From SPARQL Working Group
(In response to NH-1)
> Is there a way of describing the supported results formats in SPARQL 1.1 > Service Description, as SADDLE did? > > saddle:resultFormat [ saddl:mediaType "application/sparql-results+xml"; > saddl:spec <http://www.w3.org/TR/rdf-sparql-XMLres/> ], > [ saddle:mediaType "application/rdf+xml"; > saddle:spec <http://www.w3.org/TR/rdf-syntax/> ].
Thanks for the comments. Support for this feature will be incorporated into the next draft of the service description document. The format will be something like this:
<service> sd:resultFormat <http://www.w3.org/ns/formats/Turtle> .
The /ns/formats/ namespace goes on to define things such as media types and filename extensions.
> Another useful property that it would be useful to be able to describe are the syntaxes that an endpoint is able to parse when submitting data.
This hasn't been considered yet, but I'd be happy to bring it up for discussion within the working group.
> Finally there might be other statistical information that it would be interesting to provide: > - Number of named graphs
If the service description provides descriptions of the graphs, you should be able to count them directly.
> - Number of distinct subjects/predicates/objects/resources
This is explicitly left to other vocabularies (e.g. voiD/SCOVO).
> Some statistics relating to any type of server (these might be best served in another vocal): > - Server uptime > - Number of HTTP requests > - Number of successful Imports > - Number of SPARQL queries
We have not discussed expressing statistics such as these, and I believe they are beyond the scope of the current service description work.
Gregory Williams, on behalf of the SPARQL working group.