- From: Jeremy Carroll <jjc@hpl.hp.com>
- Date: Wed, 27 Feb 2008 14:56:50 +0000
- To: Bijan Parsia <bparsia@cs.man.ac.uk>
- CC: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, public-owl-wg@w3.org
Bijan Parsia wrote: > On 27 Feb 2008, at 14:43, Jeremy Carroll wrote: > >> Inter-op would argue for an XSLT 1.0 transform that probably can be >> done, but it may be easier to have an XSLT 2.0 transform. > > Thinking about it, the trickiest bit will be dealing with URIs and > CURIEs. Having to write a relative URI resolver will be annoying (but > it's been done...there are RDF/XML parsers in XSLT) It is easiest if we can leave relative URIs as relative. If we have to do major amounts of URI processing then I would advocate XSLT2 which has some support. CURIEs could be an issue :( I think I did write some code for them, once upon a time; and presumably Fabien's RDFa code does something with them too. > >> I tend to agree with Bijan's judgement that in principle this is >> possible (although I would expect corner cases which don't work, e.g. >> a property http://example.org/000) > [snip] > > Well, this is a case where the RDF/XML can't rep it anyway, so that's > fine (or, rather, an instance of a known limitation of the target format). > Known limitations strike me as OK. We would be doing very well if the only limitations are imposed by RDF/XML. Jeremy
Received on Wednesday, 27 February 2008 14:57:22 UTC