Summit on Serials in the Digital Environment

E-Journal Access Using the Catalog, Federated Search, and Reference Linking Systems

Karen Calhoun, Cornell University Library

Libraries are beginning to work with ILS vendors to design and implement unifying systems that help users discover, navigate, and connect to what they need within the present rich but confusing array of Web-accessible and print resources.  One such partnership was the ENCompass co-development project that began in May 2000 between Cornell University Library and Endeavor Information Systems.  ENCompass is sometimes called a portal application.

Cornell project overview and further information: http://encompass.library.cornell.edu/ENC_FAQ.htm

List of portal application functionalities (draft, LC): http://www.loc.gov/catdir/lcpaig/PortalFunctionalitiesList4PublicComment1st7-22-03.html

Portal development and the technology of the online scholarly information environment are creating new, closer ties between the data and metadata produced by libraries, publishers and other content providers, A&I services, vendors, and other intermediaries.  Portals pull e-journal metadata from vendors, the catalog, Web lists, federated search systems, and reference linking tools into new, complex relationships.  Robust library systems require the ability to reuse, repurpose, and integrate metadata from many sources.  Without this ability, they will not be capable of delivering what 21^st century library users expect: a fully interlinked information environment.

Library technical services experts have traditionally produced records describing the serial literature at the title level.  Writ large, serials control and access is also about discovery and access to the serial literature at the article level.  While continued application of the principles of shared cataloging and serials control remain crucial, it is also true that reference linking and the OpenURL (a standard way to send metadata to a link resolver) are making it more common for users to move directly from online citation to e-journal article full text, in the process bypassing the library catalog and the serial records in it. 

NISO Committee AX: OpenURL framework for context-sensitive services: http://www.niso.org/committees/committee_ax.html

In this new environment, it is appropriate to revisit the assumptions underlying the cataloging rules for describing serial publications in library catalogs—more specifically, the rules applied to e-journals.  How much should be included in a catalog record for an e-journal?  What data elements are essential to users?  Are all the data elements required by the cataloging rules for serials necessary for e-journal records?  How much of the record creation and maintenance process can be automated by re-using e-journal metadata available from third parties?  What is the role of administrative metadata, and how should it be stored and linked to descriptive metadata?

In July 2002, the Cornell library’s Technical Services Executive Group charged an Electronic Journal Maintenance Task Force to recommend (and later implement) new strategies for creating and maintaining e-journal records.  They decided to provide title access to large numbers of the library’s e-journals from aggregators by creating and adding to the catalog brief, machine-generated records, called “sleek” records. 

In early January 2004, Cornell had 21,363 e-journal records.  Of these, just over 11% (2,443) are not machine-generated; 2,698 are from ProQuest; most of the remainder are sleek records created with metadata from third parties (e.g., Serials Solutions).  The library uses separate records for each expression of an e-title.  Machine-generated records contain uniform titles with qualifiers for the name of the aggregator/provider to help distinguish them from one another.  Using coding added to the record, a program identifies the entire set of e-journal records for extraction/reuse in the library’s Web title list of e-journals. 

The methods and computer programs implemented as a result of the task force’s work have eliminated the need for human maintenance on 90% of the e-journal titles while also enabling more timely delivery of data to library staff and users.  The system is still evolving but plans are to refresh the entire 21,000+ record set on a bimonthly basis. The automated maintenance process is sufficiently flexible that should full MARC records become available for all or a subset of the titles, the sleek records could be replaced. 

Cornell University Library.  Electronic Journal Maintenance Task Force: preliminary report, 30 October 2002.  http://www.library.cornell.edu/staffweb/TSEG/e-journalPrelimReport41.html

Cornell University Library.  E-Journal Maintenance Task Force: final report, 25 April 2003.  http://www.library.cornell.edu/staffweb/EJMTFFinal.doc