W3C home > Mailing lists > Public > public-owl-comments@w3.org > November 2012

OWL 2 DL restrictions: an issue

From: Diego Magro <magro@di.unito.it>
Date: Sun, 18 Nov 2012 17:40:13 +0100
Message-ID: <50A90F6D.5000301@di.unito.it>
To: public-owl-comments@w3.org
Dear all,

I wonder whether the Global Restrictions on Axioms in OWL 2 DL 
(http://www.w3.org/TR/owl2-syntax/#The_Restrictions_on_the_Axiom_Closure) refer 
to the deductive closure of an ontology.

If not, I think that there could be some problems with the 
specifications of the restrictions for an OWL 2 DL ontology.

Consider, for instance, an ontology containing:
- two sibling classes, A and B
- an axiom stating that  A and B are disjoint
- an object property p
- two axioms stating that the domain and range of p are A and B, 
- nothing else

Such an ontology seems to be in the OWL 2 DL profile.

However, such an ontology implicitly states that p is transitive, 
irreflexive and asymmetric. Thus, the deductive closure of the original 
ontology would not be in the OWL 2 DL profile.

I checked the original ontology and its deductive closure with the 
Manchester University OWL validator 
(http://owl.cs.manchester.ac.uk/validator/), getting that the original 
ontology is in the OWL 2 DL profile, while its deductive closure is not, 
which seems rather counterintuitive.

For computing the deductive closure of the original ontology I used 
HermiT 1.3.6


Best regards,
diego magro


Diego Magro
Dipartimento di Informatica
Universita` di Torino
via Pessinetto 12
10149 Torino Italy
Tel. +39 011 6706796
Fax. +39 011 751603
Received on Monday, 19 November 2012 09:05:54 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:01:31 UTC