Re: Annotations and non-mon example

From: Jeremy Carroll <jjc@hpl.hp.com>
Subject: Annotations and non-mon example
Date: Thu, 30 Jan 2003 19:52:50 +0100

> 
> 
> 
> For those who didn't follow the WG discussion, here is a quick test case 
> showing OWL DL annotations as non-monotonic
> 
> 
> <owl:Thing rdf:about="eg:a">
>    <my:prop>foo</my:prop>
> </owl:Thing>
> 
> DL-entails
> 
> <owl:Thing rdf:about="eg:a">
>    <my:prop>bar</my:prop>
> </owl:Thing>

In the direct semantics (after translating back to the abstract syntax)
this is an entailment, in the RDF-compatible semantics this is not an
entailment.

> =========
> 
> Because, my:prop has no property declaration, and is hence treated as an 
> annotation, which is given no semantics in OWL DL.
> 
> =========
> 
> Adding the same information to both files breaks the entailment.
> viz:
> 
> <owl:DatatypeProperty rdf:about="&my;prop"/>
> <owl:Thing rdf:about="eg:a">
>    <my:prop>foo</my:prop>
> </owl:Thing>
> 
> does not entail
> 
> <owl:DatatypeProperty rdf:about="&my;prop"/>
> <owl:Thing rdf:about="eg:a">
>    <my:prop>bar</my:prop>
> </owl:Thing>

In the direct semantics (after translating back to the abstract syntax)
this is not an entailment, in the RDF-compatible semantics this is not an
entailment.

*However*, in the abstract syntax this is not an addition of new
 information.  Instead, it is a *change*.  Therefore there is no
 non-monotonicity.

> 
> ====
> 
> because this time my:prop is treated as semantically meaningful.
> 
> Jeremy

peter

Received on Friday, 31 January 2003 08:43:29 UTC