Re: Modeling large scale ontologies in OWL: Unmet needs

2006-09-21 Thread Phillip Lord

 DD == David Decraene [EMAIL PROTECTED] writes:

  DD In large scale ontologies, one link should suffice,
  DD HasPart, and whether the part is a finger, toe, nail, muscle or
  DD anything else is not a task for the property to describe, but
  DD for the target


I'm not sure why this should be true for large ontologies. It seems to
me that this is just a question of modelling style. Either way should
actually work depending on what you are trying to achieve. 

Having multiple properties allows you to provide different properties
to the properties which can be useful. If, in your example, you have a
super property hasPart, then it seems to me that it would be
relatively straight forward to reduce the information content of the
ontology so that the subproperties are no longer represented. So hand
some hasDigit Finger can be represented as hand some hasPart Finger. 


  DD In formal ontology you could express this relation
  DD on a general level of parthood: Hand HasPart 6thfinger,
  DD cardinality 0. This is not possible in OWL.

Many people have used subproperties to do something like this. It's a
poor hack for representing qualified cardinality (and doesn't capture
exactly the same semantics). But it is only a hack. As others have
said, the lack of qualified cardinality in OWL is generally regarded
as unfortunate, and it should be coming back in. 

Phil






-- 
Phillip Lord,   Phone: +44 (0) 191 222 7827
Lecturer in Bioinformatics, Email: [EMAIL PROTECTED]
School of Computing Science,
http://homepages.cs.ncl.ac.uk/phillip.lord
Newcastle University,   Claremont Tower, Room 909
NE1 7RU



Re: Modeling large scale ontologies in OWL: Unmet needs

2006-09-20 Thread Alan Ruttenberg


Hello David,

I think you are referring to the lack of qualified cardinality  
restrictions (http://www.w3.org/2001/sw/BestPractices/OEP/QCR/)


The proposed OWL 1.1 includes them and  should shortly (in the next  
few months) be supported by the major OWL reasoners, including  
Pellet, and FaCT++ (and hence Protege).


OWL 1.1 is described at http://owl1_1.cs.manchester.ac.uk/. In the  
overview search for SROIQ.


There is an OWL workshop coming up: http://owl-workshop.man.ac.uk/ 
OWLWorkshop06.html

Perhaps you should consider attending.

Let me know if I've misunderstood.

Regards,
Alan