- From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
- Date: Mon, 18 Mar 2002 22:21:28 -0500
- To: jos.deroo.jd@belgium.agfa.com
- Cc: phayes@ai.uwf.edu, jonathan@openhealth.org, www-webont-wg@w3.org
From: "Jos De_Roo" <jos.deroo.jd@belgium.agfa.com> Subject: Re: Moving forward Date: Tue, 19 Mar 2002 00:17:25 +0100 > As an aside, I have some trouble that > :John a :Person . > :John a :Student . > > would entail > :John a [ owl:intersectionOf ( :Person :Student ) ] . > > but no trouble that it would entail > :John a ( owl:intersectionOf ( :Person :Student ) ) . > > using > ------------------------------------------------------------------------ > IF | THEN > ------------------------------------------------------------------------ > ?x a ?a | ?x a ( owl:intersectionOf ( ?a ) ) > ------------------------------------------------------------------------ > ?x a ?a | ?x a ( owl:intersectionOf ( ?a / ?b ) ) > ?x a ( owl:intersectionOf ?b ) | > ------------------------------------------------------------------------ > where the ( :a / :b ) is shorthand for [ owl:first :a; owl:rest :b ] > > but maybe I haven't thought hard enough??? > > -- > Jos The range of rdf:type is rdfs:Class, so either :John a ( owl:intersectionOf ( :Person :Student ) ) . is a contradiction (if lists and classes are disjoint), or ( owl:intersectionOf ( :Person :Student ) ) a rdfs:Class . which just gets us back to the first entailment. So I don't think that your distinction is defensible. Either you believe that these extra classes should exist or you believe that they shouldn't. About the only viable intermediate point would be to have another category that would include restrictions and boolean class combinations and have these exist but not have other kinds of classes exist. peter
Received on Monday, 18 March 2002 22:22:05 UTC