Re: question about DL/Lite semantics

From: Jeremy Carroll <jjc@hpl.hp.com>
Subject: question about DL/Lite semantics
Date: Mon, 2 Jun 2003 12:13:23 +0300

> 
> 
> On one of my homework assignments after the beer session I've got a little bit 
> stuck considering the following entailments according to the direct semantics
> 
> Does
> 
> Class( <foo> annotation( <a> <bar> ) )
> AnnotationProperty( <a> )
> Class( <bar> )
> 
> entail
> 
> AnnotationProperty( <a> )
> Class( <foo> annotation( <a> individual() ) )
> 
> ?

It should.  (Well it should if ``individual'' was changed to
``Individual''.)

> If not, why not?

Well, one reason why it might not follow is that annotations might not
relate to things.  Currently, they do.

This is, however, a weak part of the semantics, and one that one probably
needs to be changed.  The fix would be to have the OWL universe be only a
subset of the entire universe.

> I also note that there is a bug in the LC mapping rules that permits 
> 
> [I]ndividual() 

> to map to nothing
> so that the following contradicition maps to a noncontradition in RDF/XML
> 
> [I]ndividual()
> EquivalentClasses( owl:Thing, owl:Nothing ) 

Ah yes, this is a problem because the requirement on typing individuals is
only for individualIDs.  The best fix is to require that an Individual(...)
axiom without an individual ID and without any type gets typed as
owl:Thing.  I will add this.

> I am still working on the rdf:List homework, and will possibly propose a bug 
> fix for this bug as a side effect.
> 
> (The equivalent expression
> 
> individual( type(owl:Thing) )
> EquivalentClasses( owl:Thing, owl:Nothing ) 
> 
> is a simple contradiction)

Agreed, and the source for my fix above.

> Jeremy

peter

Received on Monday, 2 June 2003 10:29:32 UTC