finally a bnode with a uri

OpenSocial in RDF

I've created an RDF converter for the OpenSocial field definitions.
I'm currently working on a new release of Knowee. This is another (long-promised) item on my ToDo list before I can finally concentrate on paggr (although it took too long already and hopefully won't break my neck. All the planned paid projects for bootstrapping paggr didn't happen, due to frozen budgets and politics. I hope the situation here improves soon.)

So, while I was trawling the vocabulary market, trying to gather terms for the stuff that Knowee works with (people, their profiles, contacts, accounts, and activities), I remembered OpenSocial, the effort to standardize basic interactions between social networking sites. I can use a good amount of FOAF, but OpenSocial has very handy things such as a generic "tags" field and a clean vCard mapping. And it's a super-set of Portable Contacts, too.

Today, I wrote a converter that extracts the field definitions from the JavaScript specification files, together with their labels, comments, domains, and value types. (A little too late, I found out that Dan Brickley had already done part of this a couple of months ago, could have saved me some work, d'oh.)

I've just added the osoc spec to web-semantics.org/ns. I hope it might be of use to others as well. Funnily, the "relationship" term was not part of any of the source files, maybe I still have to invent a property (a foaf:knows equivalent that also works with organizations).

Comments and Trackbacks

I don't think the conversion from the javascript files is a robust approach to define the OpenSocial Ontology. Enumeration, String arrays needs to be formalized in a better way. Also there are conceptual modeling of Organization is problematic because it mixed the organization description with the job position time frame. This should be separated to favor reusability. Currently I am not aware of any effort trying to do this manually. I can offer my help to craft this manual if needed.
Comment by Stephane Fellah on 2008-12-17 02:27:23 UTC
Stephane, you're of course right. This is less an ontology than a term set. But it may be a good first step. I moved the ENUM types into rdfs:comment for now, but could generate owl:oneOf ranges from them. The string array stuff is more tricky, at least when it comes to ordering. The converter avoids any special interpretation/changes and the vocab does the job (for me) so far, but if you'd like to contribute a cleaned-up version, even better.
Comment by Benjamin Nowack on 2008-12-17 11:27:41 UTC
0 comments are currently in the approval queue.

Comments are disabled for this post.

Earlier Posts

Archives/Search

YYYY or YYYY/MM
No Posts found

Feeds