To Use Domain/Range or Not To Use ...

... that is the question.

I am struggling over whether to use domain and range constraints in my 
ontology and could use some guidance.

Over the past year, I have picked up on various utterances which would 
seem to allude to potential pitfalls in using global domain and range 
restrictions within the context of the Web.  As far as I can tell, 
though, there is no formal documentation of the issue.  Sounds like a 
job for the SWBP ...

Some examples:

[1] "Could a user of your schema wish to apply your predicates to some 
other classes of which you have not yet thought?  If this is possible, 
range and domain constraints would seem inappropriate."

[2] "The solution of using the DAML/OWL union construct comes closest 
to giving you what you want.  However, the union construct is 
non-modular.  If a user wants to extend a union domain by  adding a new 
'subdomain' to it, that requires modifying the original union 
definition.  If this user is not the owner of the declaration that 
defines the union, then he/she is out of luck."

[3] "Indeed, many people don't use any global domain restrictions.  
Part of this is (as you correctly mention) the open world assumption in 
OWL on the Semantic Web.  Also, some reasoners behave badly inefficient 
if too many global restrictions are imposed on properties.  Thus, local 
restrictions are probably the safest bet for many applications."

--- Stephen

[1] 
http://lists.w3.org/Archives/Public/www-rdf-interest/2003Jan/0180.html
[2] 
http://lists.w3.org/Archives/Public/www-rdf-interest/2003Jan/0185.html
[3] Holger Knublauch (protege-owl list)

Received on Tuesday, 4 May 2004 21:53:14 UTC