Should academic libraries expose bib data as linked data?

Posted by laura on September 27, 2011 under Semantic web | 2 Comments to Read

There are many calls for the library world to get their act together and expose the bibliographic data in their catalogs as linked data.  But should they?

It  only makes sense if your catalog has a lot of unique data about your unique local holdings and those records aren’t represented in WorldCat.   If you’re in that boat, this blog post doesn’t apply.   I would bet that most academic libraries do their original cataloging with OCLC tools and then add those records to their local catalog.   If your stuff is in WorldCat then let OCLC do the work.   They are doing a pilot to release 1 million of the top WorldCat records as linked data.  Eventually, this is going to be broadened.

Of course, there are problems in simply waiting for OCLC to do it.  The biggest is rights and licensing.  OCLC could make linked bib data available to it’s membership.  But opening that linked data to the world is another matter entirely.        Roy Tennant, commenting on June 2  at JOHO the blog (aka David Weinberger’s blog), says:

This is a project that is still being formed. The intent of the project is to investigate how to best release bibliographic data as linked data that will provide an opportunity for us to get feedback from other practitioners about how well we’ve done it and how useful it might be. A part of this project will include considering the policy and licensing issues. There’s not yet a firm timetable, but as that is determined it will be shared with the community. We’re encouraged by the enthusiastic response folks have shown in what we’re doing. It will help influence the shape of what we ultimately do.

You may own your records.  But if they’re aggregated in WorldCat, OCLC may or may not release them depending on what licensing/usage model they come up with.

The other problem, of course, is the waiting.  Nobody knows how long it’s going to be before WorldCat is a linked data resource.  If you’re gunning to get your library catalog out there in the linked data ether you’re not going to be satisfied with the time lag.   If you’re in an organization with limited resources, however, waiting will be to your strategic advantage.  Times are tough.  Budgets are small.  Cataloging departments are understaffed.
Why spend resources when  bibliographic records as linked data will come?

Focus instead on authority records and on enhancing your repository.  This is the metadata unique to your organization.  This is where you add value for your customers.  The value for others is a nice side-effect.  Faculty, researchers,  and newly hooded PhD graduates will  require identity management tools to assist them in scholarly communication.  Think beyond bibliometrics for their publications.  Grant makers will want to track researchers.  Universities will want to track research impact.   The ground is being sown.  Witness the NSF grant to U. Chicago and Harvard announced today, which will be used to research the impact of ORCID on science policy (i.e. the ability to put it into use for things like FastLane).

When you hear the calls for libraries to expose their ILS as linked data, consider how your users are getting bibliographic information.  Most likely from the web.  WorldCat feeds into Google.  OpenLibrary, LibraryThing, and other sources of bib data abound.  They’re probably going to your catalog for holdings, if they’re going there at all.    Bibliographic linked data is a good thing.  I just don’t think it’s realistic for the academic institutions sans lots of unique stuff in Worldcat to heed the call.  I’d rather that the calls for academic libraries to participate in the linked data movement get broader.   It’s not about freeing the ILS.  It’s about pushing out metadata that exposes the work of the people in your organization, not the metadata that exposes the library materials available to the people in your organization.