W3C home > Mailing lists > Public > www-webont-wg@w3.org > September 2002

Re: SEM: All OWL reasoners will be incomplete?

From: Jos De_Roo <jos.deroo.jd@belgium.agfa.com>
Date: Thu, 5 Sep 2002 23:32:41 +0200
To: Ian Horrocks <horrocks@cs.man.ac.uk>
Cc: "Jeremy Carroll" <jjc@hplb.hpl.hp.com>, www-webont-wg@w3.org
Message-ID: <OFB0CDBF60.EF33D326-ONC1256C2B.0074BD2E@agfa.be>
[...]

> It is a completely different thing if we can infer the existence of an
> axiom asserting that a property is transitive (as would be the case if
> OWL syntax were in the domain of discourse). In this case, we need to
> perform inference (in a language which is almost certainly
> undecidable) on the ontology in order to determine if it conforms with
> our definition of what is a syntactically valid ontology. I was
> suggesting (in point 1b) that we don't want to go there!

using only 1 engine I can understand your point
however, even for such a simple problem as e.g.
http://www.agfa.com/w3c/euler/etc5.n3
we actually are using 180 engines
by which I mean that engines can produce
evidence for each other and it is not a
difficult problem in control engineering
(there's just no central point of control)
and in that way I don't see any issues
w.r.t. transitive properties

-- ,
Jos De Roo, AGFA http://www.agfa.com/w3c/jdroo/
Received on Thursday, 5 September 2002 17:33:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:52 GMT