The Red Sage Project
An Experimental Digital Journal Library for the Health Sciences
Richard E. Lucier
Assistant Vice Chancellor, Academic Information Management
University Librarian
Director, Center for Knowledge Management
rel@library.ucsf.edu
Peter Brantley
Director, Academic Information Systems
brantley@library.ucsf.edu
Library and Center for Knowledge Management
University of California, San Francisco
D-Lib Magazine, August 1995
In August 1992, AT&T Bell Laboratories, Springer-Verlag, and the University of California, San Francisco (UCSF), the only one of the nine campuses of the University of California dedicated to the health sciences, entered into a collaborative agreement to identify and study the technical, legal, economic/business, and cultural/user issues surrounding the creation and delivery of biomedical information in an electronic, networked environment. This partnership was the culmination of several months of separate discussions between Bell Labs and Springer-Verlag, and Springer-Verlag and UCSF, as well as tri-partite negotiations. The Red Sage Project is the first effort of what is envisioned as a long-term collaboration among the three organizations.
In November of 1992, the founding partners of the Red Sage Project sought to expand participation in the project through the inclusion of additional partners from the publishing community. By March 1993, UCSF had signed Memorandums of Understanding with the following publishers: Academic Press; the American Chemical Society; the American College of Physicians; the American Medical Association; the American Physiological Society; the American Society for Clinical Investigation; Appleton & Lange; Blackwell Scientific Publishers; the British Medical Association; Cambridge University Press; Chapman & Hall; the Massachusetts Medical Society; the National Academy of Sciences; Nature Publishing; Oxford University Press; Rockefeller University Press; Saunders; and John Wiley & Sons. This group of publishers is a microcosm of the publishing world, representing U.S. and non-U.S. commercial publishers, scholarly societies, and university presses.
The overall goal of the Red Sage Project is to build, deploy, and evaluate a working Digital Journal Library of the Health Sciences at UCSF, which will serve as a laboratory for the collection of valuable information as we plan our future digital library and the accompanying transition. Within this context, each of the partners comes to the project with specific objectives. UCSF, aggressively pursuing its vision of a digital library1, is seeking to explore the technical, educational, support, and behavioral issues associated with the electronic delivery of primary journals to scientists and clinicians at their place of work. The participating publishers hope to gain an understanding of how they need to modify their print production processes to produce electronic versions of journals efficiently as well as understand and quantify the significant business and marketing issues associated with the electronic delivery of journals.2 AT&T is actively exploring its potential businesses in the preparation, storage, and delivery of scientific, technical, and medical information. Together, we hope to develop a business model for electronic journals that will be satisfactory to universities and libraries, publishers, and technology providers. The absence of such a model will seriously hamper efforts to make any digital library a reality.
Early on in the discussions, UCSF decided to focus the content of this Digital Journal Library of the Health Sciences in two areas: molecular biology and radiology. Several factors drove this decision. First, in order to generate sufficient use, we realized that it was essential to create a critical mass of content in the subject areas we chose. Springer-Verlag's journal list suggested these two areas as viable possibilities. Second, there are two primary faculty customer bases at UCSF: the basic sciences community and clinical practitioners and researchers. Cutting across all the basic sciences, molecular biology seemed an ideal choice. Similarly, radiology seemed a natural choice for the clinical constituency as radiologists are proficient users of advanced technology. Further, radiology would lay the foundation for experimental imaging work in later stages of the project.
As we talked with potential users, we also decided to add a new content area that we call "high impact". These are general journals in the basic sciences and clinical arenas that are of great importance to a wide range of people, and include such titles as the New England Journal of Medicine, the Proceedings of the National Academy of Sciences, and Nature. In total, we have seventy-one titles in the Red Sage Digital Journal Library. Most of these titles begin with the January 1993 issue. Initially, all content is stored as bit-mapped page images.
AT&T's RightPages was chosen as the software platform for this Digital Journal Library. Its look and feel is modeled on a traditional library metaphor. Journal titles are arranged in alphabetical order just as they are in current periodicals section of most libraries, with the most recent issues on top (Figure 1 [79 Kbytes]). RightPages is primarily a system which aids user browsing and alerting. Mouse clicks on journal images and tables of contents allow users to easily retrieve articles of choice (Figure 2 [64 Kbytes] and Figure 3 [113 Kbytes]). Users, able to establish custom profiles for subscriptions (Figure 4 [37 Kbytes]) and key words (Figure 5 [59 Kbytes]), are alerted when new content arrives through e-mail messages and special highlighting in the user interface (Figure 6 [70 Kbytes]). An underlying text data base also allows the user to do limited full-text searching (Figure 7 [64 Kbytes]).
RightPages is designed as a client/server system. The first client software available was for Sun workstations; UCSF technical staff ported that client to additional UNIX platforms, including Silicon Graphics and DEC. AT&T made a Macintosh client available in early 1995. Plans for a Windows client are on hold as access to the Red Sage Digital Journal Library through the World Wide Web is being actively investigated.
Library staff participated in the testing and enhancement of RightPages, especially the Macintosh client, as it was implemented at UCSF. They continue to inform development at AT&T as the WWW version is implemented.
The file server for the RedSage Project is a SPARCCenter 2000, made possible with generous support from Sun Microsystems, Inc. The SPARCCenter runs Solaris, Sun's version of the UNIX operating system, and has four processors, 640 MB of RAM and a high-speed FDDI connection to the campus backbone. A large disk cache allows new and recently used articles to be retrieved quickly. Also attached to the computer is a DISC Model 350 optical jukebox, providing over 250 gigabytes of permanent storage for our electronic journal collection. The jukebox can be easily expanded to a total capacity of 500 gigabytes, or half a terabyte.
The data management software underlying RedSage is EMASS' AMASS, which after many initial troubles has developed into a highly stable and reliable foundation for the RightPages client/server software. AMASS is responsible for keeping track of where all the files are stored, and of fetching them from the jukebox if they cannot be found on the local disk cache. The overall time required to display a page image ranges from less than one second to a maximum of seven seconds when the system is being heavily used and images must be retrieved from the optical jukebox.
AMASS is a preferred database engine over both Hierarchical Storage Management (HSM) software and commercial Relational Data Base Management Systems (RDMS). Although HSM software can easily handle the size of our storage requirements, it is not optimized for retrieval speed and user response time would have been inadequate. RDMS software, such as Oracle or Sybase, begins to experience strain with gigabyte size databases; furthermore, the full gamut of database functionality, such as transaction processing, sorting and indexing, and query optimization, is unnecessary in our current environment.
The Red Sage Project officially extends for four years, beginning January 1, 1993 and projected to end December 31, 1996. After a year of exporting RightPages from AT&T to the UCSF environment and importing electronic journal content, the Red Sage Digital Journal Library became available to faculty, staff, and students at UCSF in January 1994. Only the UNIX client was available, thereby limiting the number of potential users significantly. Use was possible both from within the Library and Center for Knowledge Management as well as any campus site with appropriate hardware and network connections. In early 1995, we began distributing the Macintosh client. That fact, along with an increasing content base, has caused the usage to steadily increase. The library's education staff have developed documentation and conducted training, as necessary. Both quantitative and qualitative data are being collected for analysis and evaluation. We are particularly concerned with the Digital Journal Library's impact on work habits and routines and what changes are being wrought in how users access and utilize the journal literature in an electronic environment.
In 1994, UCSF applied for and was awarded a grant from Pacific Bell's CalREN (California Research and Education Network) Foundation for the extension of the Red Sage Digital Journal Library. The goal of the CalREN Project is to determine the feasibility and practicality of a Regional Biomedical Digital Journal Library. An ATM line links the UCSF Library and Center for Knowledge Management with Genentech, a major biotechnology firm in South San Francisco, about 10 miles from the UCSF campus. Scientists at Genentech are now able to access the Red Sage journals. We are just beginning to collect usage and network performance data to identify and evaluate the issues involved in UCSF serving as a regional knowledge server.
During the coming year, several important steps will be taken by project participants. In the fall of 1995, we expect to move to direct electronic input from publishers, a major step forward. SGML and other standards have been agreed upon by the publishers and many have altered their production processes to accommodate them. At about the same time, we will begin offering access to the digital journals through the WWW. We will also continue to offer access through the current RightPages platform. An important feature of the WWW implementation will be linkages between the journals and Medline, a database which indexes much of the important biomedical literature.
In January 1996, UCSF will begin providing access to faculty, staff, and students at UCLA, in a collaborative project with the UCLA Biomedical Library. Most importantly, proposals are being formulated and reviewed by all participants for the eventual commercialization of this experimental Digital Journal Library of the Health Sciences as we seek to ensure its continuation past the end of the project.
1. Lucier, R.E. "Building a Digital Library for the Health Sciences: Information Space Complementing Information Place." Bulletin of the Medical Library Association. July 95
2. Lucier, R.E. and Badger, R. C. "Red Sage Project." The Serials Librarian. 24:3/4, 129-134, Spring 1994.
hdl:cnri.dlib/august95-lucier-brantley