Re: Owl question

From: "Dickinson, Ian J" <Ian.Dickinson@hp.com>
Subject: Owl question
Date: Mon, 31 Mar 2003 19:58:34 +0100

> In the OWL AS&S, restrictions and Boolean class expressions are different
> productions, so they cannot be mixed in a single class description.  In
> (normative) appendix B of OWL Reference, owl:Restriction is a sub-class of
> owl:Class, and the domain of owl:intersectionOf (etc) is owl:Class.  Which
> suggests that this:
> 
> <owl:Restriction rdf:ID="Odd">
>     <owl:intersectionOf rdf:parseType="Collection">
>         <owl:Class rdf:ID="A" />
>         <owl:Class rdf:ID="B" />
>     </owl:intersectionOf>
> </owl:Restriction>
> 
> is a reasonable construct.  Given that it would be fairly easy to define
> owl.owl in such a way as to better reflect the syntactic structuring in
> AS&S, I wonder why there's a discrepancy - especially as both sources are
> "normative"?
> 
> Cheers,
> Ian

This is one of the disconnects between RDF and syntax.

The above construct is OWL Full, as is every RDF/XML construct.  It even
has a reasonable meaning, which is pretty much just an OWL class, as there
is not enough information to trigger any extra restriction meaning.  It is
not OWL DL (or OWL Lite) however, as it is not a mapping of any abstract
syntax construct.

In RDF (and RDF/XML) it is impossible to prohibit such odd
constructions. 

peter

Received on Tuesday, 8 April 2003 14:41:39 UTC