Re: Something strange in OWL-S Upper Ontology

zze-VALLEE Mathieu RD-TECH-GRE wrote:

> David Martin wrote:
> 
> 
>>De Furio Ivano wrote:
> 
> 
>>>2) Another, strange thing (at least for me) is that both ontology 
>>>http://www.daml.org/services/owl-s/1.0/Profile.owl and 
>>>http://www.daml.org/services/owl-s/1.0/Process.owl define
> 
> hasParameter
> 
>>>as a owl:ObjectProperty with a well defined domain. But, hasInput
> 
> and 
> 
>>>hasOutput, was defined as subPropertyOf hasParameter without a
> 
> defined
> 
>>>domain.
>>>
>>>I wasn't able to find that subPropertyOf implies a domain
> 
> inheritance, 
> 
>>>so maibe should be better to repeat the domain.
> 
> 
>>Yes, subPropertyOf does imply domain inheritance, so it's okay not to 
> 
> repeat the domain.
> 
> Is there a special reason not to repeat the domain (apart from avoiding
> redundancy) ? 

Not to my knowledge; that is, it's just to avoid redundancy.

> When I edit services using Protege (and OWL plugin), it seems that I
> must add the domain for hasInput and hasOutput properties. If not,
> Protege only allows me to use the hasParameter property (it does not
> include reasoning to deduce this kind of domain inheritance).

It's a good point.  This is not the first time that a tool-building 
consideration may lead us to alter our notions of what is good style.

In the 1.1 Beta release (just about ready to be announced) I'm going to 
go ahead and add the domain explicitly for the benefit of tools (not 
everywhere, but in the crucial IOPE properties)..  But this is somewhat 
provisional; that is, if there are any important objections then it's 
possible these declarations will be removed.

Regards,
David Martin

> However, this kind of troubles may be solved by the (future) OWL-S
> editor :-)
> 
> Regards,
> Mathieu

Received on Friday, 23 July 2004 17:42:16 UTC