- From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
- Date: Wed, 28 Aug 2002 10:48:12 -0400 (EDT)
- To: phayes@ai.uwf.edu
- Cc: www-webont-wg@w3.org
As an indication of how things can go wrong in a complicated specification, Pat's document is very close to implying that all RDFS classes and properties are OWL objects. Why is this? First, the extension of the denotations of rdfs:subClassOf and rdfs:subPropertyOf are transitive relations according to the RDFS model theory. Second, a resource belongs to the class extension of the denotation of owl:TransitiveProperty if its extension is transitive. Therefore, the denotations of rdfs:subClassOf and rdfs:subPropertyOf belong to the class extension of the denotation of owl:TransitiveProperty. Third, owl:TransitiveProperty is an rdfs:subPropertyOf owl:ObjectProperty and thus of owl:Property. Therefore, the denotations of rdfs:subClassOf and rdfs:subPropertyOf belong to the class extension of the denotation of owl:Property. From this, one might expect that a domain of rdfs:subClassOf and rdfs:subPropertyOf should be owl:Object. However, this is written as a closure rule, not as a semantic condition, so it is not triggered. If it was a semantic condition it would imply that all RDF classes and properties belong to the class extension of the denotation of owl:Object. peter
Received on Wednesday, 28 August 2002 10:48:24 UTC