W3C home > Mailing lists > Public > www-webont-wg@w3.org > August 2002

RE: SEM/GUIDE: subclasses of classes of Properties? (5.3)

From: Dan Connolly <connolly@w3.org>
Date: 09 Aug 2002 23:27:10 -0400
To: "Smith, Michael K" <michael.smith@eds.com>
Cc: Christopher Welty <welty@us.ibm.com>, www-webont-wg@w3.org
Message-Id: <1028950033.565.4.camel@jammer>

On Fri, 2002-08-09 at 16:24, Smith, Michael K wrote:
> Chris, Dan, Jos
> 
> I confess to agreeing with Chris: I find this example confusing.  I am 
> missing something. 
> 
> Basicly I didn't understand:
> 
>    db:KeyProperty rdfs:subClassOf owl:FunctionalProperty.
> 
> First I thought, huh, is this how we say something is a functional 
> property?

No, it's how we say that each member of a class (KeyProperty) is
a functional property.

>  I would have thought
> 
>    db:KeyProperty rdfs:subPropertyOf owl:FunctionalProperty.

no, FunctionalProperty is a class, not a property.
subPropertyOf relates two properties.


> Since nothing in any of the formalizations of RDF or OWL that I have
> seen seems to relate class inclusion to property inclusion.  
> 
> > I want FunctionalProperty to be usable like any other RDFS Class.
> 
> So, now maybe I understand the example better.  The example
> illustrates something that Dan wants. A defined relationship between
> classes and properties, such that if
> 
>  A rdfs:subClassOf B
>  B rdf:type rdfs:Property
>  
> then whatever property-related restrictions or definitions apply to B, 
> they are inherited, VIA THE SUBCLASSOF RELATION, by A.

No, it's nothing like that. I'm just using the
	C1 subClassOf C2
	X type C1
	-----
	X type C2

rule from RDFS, where C1 is KeyProperty, C2 is FunctionalProperty,
and X is fred:customer.

>  ???  E.g.
> 
>  A rdfs:subClassOf B
>  B rdf:type rdfs:Property
> ->
>  A rdfs:subPropertyOf B
> 
> Is this the point?
> 
> As a technical note, I don't understand how Jos's example can work,
> since the default namespaces for the hypothesis and conclusion are
> different in the two files referenced.  And therefore the
> fred#customer references in the two files are to different
> resources. Did I miss something?

Namespaces only apply to qnames, not to URI references.

> - Mike
> 
-- 
Dan Connolly, W3C http://www.w3.org/People/Connolly/
Received on Friday, 9 August 2002 23:27:15 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:51 GMT