Re: LANG: owl:Thing subClassOf owl:Thing

From: "Jeremy Carroll" <jjc@hpl.hp.com>
Subject: LANG: owl:Thing subClassOf owl:Thing
Date: Wed, 18 Dec 2002 17:00:10 +0100

> 
> Working through the abstract syntax it seems that virtually any OWL Lite or
> OWL DL file needs to include the somewhat interesting triples
> 
> owl:Thing rdf:type owl:Class .
> owl:Thing subClassOf owl:Thing .
> 
> Have I misunderstood? It might be a typo in the abstract syntax mappings.

This is a bug in the mapping, which I just changed.

> e.g.
> 
> suppose we want to say
> 
> <a> <p> <b> .
> 
> To get this triple,
> we have to use the rule from the mappings:
> 
> Individual(<iID>
>       <annotation1> … <annotationn>
>       type(<type1>)…type(<typen>)
>       (<pID1> <value1>) … (<pIDn> <valuen>))
> 
> which has a T(owl:Thing) in its production.
> 
> This must use the classID rule that includes
> 
> owl:Thing rdfs:subClassOf owl:Thing .
> 
> in its production.
> 
> I don't see the need for the
> 
> <classID> rdf:subClassOf owl:Thing .
> 
> triple in the second row of the table, ever.

This is now gone.

> Is this not implicit with the
> 
> <classID> rdf:type owl:Class .
> 
> triple?

It is entailed by the OWL RDFS-compatible model theory.

> Jeremy

What is *not* gone is the need to explicitly state that all individiuals
belong to owl:Thing.  I was trying to figure out how to remove this, but I
can't see a way that works.

peter

Received on Wednesday, 18 December 2002 11:51:40 UTC