Re: [call for comments] voiD - vocabulary of interlinked datasets

Hello!


> The examples look like the typical result of a SPARQL CONSTRUCT query
> against the endpoint. Would it be easier to give a URI which when used
> in a construct query against the declared void endpoint, would return
> a typical example of what can be expected? This would save people
> placing an example document somewhere and keeping its structure
> current with what the SPARQL endpoint generates if something changes.

Yes, of course - in this example this is actually the result of a
DESCRIBE query. I could have made the void:example property directly
point to the Jamendo end-point, eg.
http://dbtune.org/jamendo/sparql/?query=describe%20%3Chttp://dbtune.org/jamendo/record/33%3E

>
> If they can all be collected via a robots.txt->sitemap.xml->void
> references->SPARQL CONSTRUCT method then people won't need to submit
> or preproduce them either. The option to produce the example and place
> it somewhere would be nice though as it would suit sindice who don't
> ever want to actually call sparql endpoints (if they can help it).
>
> Your example query gives the existence of a dataset relevant to your
> query but omits to tell you what the set is. A construct or a select
> query may be more relevant to you for doing what you want here.
>

Yes, it was just an example - replacing the ASK by a SELECT ?ds
?sparql would make sense, here

Btw, I posted a follow-up on that at
http://blog.dbtune.org/post/2008/06/12/Describing-the-content-of-RDF-datasets
At the bottom of the post, there's a link to the discussion that
followed on the #swig IRC channel.

Cheers!
y