From: Deborah McGuinness (dlm@ksl.stanford.edu)
Date: 10/24/01
Sandro provided the RDF way in his response below (which i agree is better).
The typical way to handle getting existentials into DLs is to do so with
cardinality (qualified or non-qualified) restritions. I just didnt know the
RDF way so in my posting, i used the standard description logic way.
d
Date: Wed, 24 Oct 2001 15:07:10 -0400
From: Sandro Hawke <sandro@w3.org>
There is a much simpler way, since RDF includes existential variables.
<rdf:RDF xmlns:rdf="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
xmlns:example="http://example.com#">
<rdf:Description>
<example:color rdf:resource="http://example.com#red"/>
</rdf:Description>
</rdf:RDF>
This is parsed by the validator [1] to
_:j70241 <http://example.com#color> <http://example.com#red> .
As Pat's model theory [2] (and his various e-mails to me on
www-rdf-rules) make clear, the "unlabeled" nodes (like _:j70241)
should be interpretted as existential variables.
This clarification to RDF is fairly recent, and not yet official,
(it's just a working draft), but I haven't heard anyone objecting yet.
I have heard several members of the original RDF group say this is
what they meant.
-- sandro
[1] http://www.w3.org/RDF/Validator/
[2] http://www.w3.org/TR/rdf-mt/
Richard Fikes wrote:
> > > >The knowledge base contains the statements: "Pat's car is blue, and
> > > >there is something colored red." Somewhat more formally:
> > > >
> > > > RDF(PatsCar, color, blue).
> > > > exists x (RDF(x, color, red)).
> > >
> > > Sorry to be dense, but how does one state "there is something colored
> > > red" in DAML+OIL?
> > >
> > > Richard
>
> Your reply suggests that there is no direct of way of saying "exists x
> RDF(x color red)" in a DAML+OIL KB, which is what I suspected. Since
> the issue came up in a discussion of what statements in a DAML+OIL KB
> would match a query of the form "RDF(?x, color, ?y)", it is central to
> the discussion.
>
> Am I correct in concluding that there is no way of including an
> existential variable as the subject or object in an RDF statement in a
> DAML+OIL KB? (I suspect there is a way, but I don't see it.)
>
> Richard
>
> > there are a few ways, arguably, none perfect.
> >
> > the most direct is mincardinalityQ
> > state that something has mincardinality 1 (or more)
> > and the Q in this case is RED-THING
> >
> > (the bad thing about this is it requires a property to be defined for the
> > mincardinalityq statement
> > say MySpecialProperty so that i can say
> >
> > <daml:Restriction daml:maxCardinalityQ="1">
> > <daml:onProperty rdf:resource="#MySpecialProperty"/>
> > <daml:hasClassQ rdf:resource="#RedThing"/>
> > </daml:Restriction>
> >
> > (and this of course assumes red things to be defined which could be:
> >
> > <daml:Class rdf:ID="RedThing">
> > <daml:sameClassAs>
> > <daml:Restriction>
> > <daml:onProperty rdf:resource="#hasColor"/>
> > <daml:hasValue rdf:resource="#red"/>
> > </daml:Restriction>
> > </daml:sameClassAs>
> > </daml:Class>
> >
> >
> > one could also make individuals (just for the purposes of implying a
> > red-thing)
> > and say foo has min cardinality 1 on a property MySpecialProperty and
> > then have the value restriction on MySpecialProperty be RedThing and the
> > filler of MySpecialProperty
> > on foo, be bar.
> > this implies that bar is red and thus, there exists a red thing. One
> > would not need to explicitly create bar for the inference to be implied.
--
Deborah L. McGuinness
Knowledge Systems Laboratory
Gates Computer Science Building, 2A Room 241
Stanford University, Stanford, CA 94305-9020
email: dlm@ksl.stanford.edu
URL: http://ksl.stanford.edu/people/dlm/index.html
(voice) 650 723 9770 (stanford fax) 650 725 5850 (computer fax) 801 705
0941
This archive was generated by hypermail 2.1.4 : 04/02/02 EST