Search Mailing List Archives
[protege-owl] WG: Re: owl:AllDifferent
Raj.Mudunuri at medizin.uni-leipzig.de
Thu Sep 7 09:32:41 PDT 2006
Oops, I just saw this before resending the same query again to the list... sorry all, and thanq Arto...
Von: protege-owl-bounce at crg-gw.Stanford.EDU [mailto:protege-owl-bounce at crg-gw.Stanford.EDU] Im Auftrag von Arto Viitanen
Gesendet: Mittwoch, 30. August 2006 12:10
An: protege-owl at SMI.Stanford.EDU
Betreff: [protege-owl] Re: owl:AllDifferent
Mudunuri, Raj wrote:
> I'm just pondering, why the individuals of owl ontologies in Protege are
> not assumed as different entities by default... I mean, to define that our
> individuals are different, we have to explicitely say this by using the
> owl:AllDifferent construct... I can't imagine a case where we have two
> individuals and that we want the reasoner to assume that these two are
> the same... Anyhow the reasoning is done at concept level, but when we
> start assigning individuals then we really mean that all the individuals
> are different, don't we? So the reasoner can pop up a warning/error when
> it finds that two individuals have a clash where these two individuals
> seem to be the same according to the reasoner. Opposite to the way it is
> done now, I guess!?
> So, my question is, why don't we make all the individuals in the
> ontology as different by default, instead of the user making it for
> every individual of every class? Could some one tell me what problems
> could arise if we do it this way?
it comes from the semantic web idea: you have some parts of the
ontology on different web sites and youcan not change meaning. You find
thing A from siteA and thing B from siteB. Since you know nothing else,
based on your reasoner they are different and you can make lots of
reasoning on that. Then you find siteC that tells that they are
same, so all your reasoning were false. Using the current reasoner, you
cannot make those assumptions, so you have to assume that A and B might
be the same. When you then get siteD which tells that they are
different, nothing of your reasoning is false.
Arto Viitanen, CSC Ltd.
To unsubscribe go to http://protege.stanford.edu/community/subscribe.html
More information about the protege-owl