- From: Dave Beckett <dave.beckett@bristol.ac.uk>
- Date: Wed, 24 Nov 2004 14:20:12 +0000
- To: Dan Connolly <connolly@w3.org>, RDF Data Access Working Group <public-rdf-dawg@w3.org>
On Mon, 22 Nov 2004 08:52:11 -0600, Dan Connolly <connolly@w3.org> wrote: > Dave, I don't recall getting a reply to my questions > about your source proposal... > http://lists.w3.org/Archives/Public/public-rdf-dawg/2004OctDec/0256.html No, sorry, been very busy. > In particular, > > I think you're suggesting that there are 2 query results for queries > that use SOURCE and that implementations are free to return > either. Is that right? > > Or perhaps there are 2^N correct results, where N is the > number of SOURCE clauses. I had written two alternatives to allow rdf systems that supported tracking triple sources and those that didn't. Waiting to see if that was worth it. My preference is always to make it simpler, so choices to do that could be: - mandating support for SOURCE - making support for SOURCE optional. Any query with SOURCE must fail if not supported) (- removing SOURCE) > And now a question to the customers... > For the use cases where SOURCE is used, is that going to work? > i.e. when they send a SOURCE query, is it OK if they sometimes > get a source and sometimes don't? As an implementor, I'd prefer to get source info when it was available rather than utter failure. I expect there are some queries which make no sense if the source support isn't available, rather than when getting source information is seen as just extras. Dave
Received on Wednesday, 24 November 2004 14:21:59 UTC