- From: Jos De_Roo <jos.deroo@agfa.com>
- Date: Sat, 28 Jun 2003 10:56:22 +0200
- To: pfps@research.bell-labs.com
- Cc: jjc@hpl.hp.com, www-webont-wg@w3.org, www-webont-wg-request@w3.org
>Jeremy wrote: >> As a straw man, how effective would the following change to S&AS be: >> >> rewrite 5.3 so that: >> >> OWL Full neither affirms or denies that owl:Thing = rdfs:Resource, >> owl:ObjectProperty = rdf:Property owl:Class = rdfs:Class (where = is >> equality of class extension). >> >> Specifically replace: >> >> [[ >> OWL Full augments the common conditions with conditions that force the parts >> of the OWL universe to be the same as their analogues in RDF. These new >> conditions strongly interact with the common conditions. For example, >> because in OWL Full IOT is the entire RDF domain of discourse, the second >> comprehension condition for lists generates lists of any kind, including >> lists of lists. >> >> Definition: A OWL Full interpretation of a vocabulary V is an OWL >> interpretation as above that satisfies the following conditions. >> >> IOT = RI >> IOOP = PI >> IOC = CI >> ]] >> >> by >> [[ >> Definition: A OWL Full interpretation of a vocabulary V is an OWL >> interpretation as above >> ]] >> >> This seems to fix the specific counter-examples. Jeremy, how does this fix http://lists.w3.org/Archives/Public/www-webont-wg/2003Jun/0295 I can see the fix of the other ones http://lists.w3.org/Archives/Public/www-webont-wg/2003Jun/0257 http://lists.w3.org/Archives/Public/www-webont-wg/2003Jun/0031 >> Moreover it makes OWL Full and OWL DL (as in 5.4) closer; and the >> correspondence theorem is already stated as an IFF between section 5.4 model >> theory and section 3 model theory; which makes an IFF between the modified >> 5.3 and section 3 seem more plausible. >> >> Would that work? Peter wrote: >Sure, it would work. Fine; it is indeed good to keep K OWL Full entails C if and only if K OWL DL entails C certainly for OWL Full reasoners thinking DL. >> How much baby goes out with the bath water? > >However, you would then need to assert, for example, that rdfs:Class was an >owl:Class before you would get many OWL inferences from it. In essence, >you would have to do all the same typing you need to do in OWL DL. At least I haven't found such OWL inferences in the test cases that euler runs (I mean, when we drop owl:Class owl:equivalentClass rdfs:Class. owl:Thing owl:equivalentClass rdfs:Resource. we still run all the testcases that we ran before). Peter, can you give such a test case? -- Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
Received on Saturday, 28 June 2003 04:56:51 UTC