Monday, March 29, 2010

Slippery slope for linked data

I've been wondering for a while how to create a platform that used linked data natively.

There are plenty of platform building options which understand ordinary relational databases, but, for my requirements, there's nothing like Yii with a linked-data backend.

Anyway, I'd really like to see an OpenURL resolver re-use the data by making the requests it's given available as linked open data, in much the same way as SFX exposes it's data for other purposes, such as the bX service.

I also hope I can see a time when publishers, or a body representing them (CrossRef)will make the metadata for the articles they produce freely available, like Nature Publishing... hell, a lot of them expose their content to Google, so why not maximise their exposure via other means?

1 comment:

  1. I'm not sure if it is what you are looking for, but you might want to look at the Talis 'Platform' - I'm pretty sure they would be interested in work in this area. I know staff at Talis have done some work in getting a triple store to update from a stream of data (e.g. a twitter stream), so I'd have thought this idea could be transferred to pushing triples from a link resolver - if you can decide on the data model.

    I couldn't agree more in terms of publishers making metadata freely available - I really can't understand why this isn't already done as standard. In the meantime we see all this negotiation going on for aggregated search applications (e.g. Summon, Primo Central, Ebsco Resource Discovery etc.) - it would be much better and healthier to see an open market place on this IMO.