Re: [VM] RDF vocabulary configuration management with named graphs

Miles, AJ (Alistair) wrote:

>  
>
>># default graph
>>    
>>
>owl:imports ...
>		
>
...I think this is along the right lines, but the cheapo
way to do it is in plain RDF/XML, one document per
snapshotted perspective on the vocabulary, and load them up
into a multi-graph SPARQL engine.

In the FOAF case, for example, you can kinda get to all
the old versions of the FOAF vocab description index.rdf
via a (human oriented) viewcvs interface,
http://rdfweb.org/viewcvs/viewcvs.cgi/xmlns.com/htdocs/foaf/0.1/index.rdf

There are 52 documents accessible through that, eg.
(sorry about the ugly URIs),

3 years ago:
http://rdfweb.org/viewcvs/viewcvs.cgi/*checkout*/xmlns.com/htdocs/foaf/0.1/index.rdf?rev=1.2&content-type=application/rdf+xml

6 weeks ago:
http://rdfweb.org/viewcvs/viewcvs.cgi/*checkout*/xmlns.com/htdocs/foaf/0.1/index.rdf?rev=1.49&content-type=application/rdf+xml

I don't know any handy RDF vocab that would serve as a table
of contents for these, eg. by talking about dates, and that they
were the HTTP GET results against http://xmlns.com/foaf/0.1/ with
an Accept: header of 'application/rdf+xml' at those dates. But if
there were such a summary doc, it'd be fairly easy to crawl them all into
a SPARQL store, and ask questions like:

"In Aug 2004, what did http://xmlns.com/foaf/0.1/ claim was the
rdfs:domain of foaf:homepage?"

or, more ambitiously, "When did FOAF stop having an rdfs:domain
of foaf:homepage that was something other than rdfs:Resource?"
(or variations on that theme...).

"When did foaf:name become marked as 'stable'?" (and where was this
announced?...
hmm we can't support that bit yet... maybe DCMI can do better with
their decision record?).

This would I think make a great student project...

Dan

Received on Tuesday, 5 July 2005 15:18:25 UTC