NASIG Newsletter

TACTICS SESSION

Improving OpenURL Metadata
Glen Wiley, Cornell University
Reported by Wm. Joseph Thomas

The creation and adoption of the OpenURL standard solved several important problems for libraries.  The standard permitted end users to navigate from one resource, such as a citation in an online index, to another, such as the article indexed, without requiring them to execute a second search in the destination resource. By reducing linking dead ends, providing multiple access points to content in a single location, and connecting users to the most appropriate copy of the content, OpenURLs improve the visibility of resources. This improved visibility increases the usage of the library’s selected content, and may also reduce document delivery costs. Yet, problems with OpenURLs remain, and can be difficult to track and correct. To address some of these problems, librarians at Cornell University have begun working on a study of OpenURL metadata quality. Their goal is to create a systematic approach to measuring and validating metadata quality so that libraries and content providers may collaborate on solutions for metadata problems. One member of this project team, Glen Wiley, shared details about the types of problems this approach seeks to remedy, and the status of the project.

Project team members began by reviewing a log file provided by Cornell professor Éric Rebillard, director of graduate studies in the field of classics at Cornell University, and editor of L’Année Philologique on the Internet. Next, inspired by the Metadata Quality Evaluation for the Open Language Archives Community, team members started mapping OpenURL data elements to a matrix.  This process would generate a score for each OpenURL passed to the server. Failures have occurred as a result of errors in both semantic and syntactic features. Some of these were identified five years ago by Miriam Blake and Frances Knudson in their 2002 “Metadata and Reference Linking” article. Errors in passing ISSN; date and chronology identifiers; and lack of UTF-8 encoding; as well as inconsistencies in recording volume, issue, and start page were common reasons for failure to resolve. By mapping OpenURL elements to a scoring matrix and sharing the results, project team members hope to generate quantitative measures with specific feedback to which content providers can respond.

Problems in OpenURL metadata are widely perceived among librarians, as captured by responses to a survey that was part of a 2007 United Kingdom Serials Group report, “Link Resolvers and the Serials Supply Chain.” The potential magnitude of OpenURL problems grows in relation to the expanded volume of OpenURL traffic. Consider the number of OpenURL requests for just one university in a relatively short time period: between December 3, 2007 and February 8, 2008, the Cornell University Library link resolver received more than 53,000 OpenURL requests. Therefore, project team members hope that a jointly-developed approach like this can benefit libraries, service providers, and content providers. Currently, librarians at Cornell are completing the matrix for mapping and scoring OpenURL metadata quality, writing software to automate as many of the processes as possible, seeking long-term funding, and trying to determine how much the project can scale.