- From: Patrick Stickler <patrick.stickler@nokia.com>
- Date: Thu, 31 Oct 2002 09:37:50 +0200
- To: "Jeremy Carroll" <jjc@hpl.hp.com>, "ext pat hayes" <phayes@ai.uwf.edu>
- Cc: <w3c-rdfcore-wg@w3.org>
[Patrick Stickler, Nokia/Finland, (+358 40) 801 9690, patrick.stickler@nokia.com] ----- Original Message ----- From: "ext pat hayes" <phayes@ai.uwf.edu> To: "Jeremy Carroll" <jjc@hpl.hp.com> Cc: <w3c-rdfcore-wg@w3.org> Sent: 30 October, 2002 19:50 Subject: RE: rdfs:StringLiteral > > >thanks Patrick I misspoke. > > > >Should all datatypes be subclass of rdfs:Literal; or maybe only the datatype > >xsd:string should be a subclass of rdfs:Literal > > My understanding is that rdfs:Literal is the class of all possible > literal values. "literal values" sounds like an oxymoron to me. We have either lexical forms or values, and if the L2V mapping happens to map a lexical form to a value that is string equal to the lexical form,, fine, but one can still argue that both exist insofar as the semantics are concerned. Do you mean "lexical forms"? If so, then while we can say that rdfs:Literal is the class of all possible lexical forms, I don't see how that fits into RDFS Class semantics, since that is a syntactic based membership and not a semantic one. I personally consider rdfs:Literal to be a bug. Its membership is defined by M&S as a set of structural components of the graph, not the denotation of members of a class. I.e. it's a syntactic class according to the abstract syntax, not a semantic class according to the denotation of the literals. We can perhaps salvage the term by using it to reflect the semantics proposed by Jeremy for rdfs:StringLiteral, but I don't see how it is possible (or rather reasonable) to have any relation whatsoever between arbitrary datatype classes and rdfs:Literal. Patrick
Received on Thursday, 31 October 2002 02:37:52 UTC