- From: Bijan Parsia <bparsia@cs.man.ac.uk>
- Date: Tue, 5 Oct 2010 23:03:46 +0100
- To: Adrian Walker <adriandwalker@gmail.com>
- Cc: Pavel Klinov <pklinov@cs.man.ac.uk>, public-owl-dev@w3.org
On 5 Oct 2010, at 22:29, Adrian Walker wrote: > Hi Pavel -- > > A belated reply to your note. > > The question I'm worrying at is that I'd like to understand what we buy with the added complexity of OWL over rules. Actually, I don't think it is. > You wrote: I'm not sure which semantics > your system implements. The system is based on the model theoretic semantics in [1], with decidable and terminating computation as in [2]. See, you *could* answer concretely here, but you don't. You could respond to the characterisation that Pavel and I conjectured, but you don't. Looking at: http://oai.cwi.nl/oai/asset/10404/10404A.pdf I see you interpret in minimal models. It's just Datalog. So, DL Safe rules is the closest. And Pavel is quite correct. > ...it would not entail, for example, that some subset of Person love cats (e.g. "men love cats" although men are persons). > > This also seems easy using rules, as follows [3] > > some-class1 is a subclass of some-class2 > all people in that-class2 love items in some-class > --------------------------------------------------- > all people in that-class1 love items in that-class He didn't say that you couldn't create some sort of program that did something which one can loosely interpret as that inference. And, of course, it would help if you used a standard notation. (First order would be fine.) Markus posted links to his papers and his dissertation which discuss the representation problem in detail. Why not, you know, at least skim them to determine *the representation task*? Then you can try to, you know, perform it. To put it another way, where's your proof that your representation above is equi-satisfiable with the various OWL versions? If it is not, then you haven't solved it. The burden of proof is, obviously, on you. Cheers, Bijan.
Received on Tuesday, 5 October 2010 22:04:22 UTC