- From: Jos De_Roo <jos.deroo@agfa.com>
- Date: Fri, 7 May 2004 20:29:59 +0200
- To: "Pat Hayes <phayes" <phayes@ihmc.us>
- Cc: RDF Data Access Working Group <public-rdf-dawg@w3.org>
BTw Pat, in one of your earlier mails http://lists.w3.org/Archives/Public/public-rdf-dawg/2004AprJun/0325.html you wrote ??x and ?y variables and I wondered wether you intented any difference between 2 kinds of variable ?? PatH wrote: >> OK Rob, sounds fine... >> (will sleep over it once more :)) >> sounds like a separation of concerns > > Right, exactly, and I think its a good separation. OK > Still, unlike Rob, I think that we could reasonably approach the > 'inference issue' by seeing if we can come up with some way for a > query to indicate to a server (sorry I keep using this term, I can't > think of a better one: the engine or entity to which the query is > directed and which checks it against a graph: do we have a name for > this?) WHICH graph it wants the query to be directed to. Since > 'graph' here can be a virtual graph defined partly by inference > closures (and may in fact be infinite in some cases, eg when thinking > seriously about datatyped literals), a simple idea like just having a > URI for each graph, or one graph per server, seems inadequate. So can > we think of a way for a query to indicate which conceptual/virtual > graph it is intended to be evaluated with respect to? For example > (complete straw man) maybe the query could refer to a source listing > a set of rules which the virtual graph is supposed to be closed > under, or some such. Or it could be as coarse as just specifying > plain/rdf/rdfs/datatyped as options, where each of these refers to a > semantic extension and the graph is supposed to be closed under all > inferences which are valid wrt that extension. Right and that's exactly what I have been experimenting with i.e. when writing a query like e.g. (...and using N3) (<a-graph-uri> rdfs:).da:from da:select {?A a s:Publication}. that rdfs: indicates that we have to perform rdfs namespace entailment and leaving it out of the da:from argument list doesn't do that (in our tests, we actually map that rdfs: to http://eulersharp.sourceforge.net/2003/03swap/rdfs-rules.n3 and have similar stuff for xsd: owl: etc... but for those we don't make any completeness nor decidability claims :)) -- Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
Received on Friday, 7 May 2004 14:30:47 UTC