- From: TrungKien Tran <trung.tran@uni-ulm.de>
- Date: Mon, 8 Apr 2013 15:41:37 +0200
- To: public-owl-comments@w3.org
- Message-ID: <CAORpOhXS0M31NvC96ibtmN1Zj4kLPxcxi32MuLz0piTEoJqdgg@mail.gmail.com>
* Dear OWL 2 RL Working Group, According to syntactic restriction of OWL 2 RL [1], nominals/ObjectOneOfare not allowed in SuperClass Expressions. However, allowing ObjectOneOf that contains exactly one individual in SuperClas Expressions does not increase reasoning complexity. The reason is that SubClassOf axioms, which have such ObjectOneOf in SuperClass Expressions can be simulated in OWL 2 RL as follows: SubClassOf(a:SubClassExpression, ObjectOneOf(a:Individual)), where a:SubClassExpression is a SubClass Expression described in Table 2 of the profile specification [1] and a:Individual is a named individual, can be simulated by following OWL 2 RL axioms: ClassAssertion( a:Individual, a:ClassA ) (1) SubClassOf( a:SubClassExpression, a:ClassA ) (2) SubClassOf( a:ClassA, ObjectHasValue( a:hasRelation, a:AnotherIndividual ) ) (3) SubClassOf( ObjectOneOf( a:AnotherIndividual ), ObjectMaxCardinality( 1 ObjectIverseOf (a:hasRelation) a:ClassA ) ) (4) Intuitively, (3) and (4) entail that a:ClassA has at most one instance. In addition, from (1) we know a:Individual is an instance of a:ClassA. Therefore, a:ClassA has only one instance. Thus, we can derive: SubClassOf( a:ClassA , ObjectOneOf( a:Individual ) ) (5) >From (2) and (5), we have: SubClassOf( a:SubClassExpression, ObjectOneOf( a:Individual ) ) Although ObjectOneOf that has exactly one individual might not be as widely used as other axioms, it is still useful for ontology modelers. I would like to ask if there is any good reason to disallow such ObjectOneOf in SuperClass Expressions ? [1] http://www.w3.org/TR/owl2-profiles/#OWL_2_RL_2 Best regards, Kien Tran * -- ================================================= Trung-Kien TRAN Institute of Artificial Intelligence University of Ulm Tel: +49 (0)151 26 21 18 95 Homepage: http://www.uni-ulm.de/en/in/ki/staff/trung-kien-tran.html =================================================
Received on Monday, 8 April 2013 14:08:09 UTC