In my capacity as Semantic Web person, I've been really impressed with some library projects recently:
OpenLibrary,
LIBRIS and the new
Library of Congress Authorities and Vocabularies service (I mentally think of it as just "id.loc.gov" or some variant thereof).
These, and other, projects are trying to pull libraries kicking and screaming out of the technological dark ages by ensuring that both books and the categorised metadata is available as
Linked Open Data on the web. These projects are using web-native architecture (HTTP, content negotiation, 'cool URIs' etc.) and Semantic Web standards to make library data useful.
OpenLibrary is a great example: it gives every book a proper URI, it's compiled by pulling in data from a large number of academic and research libraries around the world and merged somewhat intelligently together. But it's also editable in a wiki style. If you look at
my user page, you can see the edits I've made: I've gone in and added Table of Contents data for every book I read or borrow as part of my research.
The U.S. Library of Congress provide a managed hierarchy of subjects and categories which is used by many libraries both in the U.S. and elsewhere. For instance, the library in my university college uses LoC subject headings, and many OPAC (library catalogue) systems let you use the LoC subject headings to look books up by - and they are being used in OpenLibrary also.
LIBRIS is the catalogue system for the Kungl. Bibliotekt - the National Library of Sweden - Sweden's version of the British Library or the Library of Congress. LIBRIS have also adopted Linked Data standards and are making all the data from their catalogues available in machine-readable form with proper resourceful URIs. They are really providing a model for how libraries should have catalogues. Hopefully other libraries will follow soon.
All three of these services are showing what libraries can do on the web when they take the web seriously, rather than just seeing it as an front-end to their catalogue systems. The web means allowing a complete interconnectedness between data sources: just as webpages are linked together, we are seeing that Semantic Web technologies are allowing datasets to link together into one giant web of data.
How might we be able to use this at the Citizendium? Well, I think are the sort of projects we should be partnered up with. How we do this is something I think we should explore as a community. While Wikipedia is blazing it's own user-generated path, I think that it's in the spirit of our project to link in to these projects. At a very practical level, I think that OpenLibrary is something we could start linking to on Bibliography pages with relative ease. In our metadata template, we could also start specifying how clusters are related to Library of Congress Subject Headings using the URIs provided by id.loc.gov. I think that as a relative newcomer in this field, we can start linking in with other services like this. Of course, this is just for books. For other things we discuss, we should start pointing out to machine-readable data in the same way: MusicBrainz is a database of musical recordings that operates similarly, Geonames does so for places, the BBC puts all their programmes online in a linked-data manner. And our scientific workgroups could do similarly for any number of datasets: a large number of chemical, pharmaceutical and genetic databases are going online as linked data.
In short: have a look at
Linked Data.org. Who thinks it would be a good idea if we could think of ways to linking Citizendium into any of the databases on there? Feedback welcome.
