RE: NAF and owl:complementOf semantics?

Thank you for the comment.

I referred to a paper[1] and found out that, to utilize LCW,
you need to enumerate all the instances of a class to
determine the complement of it. I think it's impractical,
because you cannot effectively enumerate all the individuals
of a class... Is there any reasoner that can generate all
the individuals contained in the complement of a class?

Regards,
Minsu

[1] http://www.cse.lehigh.edu/~heflin/pubs/#aaai-02

> -----Original Message-----
> From: www-rdf-rules-request@w3.org 
> [mailto:www-rdf-rules-request@w3.org] On Behalf Of Wagner, G.R.
> Sent: Wednesday, December 10, 2003 4:27 AM
> To: minsu@etri.re.kr; www-rdf-rules@w3.org
> Subject: RE: NAF and owl:complementOf semantics?
> 
> 
> 
> 
> > According to the S&AS, owl:complementOf(c) is interpreted
> > as O - EC(c), which is, as I understand, a set of individuals
> > which are not contained in the set of individuals of type c.
> > 
> > For me, this was easily taken as a semantic which can be
> > implemented simply by negation-as-failure. I could formulate
> > the semantic into a rule as follows:
> > 
> > if
> >    owl:complementOf(?c1,?c2) and owl:Thing(?x) and not ?c2(?x)
> > then
> >    ?c1(?x);
> > 
> > Is this a proper axiomatization of owl:complementOf?
> 
> No, because it depends on whether you have a complete
> representation of "EC(?c2)" in the scope of your KB.
> To get this, you would need to make an explicit completeness
> assumption (also called "local closed-world assumption", see
> previous messages in this list) for the class denoted by ?c2.
> 
> Gerd Wagner
> http://tmitwww.tm.tue.nl/staff/gwagner/
>

Received on Friday, 12 December 2003 00:56:53 UTC