Institutional repositories of digital data at universities and other research institutions may now receive deeper, more thorough indexing and full-text delivery through Elsevier's free, sci-tech search engine, Scirus (http://www.scirus.com). The Scirus engine already reaches content at many institutional repositories, but those joining the new Scirus Repository Search service will receive more extensive and sophisticated indexing of a wider range of content. The repositories will also have access to additional search capabilities on their own Web sites at no cost. The first university to join the Scirus Repository Search service is the University of Toronto's T-Space collection (http://tspace.library.utoronto.ca). All of T-Space's digital files and data are available to the open Web.
Marshall (Peter) Clinton, director of information technology services at the University of Toronto Libraries, said that a similar arrangement with Google preceded Scirus' arrangement by several months. He estimates that both Scirus and Google's improved service has improved access for both on-campus and off-campus users of the T-Space site.
Ammy Vogtlander, Scirus' general manager, attributes the development of the Scirus Repository Search service to the fact that "Elsevier understands that an increasing amount of valuable content is currently held in academic repositories." She indicated that working directly with institutional repositories would allow Scirus to reach unique metadata and full-text material. It will also allow Scirus to reach content in alternative formats to journal articles or reports. For example, Scirus has already indexed the full text of T-Space's complete repository, including not only articles, technical reports, and preprints, but also presentations, data sets, and multimedia digital objects. In addition to serving campus users with improved retrieval, Scirus' Repository partners will have their search results branded to identify their content in results lists that go out to the open Web.
According to Vogtlander, "We already had full indexing of various sites and institutional repositories, but now, for participating repositories, we will target key reports, have higher quality indexing, better display of results, and more accurate metadata." She found it odd that some institutional repositories, for all their important content, "offer no full text, only metadata. Scirus outperforms them often. It's time to make the technology really work. Institutions need better search technology." Clinton agreed that repository support from outside vendors offered "better penetration. Libraries with institutional repositories are hard pressed to do metadata harvesting."
Toronto's T-Space integration into Scirus has been in operation for over a month. Although Vogtlander had no specific before-and-after statistics, she guesstimated that before the Scirus Repository Search service went into effect, Scirus reached some 75 percent of T-Space information; she guesstimated it reached 95 percent after.
The new alternative formats for data will stretch beyond Scirus' current non-HTML/XML formats, e.g., PDF, Postscript, PowerPoint, etc. However, for now, Scirus will not offer special handling of images, programs, or data sets beyond better metadata.
T-Space runs on the DSpace open source software that was developed by MIT and Hewlett-Packard and that is installed on some 125 institutions worldwide. However, Vogtlander indicated that the Scirus Repository Search service would operate independently of platform, although the large research collections represented in DSpace would make logical targets for the program. The University of Toronto's Libraries also subscribe to Elsevier's Scopus, a fee-based service. Though Vogtlander assured me that they were not limiting to or targeting Scopus libraries for the new service, she did admit that they would never fail to mention when a Scopus library joined. Clinton said that Toronto's support of Scopus had "a great influence on pushing Elsevier to broaden Scirus to do more content and let us search social science and humanities databases, plus local content." Vogtlander confirmed that Scirus planned to expand its social science and humanities sources, although it remained leery of the arts.
When asked whether Scirus planned to restrict the Repository Search service to academic institutions, Vogtlander said that it would expand to any large institution or nonprofit that had valuable scientific content that it wanted opened to the Web. The Scirus Repository Search service does not enhance material that is restricted to campus users and is therefore unavailable to open Web users.
In contrast, Google Scholar has introduced linking to "appropriate copy" or restricted access content. (See "Library Collections Linked on Google Scholar for Free," http://newsbreaks.infotoday.com/nbreader.asp?ArticleID=16204.) When asked about Google Scholar's clustering and linking, Vogtlander said that Scirus is considering clustering. For now, however, Scirus users will see multiple results ranked on frequency of terms and date. Scirus also can't handle Open URL linking to library-licensed content. On the other hand, Scirus does access masses of metadata and abstracts for subscriber-accessed content.
The specter of Google doesn't worry Vogtlander. Even at campuses that use the Google Appliance, she feels that Scirus has a lot of added value to offer. "We don't just search any one site, but cross-search for different departments and even within different countries." Clinton also leads the Scholars Portal, a service that reaches beyond the University of Toronto to all Ontario's university library collections. At present, according to Clinton, they are "in the nascent stages" of working to add improved searching from Google and Google Scholar that reaches first three, but someday all, of the academic institutional repositories in the province.
Vogtlander seems to see the world in different terms from traditional Elsevier. In reference to the policy of not charging for Scirus' services, even the new Repository Search service, she stated: "We must understand the free service business. There are different business models now, and Web searching is seen as free." As to content, she said: "There is relevant content that is never published. Publishers and libraries focus too much on published content. Users want more, including research that never made it into journals, so we at Scirus include non-published research. More than half of research never makes it into journals." Clinton has the statistics to prove it. He said: "[T]oday only 2 percent of searching is done anywhere near a reference librarian. Only 8 percent of journal use is in the library, while in 1997 [and] 1998, one-third was in the library, one-third on campus, and one-third off campus." He confidently stated that "our libraries are busier than ever, but they're doing different things."
And apparently Scirus is ready to help at a very nice priceófree.