Walt Crawford

Open URL: Standards Can Be Fun!

By Walt Crawford
American Libraries Columnist

Senior analyst, Research Libraries Group

Column for August 2002

Recently, I’ve been testing OpenURL in Eureka, the Research Libraries Group’s Web-based search interface. I’ve seen how much OpenURL can benefit libraries, their users, and database suppliers—and I’ve had more fun testing OpenURL than I’ve had at RLG in years! Read on, even if your library never plans to use Eureka; this discussion applies to any OpenURL– enabled database.

What is OpenURL?

Your library or consortium has scores of databases and possibly thousands of journals in full-text form from various content aggregators. It also has many books and journals as well as access to consortial and regional catalogs showing even broader print holdings. OpenURL links article citations and other items in enabled databases to full-text resources, local print holdings, and other features that enhance access.

With OpenURL, your library or consortium runs a “resolver”—software that knows your site, the full-text or other resources you have available, and how to reach them. When a user clicks on a link, the resolver identifies what services are available for that particular item at your institution and provides the tools for whatever next step the user chooses.

If that sounds as boring as most standards-related stuff, that’s because you haven’t tried it. I wrote the specification for OpenURL in Eureka. As part of my testing, I’ve performed several thousand trials using Eureka OpenURL to reach resolvers at four of our member institutions and the demonstration programs of two software companies.

“Testing” doesn’t accurately describe this process; “methodical play while keeping accurate records” is more like it. I wasn’t expecting it to be this much fun. RLG began work on OpenURL shortly after the 2002 Midwinter Meeting in January; it went into production April 12.

Eureka supports more than a dozen citation and bibliographic databases totaling more than 100 million records. RLG does not offer full-text journals. With OpenURL, I see good reasons for RLG to avoid the full-text journal business.

Why OpenURL matters

I did two topical searches in FRANCIS, a French humanities and social sciences database, and checked the first 15 records in each result, using the resolver from one of our trial institutions. For 20 records, I found full text that came from 14 different licensed sources, with many choices for a given article from JSTOR, Ingenta, Elsevier, DowJones, Academic Index, the American Chemical Society, and so on.

It makes sense for established index and citation databases to link to existing full-text resources rather than ask libraries to license the same sources yet again. OpenURL makes that easy, and it might help libraries minimize duplication of resources in future negotiations.

If properly implemented, OpenURL is a win-win situation. Good abstracting and indexing services become more valuable by linking to local resources. Licensed resources and print holdings see more use because the link from identification to holdings is fast and easy. None of this requires fancy new numbers; the information is already there—ISSN, journal and article titles, year, volume, and so on.

I don’t want to minimize the difficulty of adding an OpenURL resolver. RLG doesn’t have such a resolver and I don’t know how difficult or expensive they are to implement. I suspect that growing competition from at least three library vendors will make them easier and more useful.

While Z39.50 continues to be a vital standard, attempts to use it to reverse the complexity of different databases and catalogs have not worked very well, partly because Z39.50 is a pure computer-to-computer standard.

OpenURL is different: It adds natural intelligence to the computer-to-computer process to navigate complexity. OpenURL resolvers don’t take metadata and immediately hand off whatever resource comes up—a process that will frequently fail because the metadata is incomplete or because the source database and the object resource follow different rules. Instead, resolvers apply as much logic as they can, then offer the user a menu showing a range of possibilities.

OpenURL involves the user’s natural intelligence in navigating the complex field of library resources, both online and in print. That offers more power than a pure computer-to-computer protocol—and, well, it makes it more fun.

For more information on OpenURL, visit the NISO Committee AX Web site. For more on Eureka, visit www.rlg.org.