CrossRef Considering Full-Text Search Service by
Paula J. Hane
Posted On May 28, 2002
CrossRef (http://www.crossref.org), a publisher collaborative that's operated by the Publishers International Linking Association, Inc. (PILA) and enables researchers to navigate online journals via Digital Object Identifier-based (DOI) citation links, is currently considering the implementation of a full-text search service. While discussions of such a service have been circulating since late last fall, CrossRef members have recently been officially informed that the CrossRef Search Project is under consideration and that a full proposal will be presented at the group's annual meeting in September. If the new service is approved, it would be developed with separate financing and would not divert resources from the existing reference-linking service.The CrossRef Search Project would aim to create a broad high-quality, full-text, cross-publisher search service. Participation would be optional for CrossRef members. This service could provide an opportunity for publishers to add value and make the access of full-text content easier for scholars. Each participating publisher would provide a search box on its Web site for users to search the full text of all articles from other participating publishers. This search function would be funded by the publishers themselves. Search results would use the DOI to link users to the full text at publishers' sites. These publishers would control access to the content and payment options, as they do now with DOI reference links. CrossRef would crawl the full-text data only for indexing (not for displaying or using in any other way) but not store the full text. CrossRef has already conducted a market research and feasibility study of the search project, which included input from primary and secondary publishers, librarians, scholars, and vendors. The response from these groups was reportedly very positive—if the service were able to meet certain criteria. According to Ed Pentz, CrossRef's executive director, the criteria include good classification technology, cross-disciplinary indexing, performance and scalability, and full-text search capabilities. He noted that the search technologies required for working across full-text scientific content are very different than working with just Web page content. The group is looking at vendors now but has not yet chosen one. Pentz said that a prototype service might be initiated to advise members about the costs and implementation issues. CrossRef, which was established in 2000 by scholarly publishers as an independent, nonprofit entity, uses open standards. It's an official registration agency of the International DOI Foundation (IDF) and is the first full-scale implementation of the Digital Object Identifier, a system for persistently identifying and exchanging intellectual property in the digital environment. The underlying technology for DOI, called the Handle System, was designed by the Corporation for National Research Initiatives (CNRI). The initial focus of CrossRef was to allow primary publishers to link their references to other primary publishers. It then enabled full-text links from secondary abstracting and indexing (A&I) databases to full text at publishers' sites. CrossRef plans to incorporate additional reference content such as encyclopedias, textbooks, conference proceedings, and other relevant literature. While some see the proposed project as quite a change from the service's original goals, Pentz says that CrossRef Search is consistent with CrossRef's goal of serving publishers and researchers by facilitating cross-publisher navigation. Currently, 122 publishers participate in CrossRef (up from 113 in April). This accounts for over 6,300 journals with over 4.8 million article records in the database. There were also 1.2 million DOI resolutions (i.e., users actually clicking DOI links). Nature, Science, ScienceDirect, and AIP are all actively linking with DOIs. A variety of organizations participate in CrossRef. Libraries use the system as part of their localized linking solutions to enrich online catalogs and databases with links to their own full-text holdings where appropriate. Intermediaries, including secondary publishers and journal-hosting services, use CrossRef as affiliates and agents, thus enhancing their own products and content offerings with DOI-based citation links. If the proposed service can provide a high level of full-text search capabilities, including indexing and categorization, it could have an impact on the perceived value and revenues of the A&I and aggregator services. "Information professionals may want to thoroughly test the promised functionality of CrossRef Search to see if such a system matches the enhanced information retrieval provided by the A&I community," said Jill O'Neill, NFAIS's director of planning and communication. "Any form of electronic retrieval must be viewed from the perspective of providing the best means possible for meeting the needs of a community of users." Others see the implementation of such a service as more controversial and complex, especially in the relationship between the primary publishers and the secondary services. Dick Kaser, former executive director of NFAIS and now vice president of content for Information Today, Inc., said: "When CrossRef first appeared on the scene, some database producers were worried that the initial production system for facilitating linking would ultimately go live and be accessible to the public for searching. Though the original founders insisted this would never happen—and in concession to secondary publishers, the metadata set was limited to just the items need to create links—it would now appear that the concept is undergoing the natural evolution that some early critics feared. Not that there is anything wrong with offering consumers a choice. But, as CrossRef publishers gravitate to where they think the money is, they may be surprised to learn that bibliographic databases have not made an easy living off providing cross-publisher search services. CrossRef is about to learn the e-facts of life." For additional information, including details of how the CrossRef system works, see the FAQ on the group's Web site (http://www.crossref.org/faqs.htm). See also Jill E. Grogg's article "Thinking About Reference Linking" in the April 2002 issue of Searcher (http://www.infotoday.com/searcher/apr02/grogg.htm). When told of the CrossRef Search Project, Grogg said: "This is an impressive project. It looks like all the key players might actually come together on this to help users." Other Developments Technology companies are currently creating software tools that interface with CrossRef. A series of nonexclusive strategic agreements with developers of library applications and new technological initiatives is furthering CrossRef's goal of providing libraries and researchers with seamless integration. Using SFX technology from ExLibris, libraries and information centers can redirect CrossRef links to the full text of article copies licensed by their institutions, thus addressing the "appropriate copy" problem. According to Pentz, Japanese journals will soon be added to CrossRef. "Membership and agent agreements were signed with the Japan Science and Technology Corp. (JST)," he said. "JST's J-STAGE hosts journals for Japanese societies, and they will be depositing for 30 journals initially. JST will also be adding DOIs to their JOIS secondary database. We are very pleased to have this international participation."
|