SDShare 1.0 final draft

For a few years now we have been using an approach based on Atom feeds to extract data from legacy sources and move them into RDF triple stores. The same approach also lets us pull data from triple stores for conversion back to legacy stores, indexing in search engines, and more.

We've now finally written it up as a proper specification so that other people can use the same approach. We have a few tools, open source and commercial, but most of them need to be updated to the final draft of the spec (published just now).

We think this is a mechanism that is both very powerful and really simple. For example, to extract data from a legacy database, all you need to do is expose a few Atom feeds, and some RDF views of the data. A general SDShare client[1] can then do the work of actually keeping the RDF store in sync. Obviously there can be multiple sources.

We looked at the Linked Data Platform use cases, and use cases 1.1.1, 1.1.2, 1.1.3, 1.1.5, and 1.1.8 (at least) look to us as though they can be met by SDShare. In fact, we've spent the last few years implementing systems doing that for commercial clients.

So we think this spec is very relevant for a large number of real-world use cases. 

We'd very much welcome any and all comments and reactions to the spec, which you can find at:
  http://www.sdshare.org/spec/sdshare-current.html




[1] http://code.google.com/p/sdshare-client/
[2] http://www.w3.org/2012/ldp/wiki/Use_Cases_And_Requirements

--Lars M.
http://www.garshol.priv.no/tmphoto/
http://www.garshol.priv.no/blog/

Received on Tuesday, 10 July 2012 12:39:09 UTC