- From: Frank Manola <fmanola@mitre.org>
- Date: Tue, 04 Mar 2003 08:45:05 -0500
- To: pat hayes <phayes@ai.uwf.edu>
- CC: w3c-rdfcore-wg@w3.org
pat hayes wrote: > >Frank Manola wrote: snip >> Note that we're talking about *syntax* here (this is the abstract >> syntax section), rather than differences about what RDF chooses to use >> these things for (which seems to me a separate issue). So are RDF URI >> references *syntactically* URI references or not? (I understand that >> not all URI references, e.g. relative URIs, are legit as RDF URI >> references > > > Hmm. I presume there was a good reason for that restriction. Seems on > balance that there is no obvious reason why RDF should exclude any class > of identifier. Logically, the 'RDF URIrefs' could be any set of strings > which are distinguishable from literals and bnodes. I cannot see any > good pragmatic or semantic reason to exclude any URIrefs from RDF. > > Has this got something to do with bloody XML ? > I believe we don't have relative URIs in the abstract syntax because, in the abstract, we don't define anything for them to be "relative" to (like a "graph id" or something), so there's no way to "complete" them. URIrefs in the abstract model (in triples) have to be absolute. In the XML, we *can* have relative URIrefs (like those names for schema classes like #Person we use all the time), because there we XML mechanisms to turn them into absolute URIrefs. --Frank -- Frank Manola The MITRE Corporation 202 Burlington Road, MS A345 Bedford, MA 01730-1420 mailto:fmanola@mitre.org voice: 781-271-8147 FAX: 781-271-875
Received on Tuesday, 4 March 2003 08:25:05 UTC