- From: Ian Horrocks <horrocks@cs.man.ac.uk>
- Date: Tue, 25 Mar 2003 19:05:42 +0000
- To: Dave Reynolds <der@hplb.hpl.hp.com>
- Cc: lchewhun@dso.org.sg, www-rdf-logic@w3.org
On March 25, Dave Reynolds writes: > > Ian Horrocks wrote: > > > > A couple of points. DLs certainly CAN represent predicates involving > > multiple variables and predicate chains, it just happens that the > > particular DL on which DAML/OWL are based does not include these > > features > > Apologies for my imprecise language - I should indeed have said "the DLs behind > DAML/OWL". > > > (there are a variety of reasons for this, not the least of > > which is the fact that the RDF and XML type systems do not specify > > n-ary predicates). > > I assume there are also complexity arguments. Predicate chains would be > compatible with the RDF substrate but I assumed that the combination of role > composition with the other operators in OWL would have pushed up the > computational complexity further. Is that correct? There are complexity and decidability issues. These are quite involved and may related to the datatypes and predicates themselves, as well as interactions with the class description language - full details in [2,3]. I believe that it would be pretty straightforward to extend OWL with n-ary predicates - see [1]. Ian [1] http://www.cs.man.ac.uk/~panz/Zhilin/download/Paper/Pan-Horrocks-datatype-2002.pdf > > > See [2,3] for details about DL support for datatypes > > (called concrete domains in DLs). > > Thanks for the references. > > Dave
Received on Tuesday, 25 March 2003 13:05:07 UTC