- From: Dan Brickley <Daniel.Brickley@Bristol.ac.uk>
- Date: 1998年12月21日 20:20:45 +0000 (GMT)
- To: www-rdf-comments@w3.org
- Message-ID: <Pine.GHP.4.02A.9812212011040.14803-100000@mail.ilrt.bris.ac.uk>
Hi The RDF M&S spec defines triples as triple(res,property,value) where value may be a URI reference or a literal, and res and property are URI references. The output from SiRPAC doesn't always produce URIs in the appropriate places: for example, generated node identifiers (genid_6 etc) for "anonymous" nodes. Also, relative URIs (and namespace refs, though I'm unsure about legality of trying that!) and resources identified only by ID within the RDF/XML do not produce triples that contain URIs. I'm not sure what's best to do here. Seems a bit over-the-top to create uuid: URIs for each anonymous node. The URI specification[1] defines four rules for derriving absolute from relative URIs + context, and includes warnings about 'meaningless' relative URIs. It includes provision for 'application specific' rules for creating full URIs as a last resort. This might be worth considering in cases where the RDF/XML doc being parsed has no obvious URI. Perhaps an additional parameter for SiRPAC might be 'baseURI', so that the calling context could pass in a URI against which relative URIs should be considered? cheers, Dan [1] http://www.ics.uci.edu/pub/ietf/uri/rfc2396.txt -- Daniel.Brickley@bristol.ac.uk Institute for Learning and Research Technology http://www.ilrt.bris.ac.uk/ University of Bristol, Bristol BS8 1TN, UK. phone:+44(0)117-9288478
Received on Monday, 21 December 1998 15:20:49 UTC