Re: Concrete and abstract domains disjointness

 You can also understand the separation from an intuitive point of 
view. You can assume that the type system already providesways for 
definition and derivation of datatypes. You don't need the DL language 
to form new datatypes (DLs are not defined for such a purpose). 
However, you still need some sort of connection between the object 
domain and the domain of datatypes. This connection is provided by the 
datatype properties. this connection can be used to "import" 
information from the type system and use it to build new concepts 
(through restrictions on datatype properties)

 Regards

 Bernardo Cuenca Grau

> 
> Dimitrios A. Koutsomitropoulos wrote:
> > 
> > Can somebody explain some formal reason why the concrete and 
abstract
> > domains (i.e. the datatype and individual sets) have to be disjoint 
in OWL
> > DL?
> 
> I think in addition to the other postings, I would suggest procedural 
> reasons. Those who like this separation could point to 
implementations 
> (both real and theoretical algorithms) based on this idea, and had 
the 
> advantage that Daml+OIL has this separation. Anyone wishing to 
challenge 
> that really needed to point to working systems, with academic 
> creditionals, that was as credible. I don't think this point was 
> seriously challenged.
> Politically, those who were most likely to want to challenge this 
> separation were happy enough with OWL Full.
> 
> The point of deliberately ignoring your request for *formal* reasons, 
is 
> that any such reasons will be the views of some (but not all) of the 
WG. 
> OWL was determined using a process, which was (roughly) to take 
DAML+OIL 
> and raise issues against it. IIRC this issue was not (formally) 
raised, 
> so the separation is in OWL because it was in DAML+OIL (not for 
> technical reasons)
> 
> 
> Jeremy
> 
> 
> 
> 
>

Received on Monday, 10 May 2004 14:17:15 UTC