On Wed, 17 Nov 2004 07:38:21 +0900, Martin Duerst <[EMAIL PROTECTED]> wrote: > At 23:32 04/11/16, Henry Story wrote: > > > 4- Have a very simple example that shows where one needs to place the > > rdf:parseType="Resource" attributes to turn the document into RDF -- > > if this is needed at all. > > In the simpler cases (and I think this may apply to Atom), there > are even well-known formal ways to describe this. It's just default > attributes, so any DTD can do it. There would be a standard DTD > for Atom (without these attributes), and an RDF-oriented (with > these attributes). With that RDF-enhanced DTD, it might go straight > into an RDF parser.
If I remember correctly, the use of DTDs in this way have passed through the list before (pre-IETF). I think the idea was discounted at the time for one or two practical reasons (sorry, I forget why, and Googling didn't help), but given the present circumstances it may well be worth revisiting. If a DTD could be used to provide a transparent overlay in this way without breaking anything else, that'd be great. Henry's approach of having an implicit mapping to RDF that can be made explicit through the insertion of one or two other-namespace attributes really does sound a low-cost, potentially uncontroversial way of making an RDF interpretation available. I think this may also offer a good reference for structural interpretation of extension elements too (as in Quark, Strangeness and Charm). Cheers, Danny. -- http://dannyayers.com
