- From: Christopher B Ferris <chrisfer@us.ibm.com>
- Date: Fri, 26 Jul 2002 16:04:18 -0400
- To: www-tag@w3.org
Mark Baker wrote: > On Fri, Jul 26, 2002 at 09:48:00AM -0700, Joshua Allen wrote: > > Can you please elaborate on how you believe RDF to be related to > > *accessing* a resource? > Well, RDF is what you get when you access a resource. No, it is not. It might be, but only if the origin server decides that "RDF" is the representational form that will be returned for the resource identified. Of course, there are a number of *different* representations of RDF (RDF/XML, N3, NTriples). Given that there appear to be no registered media types (and their formal specification) for any of these, it isn't at all clear to me that the fragment identifier has any meaning at all, given that its interpretation is based on the rules defined for the media type of the representation retrieved (RFC2396). If the "RDF" is returned as text/plain, then the fragment identifier is meaningless. If it is returned as 'application/xml', it is equally meaningless since RFC3023 doesn't specify any fragment id mechanism, but defers to the WD of XPointer which doesn't seem to have any rules for naked fragment identifiers. These seem only to have conventional interpretation as meaning, "some element that has an attribute that is of type ID that has the same value as the fragment identifier". Of course, RDF doesn't have a DTD, and hence it isn't at all clear to me that even if expressed as: http://mumble/#xpointer(id("foo")) or http://mumble/#id("foo") That it would have any meaning that an XML parser could infer on its own because absent any DTD processing, there's nothing to tell the parser that rdf:ID is supposed to be an XML ID. http://mumble/#xpointer(/decendent::*/[@rdf:ID="foo"]) might be meaningful if the namespace qualifier of the RDF namespace happened to be 'rdf', so in truth, the only possible meaningful fragment identifier would be something like: http://mumble/#xmlns(rdf=http://www.w3.org/1999/02/22-rdf-syntax-ns#) xpointer(/decendent::*/[@rdf:ID="foo"]) And then, there's the issue of getting the same fragment identifier rules to apply to all possible representations of the identified resource, which would seem to exclude using XPointer since that's XML specific and RDF is not exclusively represented as RDF/XML. But, I digress:) Cheers, Christopher Ferris Architect, Emerging e-business Industry Architecture email: chrisfer@us.ibm.com phone: +1 508 234 3624
Received on Friday, 26 July 2002 16:09:14 UTC