- From: Jonathan Borden <jonathan@openhealth.org>
- Date: Fri, 5 Apr 2002 15:17:41 -0500
- To: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>, "Jeff Heflin" <heflin@cse.lehigh.edu>, "Ian Horrocks" <horrocks@cs.man.ac.uk>
- Cc: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>, <www-webont-wg@w3.org>
Jeremy Carroll wrote: > > I think this indicates an interesting technique for resolving whether the > lite syntax is either: > - transformed into triples by already being in RDF/XML (Jonathan's > preference?) > or: > - in an XML concrete syntax corresponding closely to the abstract syntax > (Jeff's preference?), and then (my preference) transformed into RDF/XML by > XSLT, and then into triples by RDF/XML. No, really my position is that I would like to see an actual XML grammar proposal i.e. a DTD, a RELAXNG schema, an XML Schema so that this can be properly evaluated. I am not for making tiny little changes to what might be expressed in RDF/XML. That said, unless something very close to "daml:collection" is allowed by RDF, then I wouldn't favor an RDF/XML syntax. XML has the clear advantage over RDF precisely given the fact that child elements are true _sequences_ i.e. ordered and finite. > > The technique is: > - the initial (internal) draft is for an XML concrete syntax (uninfluenced > by RDF). > - advocates of RDF/XML as the syntax demonstrate how with little change it > can be made conformant RDF/XML > - we also explore how to transform the initial draft XML concrete syntax > into an adequate set of triples. > > I confess to being nervous about the pure RDF/XML option, because of the > round tripping problem. A possible solution is to follow Jonathan's > suggestion but also (silently) requiring the reification of all the triples > in the file and sticking them in a bag. This allows the preservation of > (XML) document order. (It's a bit of a hack). > I've also demonstrated an actual concrete non-XML syntax which closely follows the abstract syntax http://www.openhealth.org/WOWG/OWLNonXMLsyntax.text but haven't received much comment on this. I also think there is a real viable option of treating the RDF/XML syntax _as XML_ in that it can be edited and roundtripped in its XML form, and need not always be parsed into triples ... i.e. RDF/XML is actually XML and can be used as XML. The advantage of this is that folks who want to use XML can use XML and folks that want RDF can use RDF. Again this is all predicated on things like daml:collection (the unexpanded form) etc. which would make RDF usable. Without something like this, I would favor another XML syntax. Jonathan
Received on Friday, 5 April 2002 15:22:34 UTC