Re: semantics document revised

From: Dan Connolly <connolly@w3.org>
Subject: Re: semantics document revised
Date: 19 Sep 2002 09:04:10 -0500

> 
> On Wed, 2002-09-11 at 11:45, Peter F. Patel-Schneider wrote:
> > I fixed a couple of problems in my semantics document, and reordered some
> > of the material.
> 
> I wonder if it's essential to restrict these this way:
> 
> |owl:FunctionalProperty rdf:subClassOf owl:Property .
> |owl:SymmetricProperty rdf:subClassOf owl:ObjectProperty .
> |owl:InverseFunctionalProperty rdf:subClassOf owl:ObjectProperty .
> 
> Can we either
>   (a) not restrict these this way.
>     - note that folks can still declare their properties
>     to be both a FunctionalProperty and an owl:ObjectProperty.
>   (b) specify analogs that aren't restricted.

Either might work. 

> i.e. I'm still interested in replies to...
> 
>  SEM/TEST: restricted complement, unrestricted InverseFunctionalProperty
>  Dan Connolly (Fri, Sep 06 2002) 
>  http://lists.w3.org/Archives/Public/www-webont-wg/2002Sep/0075.html
> 
> 
> Likewise for disjointFrom...
> 
> 
>         1. The class extension of owl:Thing is a subset of IR, does not
>            contain any RDF or OWL structural resources, and is disjoint
>            from the extensions of classes, properties, lists,
>            collections, statements, and literals. 
>             1. for each name N in the rdf:, rdfs:, and owl: namespaces,
>                IS(N) is not in ICEXT(IS(owl:Thing)) 
>             2. ICEXT(IS(owl:Thing)) is disjoint from
>                ICEXT(IS(rdfs:Class))
>             3. ICEXT(IS(owl:Thing)) is disjoint from
>                ICEXT(IS(rdf:Property))
>             4. ICEXT(IS(owl:Thing)) is disjoint from ICEXT(IS(rdf:List))
>             5. ICEXT(IS(owl:Thing)) is disjoint from
>                ICEXT(IS(rdf:Collection))
>             6. ICEXT(IS(owl:Thing)) is disjoint from
>                ICEXT(IS(rdf:Statement))
>             7. ICEXT(IS(owl:Thing)) is disjoint from LV
> 
> It seems highly ironic, to me, that we have to state that informally.

The above is a formal as things get.

> If owl:disjointFrom were specified over the whole domain of
> discourse, we could write, formally...
> 
>   owl:Thing owl:disjointFrom rdfs:Class.
>   owl:Think owl:disjointFrom rdf:Property.

I believe that this would work.    However, I prefer the former way of
stating the conditions.

> etc.

> -- 
> Dan Connolly, W3C http://www.w3.org/People/Connolly/


peter

Received on Thursday, 19 September 2002 10:55:14 UTC