- From: Jan Grant <Jan.Grant@bristol.ac.uk>
- Date: Fri, 10 Jan 2003 16:58:12 +0000 (GMT)
- To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
- cc: www-rdf-comments@w3.org
On Fri, 27 Dec 2002, Peter F. Patel-Schneider wrote: > Integrated Review of the RDF Core WG LCC Documents (as of 26 December 2002) > RDF Test Cases > W3C Working Draft 18 December 2002 > > > > The test cases RDF encoding depends on an external-to-RDF correlation > between URI references and documents. It would be much better to create > resources for these documents, with a property whose (literal) value is the > URI of the document. Thanks for this, Peter. Back when I first sketched out the manifest format I was of the same opinion: [[ Base URI and relative URI as literal properties <test:RDFDocument test:baseURI="absolute-address-of-test-document" /> This is arguably preferable to the [current situation], since it treats the address of a document as being explicitly differentiated from a URI which may denote it. This is particularly important since the lexical content of the base URI will be of interest to a parser. The downside of this choice is that it may prove confusing (even contentious) to the RDF community. However, the discussion of the distinction (if any) between denotation and dereferencing of a URL has to happen sooner or later. ]] In the end, I was persuaded that the (tacit) denotation of a document by a resource node labelled with a URL that, when dereferenced, produces (a representation of) that document, is a common assumption made by users of RDF (DanC primarily persuaded me of this - I recall his comment was that the alternative "smells bad" :-) ). I've not pointed out this relationship explicitly in the test case document. I've heard people (DanC in particular) repeatedly state that, if it looks like a URL (as opposed to a URIref), you ought to be able to get something by going there with a browser. My personal opinion is that W3's TAG ought to say something on this matter, because it's an important one that ought to be made explicit. As it is, sneaking it in in a test case document doesn't do the matter justice - I'd rather see this dealt with elsewhere, since the relationship between denotation and documents available by dereferencing would appear to be an important architectural issue. Cheers, jan -- jan grant, ILRT, University of Bristol. http://www.ilrt.bris.ac.uk/ Tel +44(0)117 9287088 Fax +44 (0)117 9287112 http://ioctl.org/jan/ stty intr ^m
Received on Friday, 10 January 2003 12:02:14 UTC