Vancouver, British Columbia
May 15-17, 1996
Susan F. Rosenblatt, Deputy Director of Libraries
University of California at Berkeley
The SCAN Project (Scholarship from California on the Net) is a collaborative project between the University of California Press and the Library at UC-Berkeley as well as the Division of Library Automation of the Office of the President and the libraries at UC-Irvine and UC-Los Angeles. SCAN, a pilot project designed to facilitate broad scholarly access to humanities journals, books, and related materials through publication on the Internet, has been one of the most interesting projects the Library has been involved in, and has greatly strengthened the relationship between the Press and the Library. The ultimate purpose is to draw together the resources of the university community to explore the potential of technology to support new methods of scholarly communication in teaching, learning, and research. It furthermore is designed to explore core issues relating to the future of scholarly publishing and university presses. SCAN has received generous funding from the Mellon Foundation, and is hosted on Berkeley’s Sunsite for Digital Libraries, donated by Sun MicroSystems.
SCAN’s goal is to develop an economically viable publishing model for humanities scholarship that integrates electronic dissemination, library access, and scholarly use.
Progress to Date
We have created a prototype electronic edition of an existing print journal, Nineteenth-Century Literature, to test networked access from remote and local workstations, ease of use, searching tools, and cost recovery mechanisms. Five years of back issues of Nineteenth Century Literature have been mounted as Nineenth-Century Literature—Electronic Edition; current issues are mounted as they are published.
Later this year we will be mounting the full text of several books in nineteenth century studies, as well as primary source materials. Additional journals, books, and primary materials in nineteenth-century studies, history, and classics will be mounted in 1997.
SCAN has four major objectives:
It will mount on the Internet a base of electronic humanities journals, monographs, and digital reproductions of primary source materials in literary studies, classics, and history. This base will ultimately form the core of a database from which pieces can be combined and repackaged by individual users to meet a variety of specialized user needs.
In order to achieve the general goal of creating a large coherent database, the partners will develop and implement a suite of consistent Standard Generalized Markup Language Document Type Definitions (SGML DTDs) for these materials by using and modifying existing DTDs as necessary.
The project will conduct user studies among faculty and students to evaluate the use of SGML-based authoring tools, end-user systems for sophisticated searching and navigation, and the most appropriate content and access structures for such online information.
Cost recovery experiments will monitor and document both costs and ways for publishers to charge for access to electronic journals, monographs, and databases.
In planning and carrying out the project, the Library and the Press wished to explore five general assumptions:
First, we believed that the complementary missions, roles, and operational capabilities of the Press and the Library would strengthen the project and increase the likelihood of developing sustainable new scholarly paradigms for the Press and Library alike.
Second, we believed that the economic pressures faced by both the Library and the Press threaten the process of scholarly communication itself, and that these pressures could best be mitigated through collaborative efforts.
Third, we believed that new technologies offered possibilities to improve the delivery of scholarly information; to reduce its cost not just to libraries, but also to end-users; to protect the scholarly monograph; and to sustain the economic viability of university presses.
Fourth, we believed that dissemination of research can be broader, more rapid, more sophisticated, and more integrated, and that improvements in the dissemination of scholarly information in the humanities can improve both research and teaching.
Fifth, we believed it likely that in the electronic world the roles of author, publisher, librarian, and user of scholarly texts might begin to change, and, in some cases, to merge. We wanted especially to begin exploring new definitions of these various roles, particularly those of publisher and library, and to learn how the synergies of collaborative endeavors might change the entire process of scholarly communication.
We are now almost half-way through the project, and can begin to test these assumptions and others against some real-world experience.
Today, I would like to focus on three themes:
The economics and costs of the project—what we didn’t know when we started, but have begun to learn.
How technology helps and hinders our progress, especially dealing with legacy systems, being overtaken by events, and how the need to choose among technical alternatives forces a reevaluation of values.
The partnership itself: Does collaboration work? Is there enough congruence of mission and values between the Press and the Library to make the partnership worthwhile?
We started the project with the assumption that electronic publishing could help both publishers and libraries to respond to economic pressures in the print world of scholarly publishing. There were several ways in which we thought this would happen.
We expected that the effective use of technology would reduce production and distribution costs incurred by the Press, that electronic publishing would expand the market for scholarly journals, that new pricing models would reduce costs to libraries, and that, possibly, new subsidiary and derivitave products could be developed through the inclusion of primary resources in the database created for the project.
As of now, production costs for the Press have only increased through the use of technology, and clearly we underestimated the cost and pace for implementing an operational system for online scholarly publishing. The technological efficiencies that the Press had, over the years, put in place to reduce production costs for the print journals have actually increased the costs of creating a digital publication workflow, and the Press must now either make sizeable one-time investments to retool the print production stream or retain both production streams as long as there is a print edition. This is a Hobson’s choice that compromises the cost-effectiveness of online publishing as long as the necessity of retaining both forms of publishing continues.
We also underestimated start-up costs. This was particularly serious for the Press, but also a problem for the Library. The difficulties of implementing a research project to explore the transformative powers of technology at the same time that both organizations were undergoing unprecedented budgetary retrenchments cannot be overstated. In many ways, the Library’s problems with this project were not as severe as those faced by the Press. Most importantly the gift from Sun Microsystems reduced the capital costs we had expected to contribute to SCAN by subsidizing the hardware for the project. In addition, since our externally-funded digital library projects complemented our SCAN project commitments, meeting the new project’s goals has been congruent with our general direction in digital library research and demonstration. In contrast, for the Press the SCAN project was its first real-life, real-time, digital project. Moreover, the stakes are greater for the Press, as its very future depends on its ability to reduce costs, develop new markets, and support research.
Both organizations are understaffed and underfunded; to test fully the transformative possibilities of technology for scholarly publishing, we should have invested far more heavily at the outset to redesign the entire production flow, capitalizing on technologies for electronic publishing without jeopardizing the quality or increasing the ongoing costs of the print journals. But the downsizing and re-engineering processes that both the Press and the Library have undertaken as a result of six years of budget cuts reduced our flexibility to invest budget resources in or to add staff to the project.
Since the “real-world experience” of scholars using digital libraries will require a critical mass of information online, this project can only hint at the possibilities for changes in the patterns of use of scholarly information by scholars, in the introduction of new markets, and in pricing models based on use. If there were money to move more decisively and rapidly with a large-scale project, we would also have a better test-bed for the use studies and, most importantly, for the economic modelling itself.
Despite the small scale of the project, we are beginning to gather information about comparative usage of the electronic and print editions. Although one of our key assumptions was that electronic publishing would generate new markets for the product, we have been surprised at the results to date. We have found that usage of the electronic version of the journal is vastly greater that that of the print edition. The online version receives between 5,000 and 6,000 “hits” per month, while the print version receives approximately 40-50 recorded uses each quarter (it should be noted that print edition usage cannot be measured as reliably as can online usage). Implicitly, usage of the print version is primarily by faculty and students, since the general public is a very small portion of our clientele. In contrast, a significant proportion of the users of the online version come from non .edu domains. In addition, approximately 20% of the usage of the online edition comes from abroad.
At present, we have only quantitative measures of use; in a later phase of the project, we will conduct interviews and create focus groups to ascertain how patterns of usage may be changing.
Since use of the online version is currently free of charge, we cannot say for certain that there are expanded markets for the online scholarly journal. It is possible that the demand will be quite elastic, and usage will decline significantly as charges are imposed. However, the preliminary data would indicate a strong likelihood that there will be a greater volume of use for the online journal than there has been for the print version—and therefore a larger market over which costs can be spread. Licensing of the online journal will begin soon, and we will then be able to test various pricing schemes to learn how we might best capture a large base of users and generate revenues.
The SCAN project has dramatically demonstrated to us that, paradoxically, even the most efficient use of technology can increase costs and slow the process of change. The Press had used automated production systems for many years, and these systems were highly customized to the needs of the individual journals and the capabilities of the printers. Since the text of the print journals was in machine-readable form, we anticipated few problems creating online versions from the print source. Unfortunately, not only has it proven very difficult to create the online journal from the print production stream, the Press has concerns about compromising the quality of the print journal either through revision of the print production stream to facilitate the online publication or by changing the print production to make it a by-product of the online publication. In retrospect, we should have anticipated the problems that these legacy systems would present.
Not only have we had to deal with the legacy systems for the print journal, we have had to develop migration strategies for the electronic edition as well. When we began planning SCAN, Gopher was the latest net technology, Berkeley had invested considerable staff resources in implementing a Library Gopher, and the Press was beginning to publish electronic journals on Gopher. Because of the commitment of both the Library and the Press to structured text, we planned to use SGML as the formal publishing mechanism, although we knew that we would need to support Gopher-based text access as well in the short run. Between the planning and proposal-writing stages and the present, the World Wide Web has overtaken Gopher, and PDF has become an industry “standardÓ#148; for publishing. In less than two years, we have had to develop migration strategies from plain text Gopher, through HTML, to SGML, and must now consider the advantages of PDF. We now know that migration plans have to be an integral and ongoing part of all of our strategies.
The technology is moving faster than we can implement it, and in some cases costs more than the community can afford. In addition, since each alternative offers a different mix of advantages and disadvantages, the Press must continuously reevaluate its values and strategies.
The recent emergence of PDF is a good case in point. The Press has been committed to the use of structured text for its publications because it believes that scholars will be better served by the searching and navigation potential it offers. A deficiency in the current technology for structured text is that the Press has less control over the “look and feel” of the product, one of the values that the publisher adds to the scholarly work. In contrast, PDF provides a stable visual presentation, can easily be used to make printed copies, and is less expensive to implement for electronic publishing. Moreover, the use of PDF could partially solve the print production legacy system problem. The Press must therefore make difficult choices about scholarly and publishing values, costs, and implementation strategies.
Likewise, the problems of converting backfiles to digital form can now be relatively easily and inexpensively addressed through JSTOR, an option that did not exist when we planned SCAN; current issues could be published in PDF, HTML, or SGML. Despite the ease and low cost of JSTOR, the Press would prefer to publish both backfiles and current issues in a single system using structured text that users could easily navigate through with a consistent interface.
Distribution options are a third example of the proliferation of technologies that has occured since we began SCAN. Our original intention was to use the Library’s server for SCAN, with the expectation that the Press would ultimately obtain its own server and assume ongoing responsibility for distribution, archiving, etc. The Library is quite willing to distribute the Press’ electronic publications indefinitely, although the Press would still prefer to have control over its own production and distribution. Just emerging are attractive third party distribution and archiving services which we must now analyze from cost/benefit and service perspectives.
The need to constantly reevaluate the technical choices we have made, to develop migration strategies to enable changes in preferred technology, and to rethink our goals and values as a part of technology evaluation has been more timeconsuming—and challenging—than we expected.
Developing a closer relationship with our partners at the University Press has been one of the most rewarding aspects of SCAN. We had thought that there would be sufficient congruence of mission and values between the Press and the Library to ensure a successful project. We share a commitment to preserve and make accessible the fruits of scholarly inquiry, and are committed to lead in efforts to insure that scholarly publishing survives. Through the project we have developed a much more sophisticated understanding of one another’s values, commitments, problems, vision, strategies, and strengths. We have found that we have complementary knowledge, skills, and technical expertise. The Press’ expertise in publishing, marketing, and scholarly communication helps the Library in its “publication” of digital library resources. The Library’s experience in running production systems and its track record in carrying out research and demonstration projects using structured text and developing “best practices” for encoding and navigation can help the Press to move more rapidly to an electronic publishing environment. The Press’ interest in and commitment to conducting effective research on the economics of electronic publishing will help both of us. Although there is a significant overhead cost in cooperation, the end result is worth the investment. By sharing our expertise and knowledge—and our vision for the future with one another—we avoid mistakes we would otherwise have made.
We have found that we have complementary needs that can be more cost-effectively addressed in concert. For example, we both need to have secure systems for authentication (of users and documents), security, electronic commerce, digital signatures, and rights management. We are finding that our roles are changing—and would have even without SCAN, but SCAN helps us to formally develop those changed roles. For example, the Library is becoming an author, editor, and publisher—initially of web resources and digital surrogates of primary resources. It is also becoming a distributor of both its own and the Press’ publications. The Library is learning from the Press how to become a publisher. The Library has thus far made more fundamental changes in organizational structure to support digital library initiatives than the Press has, therefore, Library staff can assist with mark-up and other tasks for which it is well-organized. Both the Library and the Press have learned from SCAN how far we have to go to retool our organizations to be effective in the digital world, and how great the staff training and development needs are if we are to have sufficient numbers of staff with the requisite technical expertise coupled with an appropriate level of understanding of scholarly publishing and sufficient research library knowledge.
We continue to gain important insights about ways in which our vision and values are not shared, and these differences will affect both how we interact with each other operationally, and how we develop mutually beneficial strategies for preserving scholarly publishing and research libraries.
In sum, SCAN has been an important and rewarding learning experience for both the Library and the Press—and it has been more fun that we thought it would be. Through the project we have a better understanding of the system of scholarly communication, and a greater commitment to working together to develop solutions that will preserve it.
Copyright © 1998 by Susan F. Rosenblatt