- From: Birte Glimm <birte.glimm@uni-ulm.de>
- Date: Tue, 4 Oct 2011 14:39:25 +0200
- To: SPARQL Working Group <public-rdf-dawg@w3.org>
Hi all, I am working through Michael Schneider's comments and I see two possibilities to address http://lists.w3.org/Archives/Public/public-rdf-dawg-comments/2011Jul/0013.html The problem he identifies is that we call the D-entailment regime D-entailment, but unlike the D-entailment specification in the RDF model theoretic semantics document, we fix a datatype map. Thus, we kind of change D-entailment, which is not our business. There seem to be two solutions for this: 1) Change the name of the regime; Michael suggests XSD-entailment, but since we also have rdf:PlainLiteral or rdf:XMLLiteral, I don't find that very fitting, so we would have to find a better name. 2) We do not specify any concrete datatype map. In this case, implementors would have to specify, e.g., in their system documentation, what kind of datatypes they support. In this case, we also have to think about canonicalisation. For all datatypes in the regimes there is a wel-defined canonical form, which has to be used. Otherwise, we again have the problem of infinite answers due to infinitely many lexical forms for some data values. If we leave the choice of the datatype map open, we might have to require that either each datatype is such that a canonical form for each data value is specified or, for datatypes where this is not the case, implementors have to make sure that only finitely many answers are returned and describe, e.g., in their system documentation, how that is achieved. I tend to prefer option 1 as this is also less likely to require another LC. Birte -- Jun. Prof. Dr. Birte Glimm Tel.: +49 731 50 24125 Inst. of Artificial Intelligence Secr: +49 731 50 24258 University of Ulm Fax: +49 731 50 24188 D-89069 Ulm birte.glimm@uni-ulm.de Germany
Received on Tuesday, 4 October 2011 12:39:56 UTC