- From: Pat Hayes <phayes@ihmc.us>
- Date: Sat, 17 Nov 2012 01:17:08 -0800
- To: Richard Cyganiak <richard@cyganiak.de>
- Cc: Pierre-Antoine Champin <pierre-antoine.champin@liris.cnrs.fr>, RDF Working Group WG <public-rdf-wg@w3.org>
On Nov 16, 2012, at 6:35 AM, Richard Cyganiak wrote: > On 15 Nov 2012, at 22:20, Pat Hayes wrote: >> Also, nobody in 2004 objected to the current scheme with quite Richard's degree of persistence :-) > > I didn't even *object*, I just wanted to know why it was done that way ... > > Glad to hear that there would be an acceptable way of making it work. I'm still a bit worried about consequences for OWL. Peter might be the best person to comment on that particular issue. AFAIK the consequences should not be dire, but they might be enough to require an explanatory Note to be issued. We might also allow RDF to be weakened so as to permit consistent usage of ill-formed literals as being legal. The semantics of such a weakened RDF would treat ill-formed literals just like literals with an unknown datatype, as denoting something but nobody knows what. The entailment :a :p "bad"^^xsd:number. ==> :a :p _:x. is valid in both cases: in the strict RDF, because the antecedent is always false; in the weaker RDF, because the consequent is true when the antecedent is. So the actual inference rules would still work, but reasoners might act differently when they detect an overt contradiction. This would give enough slack to existing OWL/RDF reasoners to just carry on as they are now, with modifications only to the text of their documentation, maybe. Anyway, an idea for possible discussion. Pat > Best, > Richard > ------------------------------------------------------------ IHMC (850)434 8903 or (650)494 3973 40 South Alcaniz St. (850)202 4416 office Pensacola (850)202 4440 fax FL 32502 (850)291 0667 mobile phayesAT-SIGNihmc.us http://www.ihmc.us/users/phayes
Received on Saturday, 17 November 2012 09:17:45 UTC