Hi Peter! Peter F. Patel-Schneider wrote: >> >5/ The RDFS-compatible semantics for OWL 1.0 would be incorrect if >> > owl:FunctionalProperty rdfs:subClassOf owl:ObjectProperty >> >> Why is this an error? All instances of the class >'owl:FunctionalProperty' >> will certainly be instances of the class rdf:Property, right? And in >OWL >> Full, the classes 'rdf:Property' and 'owl:ObjectProperty' have >identical >> class extensions, according to sec. 5.3 of the AS&S. So the above >axiomatic >> triple is equivalent to >> >> owl:FunctionalProperty rdfs:subClassOf rdf:Property > >Yes, in OWL 1 Full, but not in the RDFS-compatible semantics for OWL 1 >DL. > >In OWL 1 DL there can be functional data properties, which are not >object properties. Ok, this clarifies things. Thanks! The WG will have to decide whether we want to have such an "RDF style" OWL-DL clone or not in OWL 2. I cannot see a demand in the charter for such a language. And I do not remember that there was any advocacy in the WG or by someone else in favor for it (although I remember that there was once a mail from a HP employee bringing this language into play). And I, personally, do not see any value in such a language, but see a lot of disadvantages instead. Anyway, I will raise an issue. And, independently on the outcome, I am going to change the axiomatic triples to applying the RDFS vocabulary. Cheers, Michael
This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 21 April 2008 08:46:12 GMT