On day 2 by far the most striking paper for me was by Michael Durbin [PDF]: Terminology Services in a Digital Repository. He described a set of tools for creating controlled vocabularies and making them available to services such as metadata forms. This is an important area or repositories – if we can share terminology using a linked-data approach then we can improve discoverability. There’s a page about the service at OCLC’s site.
At the moment in the IR world we have a variety of ways to refer to the same thing so you get “Article, Journal” and “Article” and “Journal Article”, and so on, all referring to (more or less) the same thing. Using a terminology server, we could all use the term from the bibliographic ontology with the URI: http://purl.org/ontology/bibo/Article. That doesn’t mean the users have to see that URI, you can still choose what you want to show on a web page as part of a description, but when services such as Trove or ARO aggregate repository content we can be much more confident that the right kinds of resources are being grouped together.
What the users depositing content will see need not change much, they will still be presented with drop-down lists and auto-complete fields, but behind the scenes repositories will not be storing plain-old text strings as metadata, but a combination of an identifier (a URL) and (maybe) a text string.
There are a similar efforts elsewhere. Chatting to Neil Jefferies of Oxford, during the Duraspace-hosted drinks on the third day I learned that Oxford has a similar service under development. In my last post I mentioned BibApp as another important service which can do a similar thing for people’s names. And I know that ANDS have been exploring the idea of an Australian vocabulary service.
In the ADFI Software R&D team’s forthcoming work for ANDS under the metadata stores programme we will be making a service which will work for both names and terms. I hope that as this new class of software becomes more common and the various efforts mature we can start to work out some standards for interoperability.
Copyright Peter Sefton, 2010. Licensed under Creative Commons Attribution-Share Alike 2.5 Australia. <http://creativecommons.org/licenses/by-sa/2.5/au/>
