Re: Axiomatic Semantics of OWL

On 18 Oct 2004, at 10:08, Jos de Bruijn wrote:
>> I wonder whether it has been proved that the proposed axiomatisation 
>> - restricted to the OWL-DL fragment - is proved sound and complete 
>> with respect to the published normative model theoretic semantics of 
>> OWL-DL.
>> In other words: how can we be sure that if we use your axiomatisation 
>> for just OWL-DL we get the same behaviour as if we were just using a 
>> standard DL reasoner (already proved to be sound and complete wrt the 
>> published normative model theoretic semantics of OWL-DL)?
>
> I would assume this would not be the case, since it was already 
> demonstrated in [1, Appendix B.2] that OWL Full semantics allows for 
> more entailments than OWL DL semantics for an OWL DL ontology. So, if 
> it properly implements the OWL Full semantics, it does not properly 
> implement the OWL DL semantics.

Ah-ah. I forgot this. This reminds me that really there is no 
interoperability nor smooth transition  between OWL-DL and OWL-Full. 
These are basically two different ontology languages with little in 
common. The same OWL-DL ontology will behave differently when used by 
an OWL-DL tool wrt an OWL-Full tool.
Not bad for a standard that should enable interoperability.
cheers
--e.