- From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Date: Thu, 01 Apr 2004 11:01:25 +0100
- To: Nick Gibbins <nmg@ecs.soton.ac.uk>
- Cc: SWBPD list <public-swbp-wg@w3.org>
Nick Gibbins wrote: > Jeremy Carroll wrote: > > >>>I'm not sure if this counts as best practice, but we use a simulated >>>stochastic process to determine which Semantic Web language is being used >>>by a particular file. You can see a demo of our technique at: >>> >>>http://triplestore.aktors.org/demo/OwlBall/ > > >>I think that qualifies under the "sorting out the mess that we've made" >> >>Isn't >>http://www.w3.org/2002/07/owl >>intended as an RDFS file, not OWL Full? >>(The semantics differ ...) > > > I think that this rather proves Ian's point - I'm using Sean's species > validator behind the scenes (and a few simple and no doubt defeasible > heuristics), and it reports it as OWL Full. Maybe the question should > instead have been "what is the best practice for *indicating* which > semantic-web language and encoding is being used"? > Named graphs could do that ... but that's still a research topic. (Discussed on SWIG) If we have a URI for the graph, then we can make statements about it, such as its intended assertional status, its intended semantic theory etc. Jeremy > >>It made me laugh though. > > > So justifying the (minimal) effort...
Received on Thursday, 1 April 2004 05:04:19 UTC