- From: Peter F. Patel-Schneider <pfps@research.bell-labs.com>
- Date: Wed, 10 Sep 2008 14:23:12 -0400 (EDT)
- To: sattler@cs.man.ac.uk
- Cc: public-owl-wg@w3.org
From: Uli Sattler <sattler@cs.man.ac.uk> Subject: Review of RDF mapping Date: Mon, 8 Sep 2008 19:12:31 +0100 [...] > - in Table 10, if we have > > _x owl:inverseOf *:y1, > _x owl:inverseOf *:y2 > > then OPE(_x:) is, say InverseOf( OPE(*:y2)), and thus I am afraid the > we will miss that :y1 and :y2 are equivalent?! Yes, this is a buglet in the reverse mapping that arises because owl:inverseOf can arise from both axioms (InverseProperties) and object-property expressions (InverseOf). The solution, which I have put in the document, is to have an additional side condition in Table 10 that OPE(_:x) is empty. This prevents the above valid pair of triples from not being recognized. > The same observation > can be made in Table 12 with ComplementOf: perhaps things like > CE(_:x) need to be sets, and you then use the other rules on all > elements in these sets? The problem only occurs with owl:inverseOf because it is used for two things. Other properties are not overloaded in this fashion and don't exhibit this problem. > Or is this related to a confusion mentioned > above of "The rules from the following sections are not allowed to > redefine the value of any of these functions for some x." ? Not really, this wording (which I have modified to make clearer) is what make the two example triples above result being rejected (before the recent change). [...] peter
Received on Wednesday, 10 September 2008 18:24:03 UTC