W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > July to September 2009

Re: Options on SPARQL endpoint descriptions...

From: Gregory Williams <greg@evilfunhouse.com>
Date: Mon, 17 Aug 2009 22:30:15 -0400
Cc: Lee Feigenbaum <lee@thefigtrees.net>
Message-Id: <3D1366B4-61C9-4493-964B-10D58895F825@evilfunhouse.com>
To: SPARQL Working Group <public-rdf-dawg@w3.org>
On Aug 12, 2009, at 10:24 PM, Lee Feigenbaum wrote:

> Greg Williams wrote:
>> On Tue, Aug 11, 2009 at 11:01:00PM +0100, Axel Polleres said:
>>> If we rule out Option 5 entirely, we have no defined way of  
>>> querying the service description directly over the very SPARQL  
>>> endpoint, do we want that? I feel somehow this could be useful,  
>>> but don't really have a strong opinion about it either. We could  
>>> recommend some variant of Option 5 on top of one of the others  
>>> informally and simply not enforce everybody to support it?
>> Doesn't option 7 (conneg) give you this for free? By using the  
>> service
>> URI in a FROM clause, the service description RDF could be pulled  
>> in by
>> the endpoint (assuming it didn't keep it loaded) and be queried.
>
> Only for stores that dereference HTTP URIs that are part of a  
> query's dataset. Some don't (or aren't always configured to),  
> instead treating graph URIs simply as identifiers within a quad store.

Is it unreasonable to hope that if we went with conneg, stores that  
*didn't* dereference HTTP URIs but *did* provide a service description  
might keep the service description in the quad store? Also, I support  
Steve's point that the service description could indicate whether (and  
how/where) a description could be queried.

.greg
Received on Tuesday, 18 August 2009 02:31:00 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:39 GMT