Re: Semantic layering bug (not again!)

From: Jeremy Carroll <jjc@hpl.hp.com>
Subject: Semantic layering bug (not again!)
Date: Mon, 22 Sep 2003 13:39:38 +0300

> While doing the action concerning adding tests exhbiting the change agreed 
> concerning owl:Ontology I noted that there is a further bug on semantic 
> layering.
> 
> We agreed:
> 
> http://www.w3.org/2001/sw/WebOnt/webont-issues.html#I5.3-Semantic-Layering
> 
> 
>  RESOLVED: to close the layering issue (5.3) as described in Consensus on 
> semantic layering, with the change that KB large-OWL-entails C if 
> KB-fast-OWL-entails C (but not only if -- Large OWL is now known as Owl Full, 
> Fast Owl is now known as OWL DL). 
> 
> 
> Unfortunately the following test contradicts that (also visisble in new tests 
> Ontology-002, and Ontology-003).
> 
> ===
> 
> <Empty>
> 
> entails
> 
> _:a rdf:type owl:Ontology .
> 
> ====
> 
> I think this is an OWL Lite entailment but an OWL Full non-entailment.
> Technically this follows from the decision we made to make such triples 
> optional - I do not think making them compulsory would be the right fix.
> 
> Jeremy

Yuck.  

There is a (two-part) solution.

1/ Make the ontology typing triples compulsory.
2/ Expand the notion of OWL entailment to allow for directives (excluding
   ontology annotations) that are not part of an ontology and to require
   that there are sufficient ontologies on the RHS.

peter

Received on Monday, 22 September 2003 11:32:07 UTC