- From: pat hayes <phayes@ai.uwf.edu>
- Date: Thu, 31 Oct 2002 10:14:16 -0600
- To: "Jos De_Roo" <jos.deroo.jd@belgium.agfa.com>
- Cc: w3c-rdfcore-wg@w3.org
>[...] > >> I actually said in the MT document that those rules weren't intended >> to define a process or to be directly implemented. I wish Id never >> mentioned closure rules: it was only intended to be a way of relating >> the semantics together. > >right, I understand that, but still, inference rules can be inferenced >why then not use e.g. owl properties for some of the rdfs-rules >e.g. >rdf:first a owl:FunctionalProperty. # and just have rdf:List >rdfs:subClassOf a owl:TransitiveProperty. # instead of rule5 >rdfs:subPropertyOf a owl:TransitiveProperty. # instead of rule3 Sure, I have no problem with that. However it just puts the real issue under the rug, since those owl properties now have to be reasoned about, and the question re-appears as how best to think about owl:TransitiveProperty and so on. It you follow Ian's translation of OWL (DLs generally) back into FOPC, the 'rules' just re-appear in any case; TransitiveProperty becomes the implication you would expect, and so on. You know, all we are all doing is taking little pieces of logic and inventing about 7 different notations for them and then writing long papers about how to translate between the notations we have invented :-) Pat > >I really can't see any problem with such an *implementation* >at least not as far as I did tests with that (for all our testcases) >(in a similar way, owl-rules could make use of e.g. math properties) > >-- , >Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/ -- --------------------------------------------------------------------- IHMC (850)434 8903 home 40 South Alcaniz St. (850)202 4416 office Pensacola (850)202 4440 fax FL 32501 (850)291 0667 cell phayes@ai.uwf.edu http://www.coginst.uwf.edu/~phayes s.pam@ai.uwf.edu for spam
Received on Thursday, 31 October 2002 11:14:21 UTC