- From: Bijan Parsia <bparsia@cs.man.ac.uk>
- Date: Tue, 10 Mar 2009 15:31:30 +0000
- To: W3C OWL Working Group <public-owl-wg@w3.org>
On 10 Mar 2009, at 14:12, Peter F. Patel-Schneider wrote: > From: Sandro Hawke <sandro@w3.org> > Subject: Re: Response draft for Jan Wielemaker JR8-2/54 > Date: Tue, 10 Mar 2009 09:37:40 -0400 > >> >>> From: Ivan Herman <ivan@w3.org> >>> Subject: Response draft for Jan Wielemaker JR8-2/54 (was draft >>> responses for >>> LC comment FH3/29) >>> Date: Tue, 10 Mar 2009 11:19:19 +0100 >>> >>>> Based on the email discussion yesterday I have made a draft for a >>>> possible (separate) answer to Jan: >>>> >>>> http://www.w3.org/2007/OWL/wiki/LC_Responses/JR8-2 >>>> >>>> I hope this summarizes the discussion. >> >> Can we just remove the long paragraph which mentions TriX? I think I >> already made the case for why its argument is false, The argument is absolutely true, at least in the sense that all proposed (as in worked out in some detail) formats fall, to my knowledge, to this argument. >> but I'll repeat it >> if someone wants. > > Please do so. I don't remember anyone falsifying the claim that > triple > serialisations of OWL ontologies are not unfriendly to XML tools. Even remotely. You've claimed that there *is* a case. But nothing you've pointed to *demonstrates* the case. It's certainly impossible with RDF/XML, since you need an unbounded set of QNames (to represent properties). It seems unlikely, and very much a hard research question, whether you can do it otherwise. I'm particularly concerned that you've already admitted a verbosity tax. It doesn't take a *lot* of extra cruft to make an XML format rather hard to work with. Cheers, Bijan.
Received on Tuesday, 10 March 2009 15:27:53 UTC