and checking the "raw material," as well as the bitmapped images, has necessitated a heavy investment in quality control. More generally, the flow chart now used to monitor progress in converting paper issues of back files into the final electronic database involves more than 30 distinct cells, some with their own subsets of instructions!
Complications notwithstanding, we have now accomplished our original objective: electronic replications of all of the pre-1990 issues of the ten economics and history journals chosen to begin the project are now incorporated into the JSTOR database. Bitmapped images of every published page are linked to a text file generated with optical character recognition (OCR) software which, along with newly constructed indexes, allow for complete search and retrieval of the published material. The database is being used at the first five college test sites (Bryn Mawr-Haverford-Swarthmore Colleges, Denison University, and Williams College), as well as at the University of Michigan, Harvard University, and the Center for Advanced Study in the Behavioral Sciences.
In all essential respects, JSTOR works--and, indeed, works very well. In particular, the decision to scan the pages at high resolution (600 dots per inch) has proven to be sound; the images are of truly archival quality, and even the most complicated figures and equations can be printed beautifully (provided that the right printer applications are used in conjunction with the specified printers). Users can perform full-text searches on the database as well as searches by abstract, author citation, and article title. Also, they have the convenience of being able to access the journals using standard PC equipment at any time (this "library" never closes) and from any location that can access a campus network. Issues of journals are never "out," and are always available in pristine condition. In sum, the addition of these powerful search and printing capabilities makes the JSTOR system much more than just a way for libraries to save capital costs; it has become a scholarly tool of enormous value. (Note 2)
These accomplishments duly noted, those of us who have been closest to JSTOR are more aware than anyone else of the many respects in which it can be improved. Work is ongoing to replicate the database at another site, to test the feasibility of a score of technical enhancements, and to experiment with adding Standard Generalized Markup Language (SGML) tags to at least some back issues (to permit more structured searches and the use of hypertext links). JSTOR is committed to remain up-to-date, to continue to take advantage of advances in electronic technologies, and not to allow itself to "ossify" in a highly technical field that continues to change at a dizzying pace.
The most important open questions are by no means solely technical. Studies are being made of the ways in which JSTOR affects the way readers use journals and, by extension, the very process of scholarly communication. (For example, given access to the highly convenient search capabilities of JSTOR, will students, as well as faculty members, make fuller use of older journal literature?) Work is also underway on the development of viable economic models which will assure cost-effective access for users while protecting the revenue streams of publishers and allowing JSTOR to recover its costs.
New directions for JSTOR in 1995
Three substantive decisions concerning future directions for JSTOR were made during 1995: (1) to include more fields and more journals; (2) to "roll out" the database to an unlimited number of sites; and (3) to explore linking current issues to backfiles. Taken together, these judgments drove another decision, already noted, that was organizational in character: to preserve JSTOR's nonprofit status, but to separate it from the Mellon Foundation.
More fields and more journals. There was never any reason, and never any intention, to include only the fields of economics and history in JSTOR. These fields were useful starting points, but it was evident from the beginning that one of the major advantages of the JSTOR concept is that it allows scholars, and particularly those interested in topics that span disciplines, to search for materials within a number of related fields simultaneously. For example, two topics now being investigated by staff members at the Foundation--the nature and consequences of affirmative action policies in college admissions and the changing role of intercollegiate athletics in American higher education--require searches of journal literature in a wide variety of fields, including economics and history, but by no means limited to these fields alone. (Note 3)
One of the nicest compliments paid to JSTOR is that, as the great potential of the project became ever more evident, scholars from many disciplines urged us to add journals--almost always in "their" fields! Some urged us to add leading journals in fields adjacent to history and economics, such as political science, sociology, finance, and area studies. And we agree that there is much to be said for assembling "clusters" of related fields. Others argued out that we should include fields, such as mathematics, which are important in and of themselves and are also relevant to scholars in other disciplines. (Note 4) Finally, colleagues at the Foundation made the obvious point that it was odd, and even awkward, for the Foundation to appear to ignore journal literature in fields in which it continues to be an active grantmaker. Ecology, population studies, various fields within the humanities (art history, philosophy, literary criticism, and classics, among others), and higher education all come to mind immediately.
Once we had convinced ourselves, and others, that JSTOR worked, expanding the database along the general lines just suggested made eminently good sense. In considering other journals, we have been avowedly opportunistic, since we did not believe that there was any single "right" way to proceed. Our experience has taught us that there is much to be said for assigning highest priority to fields, and to publishers, that evince a strong belief in the JSTOR concept and genuinely want to work with us in a partnership mode.
Conversations are now underway with publishers in a number of fields. We expect that, as time and circumstances allow, we will add fields, and journals within existing fields, to the initial JSTOR database. To cite just two examples, agreements have already been signed to include the Annals of Mathematics and the three journals published by the Ecological Society of America. Also, we would like to include leading journals published outside the United States.
"Rolling out" JSTOR to an unlimited number of sites. The responses to date of JSTOR users (including those who have attended demos as well as those at the test sites) have been extremely positive. Representatives of the library community have also been enthusiastic advocates of JSTOR. These expressions of support have led us to conclude that we should begin now to plan for general access by authorized users to at least the backfiles of an initial set of journals. In short, we are committed to making the transition from an initial test phase of the project to an ongoing operation. It is encouraging to report that we have already received commitments from the American Economic Association, the Econometric Society, the American Historical Association, the Ecological Society of America, and the publisher of the Annals of Mathematics to make the backfiles of their journals available through JSTOR at an unlimited number of sites.
Presuming that agreement as to how to proceed can be reached with the rest of our initial publishing partners, and with other publishers, we hope that it will be possible to have some number of core journals ready to "roll out" to libraries and other subscribers by the fall of 1996. But we also recognize that much work remains to be done, and that this schedule could prove to be overly ambitious. In addition to working with publishers and continuing to make technical improvements to the database, it will be necessary to define an initial set of relationships with subscribers, which will have to include agreements on subscription prices (since JSTOR must recover its costs) and proper means of authorizing and restricting access to relevant parts of the database. (Note 5)
Linking current issues to the backfiles. A third major decision made during 1995 was to extend the initial conception of JSTOR in one major respect. As we came to appreciate its capabilities more fully, it became obvious that there was no fundamental reason to focus solely on backfiles (if one is prepared to address the complicated questions concerning pricing and revenue streams, which the inclusion of current issues presents). There are evident benefits to all parties in linking electronic archives of backfiles to electronic versions of current issues of the same journals. Such an arrangement would allow users to search the entire run of a journal, from the most recent issues back to the earliest ones, using common software. For both users and libraries, this approach would also solve the problem of how to update the historical archive, since updating would occur annually and automatically if linking were achieved. Moreover, libraries would have the option of avoiding the costly and sometimes arduous process of binding current issues and then finding space for them in the stacks.
Publishers can also benefit in significant ways, as they seek to make a smooth, affordable, and timely transition to the fast-moving world of electronic publication. For those scholarly associations and other publishers that have not yet invested in developing the technology and infrastructure to build and maintain electronic versions of their journals, JSTOR offers the opportunity to take full advantage of years of development by a nationally recognized team at the University of Michigan. In addition, through participation in JSTOR, publishers can gain assistance in distributing their journals in electronic form to authorized subscribers (libraries and individuals) and learn from others which pricing models and restrictions on access work best in protecting revenue streams. Over time, we expect participating publishers to be able to reduce distribution and inventory costs.
Moreover, we are optimistic that JSTOR will increase the attractiveness of the participating journals to libraries and scholars worldwide, both those that currently subscribe to the paper version and those that do not. For the current subscribers, there are the many benefits of the JSTOR database previously discussed (including cost savings for libraries and the powerful search engine). In addition, many of the members of the second group--those that are not current subscribers--may lack the space and infrastructure to process and store paper copies of journals, but would benefit and might choose to subscribe if offered convenient and affordable electronic access.
The Ecological Society of America (ESA) was a major partner in the decision to develop mechanisms for linking current issues to backfiles. The leadership of the Society expressed strong interest in entering into such an arrangement with JSTOR, and agreement has now been reached on an experimental plan to link electronic versions of current issues of the Society's three journals (including its main journal, Ecology) to the electronic backfiles. Other publishers have expressed strong interest in similar arrangements, and discussions are ongoing. To be sure, there are complex technical, economic, and administrative problems to be addressed, and it almost certainly will take longer to work out these problems to everyone's satisfaction than it will to move ahead with the backfiles alone. But we are persuaded that linkage is the right approach, and that it is, in fact, likely to prove to be an inevitable consequence of the technical possibilities associated with digitization.
Creation of a new not-for-profit entity
The major organizational decision of the year--to locate JSTOR within a separate, self-contained not-for-profit entity--was based on our recognition of the rapidly increasing complexity of the project, its great potential, and the need to demonstrate sooner, rather than later, that JSTOR can be self-sustaining. The interrelated decisions to add fields, to make JSTOR widely available without undue delay, and to explore linking current issues to the backfiles, combined to give urgency to this change. On July 31, 1995, the new JSTOR entity was incorporated, and in January 1996 it received a tax ruling classifying it as a 501(c)(3) Public Charity. At this writing, JSTOR is still being "incubated" at the Foundation, but it is expected to move into its own space at the New York Public Library's new Science, Industry, and Business Library later this spring.
As a not-for-profit entity, JSTOR is prohibited by law from accumulating funds beyond its legitimate needs, and, in any case, it would have no interest in doing so. Its sole purpose is to serve the scholarly community by increasing the availability of scholarly journals and enhancing their usefulness, while concurrently reducing library costs. In these respects, the posture of JSTOR differs fundamentally from that of commercial vendors. Unlike commercial entities, the test of success for JSTOR is not merely the proverbial "bottom line," but how well it facilitates teaching and scholarship by improving the mechanisms of scholarly communication. JSTOR works in concert with publishers as well as with libraries and the scholarly community at large, as it seeks to advance the common interests of all parties. It operates on the premise that it is only by adopting a "system-wide" perspective, which recognizes the legitimate needs of both the providers of scholarly materials and their users, that socially optimal arrangements can be put in place. (Note 6)
At the same time that it is appropriate to emphasize the broad social and even philanthropic objectives of JSTOR, it is no less important to recognize that fiscal discipline is essential. The Trustees of the Foundation have always believed that JSTOR would have to be self-sustaining eventually. Perpetual subsidy would be both unrealistic and unwise: projects of this kind must make economic sense once they are up and running. If users and beneficiaries, broadly defined, are unwilling to cover the costs, one should wonder about the utility of the enterprise. In this important respect, we are strong believers in "market-place solutions"--provided that what economists call "externalities" can be captured.
JSTOR will, of course, need to receive subscription revenue in order to cover its costs, which we believe is an achievable goal once a proper scale of operations is attained. It is hoped--though, again, there are no guarantees--that the appropriation of $1.5 million for working capital approved by the Trustees of the Foundation in December will be sufficient to allow JSTOR to become self-supporting within roughly two years.
Finally, a word about staffing. The effective staffing of JSTOR is obviously critical to its success. At their first meeting in July, the Trustees of JSTOR elected Kevin M. Guthrie to the position of Executive Director. (Note 7) JSTOR has also retained the interest and commitment of Ira Fuchs, who serves as Chief Scientist of JSTOR while remaining Vice President for Computing and Information Technology at Princeton. Richard De Gennaro, who will retire as Librarian of Harvard College at the end of August 1996, has also agreed to work with Mr. Guthrie and Mr. Fuchs in the further development of JSTOR. It remains only to express our personal thanks to colleagues at the University of Michigan (especially Dan Atkins, Randy Frank, Wendy Lougee, Spencer Thomas, Kristen Garlock and Sherry Piontek) who continue to work so hard on the vast array of technical and library-related aspects of JSTOR.
* * * * *
In thinking about the future of JSTOR, we believe that we are still at the beginning of an enterprise that is full of promise, as well as replete with questions. It is also an enterprise that is seeking to operate within a rapidly changing environment. At this juncture, no one can predict the eventual outer boundaries of JSTOR, or whether, at some date, it will take on even broader functions itself or become part of another organizational structure with an even more ambitious scholarly mission. Under present circumstances, the appropriate "rules of the road" appear to be: proceed with all deliberate speed, but with at least one eye open for opportunities not even dimly perceived now.
In our work to date, we are grateful for the extraordinary help rendered by so many thoughtful people. We hope that JSTOR can continue to count on our many friends, and especially our friendly critics, for advice. Jacob Viner, the distinguished economist and discerning observer of academia cited earlier, was fond of reminding his students as well as his colleagues: "There is no limit to the amount of nonsense one can propound, when one thinks too long alone."