- From: Gerardo Horvilleur <mago@mail.internet.com.mx>
- Date: Wed, 25 Apr 2001 10:07:41 -0500
- To: www-rdf-logic@w3.org
> > > (b) If P is applied to a resource r, r will belong to C > Which interpretation of range constraints do you prefer? > > [ P, rdfs:range, C] > is intended to express: > > (a) P may be applied to any resource that belongs to class C > > or > > (b) If P is applied to a resource r, r will belong to C What happens if you also have [P, rdfs:range, C2]? With the current "union" semantics then (b) would have to be restated as: (b1) if P is applied to a resource r, r might belong to C if we change to "intersection" semantics then (b) could be restated as: (b2) if P is applied to a resource r, r will belong to C and C2 In my opinion the correct interpretation is (a). Depending on the chosen semantics an application/inference engine could have (b1) or (b2) stated as a rule. Using just plain (b) by itself might be wrong because it only considers the special case where P has a single range. On the other hand: a property with multiple ranges seems a little weird to me. Does anyone have a good example of why it might be useful? -- Gerardo Horvilleur mago@mail.internet.com.mx .
Received on Wednesday, 25 April 2001 12:06:23 UTC