A small number of folks, including our own Rob Styles, recently flagged up the work by Ed Summers  in producing:

an experimental service that makes the Library of Congress Subject Headings available as linked-data using the SKOS vocabulary.

The results of this work can be found at lcsh.info. At first look a deceptively simple site, with a wealth of information and relationships lurking beneath the surface.  Viewing the site in your web browser, although interesting, is not the point.  It is designed to be used by other applications using the subject headings an the relationships between them.  Take up the offer on the site of browsing some of the subjects using a linked data browser – I found Zitgist easy to use, follow the narrower, broader, and related terms, it is amazing where you end up.

Ed has used SKOS (Simple Knowledge Organisation Systems), a formal language built on RDF to describe the concepts in the LCSH.

Recently Alistair Miles from the University of Oxford, key developer of SKOS gave a presentation at the Library of Congress on SKOS in the context of Semantic Web Deployment.  From the slides it looks like one of those events that you wished you had been around to attend.

It is initiatives like this that are the early green shoots of the benefits of Semantic Web appearing in and around library data.   With a reliable unique URI for each of the concepts in LCSH which leads you to a SKOS encoded definition for that concept, which then includes URIs for both broader and narrower terms in the subject heading hierarchy, why would anyone bother encoding such information in to their own application? 

For the moment lcsh.info is an experimental site which hopes to inform developments inside the Library of Congress.  If LC do follow this lead and open up the LCSH as reliable Open Linked Data and applications start to use it, you will rapidly end up with a network of applications that are semantically linked together by the fact they share the same URIs to define the same concepts.

There is much more to the Semantic Web, but just extrapolate this little bit of interlinking across other authoritative data sets in the library world  such as authors, publishers, and the like, then on in to data sets in other domains that share things like geographical locations, movie databases, Wikipedia etc., and you will end up with something significantly more powerful than the Web we have produced so far.

Early days yet, but pioneers like Ed, Alistair, and a growing band of linked data enthusiasts many of whom were to be found at the Linked Data Workshop at WWW2008 co-chaired by our own Tom Heath, deserve a hat tip.