- From: Seaborne, Andy <andy.seaborne@hp.com>
- Date: Fri, 15 Oct 2004 09:29:12 +0100
- To: "Dan Connolly" <connolly@w3.org>, "Steve Harris" <S.W.Harris@ecs.soton.ac.uk>
- Cc: "RDF Data Access Working Group" <public-rdf-dawg@w3.org>
Dan, Tests dawg-source-simple-00{4,5} illustrate queries involving SOURCE over a graph with a single named container in a collection of named containers (CNC) - which is the whole graph / [test 4] - and query over a single RDF graph - no containers, just a plain RDF graph [test 5]. There are two possible ways to go on the case of: SOURCE ?src (?x ?y ?z) and no CNC, just a plain graph, no name, no concept of named container. 1/ It can match with ?src unbound, or it can fail. 2/ It can fail Both would be reasonable: the first is more liberal, but if you are sending a query expecting a CNC and there isn't one, then it's a bit like sending a query expecting FOAF query to graph with RSS data. Dan wrote: > I don't see how test dawg-source-simple-005 works. > simple-data-1.n3 isn't referenced in the input data, I'm not sure which version of the tests you were looking at. The original version of the tests that I checked in, had no data files assigned in the manifest except for test 5 (plain graph); tests 1-4 got their data from the FROM clause ; test 5 does not have a FROM clause and got its graph to query from the query context (here, the manifest). I used this to be able to differentiate between a query over a CNC and query where is just a graph, no a CNC. Steve changed my tests to make the page generation work by adding back in the data files for queries 1-4 in the manifest. This broke them and I have fixed up something for now. It can only be a temporary solution, because the test suite can't express the difference between querying a plain old graph and a CNC as currently setup. Andy -------- Original Message -------- > From: Dan Connolly <> > Date: 14 October 2004 18:53 > > On Thu, 2004-10-14 at 11:36, Steve Harris wrote: > > On Thu, Oct 14, 2004 at 04:04:16PM +0100, Andy Seaborne wrote: > > > Steve, > > > > > > I'm not sure what the process is for tests. Do I send them to you > > > for inclusion or do I put then in CVS and let you know? I checked > > > into to CVS the tests for: > > > > > > http://lists.w3.org/Archives/Public/public-rdf-dawg/2004JulSep/0541.html > > > > > > Feel free to "organise" them into wherever they should go. > > > > OK, I've built the HTML and commited. > > I don't see how test dawg-source-simple-005 works. > simple-data-1.n3 isn't referenced in the input data, > as far as I can tell. How does it become part of the solution? > The query engine doesn't have access to the test > manifest, does it? > > Have the definitions in the relevant part of the spec > been enhanced to show how SOURCE solutions are defined? > No, not yet. http://www.w3.org/2001/sw/DataAccess/rq23/#source > $Revision: 1.117 $ of $Date: 2004/10/13 11:32:06 $ > > I hope they get elaborated soon. > > -- > Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Friday, 15 October 2004 08:29:45 UTC