W3C home > Mailing lists > Public > www-webont-wg@w3.org > May 2003

Re: owl:All(something) ??

From: Deborah L. McGuinness <dlm@ksl.Stanford.EDU>
Date: Mon, 12 May 2003 09:41:40 -0700
Message-ID: <3EBFCEC4.6080401@ksl.stanford.edu>
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
CC: hendler@cs.umd.edu, www-webont-wg@w3.org
i would oppose taking all different out
and would instead propose an ordering for additions (possibly all of 
which go in after last call).

next addition - allDisjointClass.  I agree with jeremy and ian that 
there are typically small numbers of classes involved and that it is not 
needed as often as the unique names assumption, but i do find that it is 
typically useful in most kbs.   even if a kb has been provided without 
disjointess statements, we may find that some applications first have a 
human do a pass over the kb and add disjointness at the top layers of 
the ontology just to limit options in an interface that uses the 
ontology or limit reasoning paths that can be pursued.

following that put in the allSame   (it is a convenience that people 
will look for and expect since we have different.)  

i must admit, i never thought about disjoint properties before and have 
never been asked for it as an expressive construct.  I have on the other 
hand been asked for the equivalent of allDisjointClass and have used 
this when i had it in my underlying language.


Peter F. Patel-Schneider wrote:

>From: Jim Hendler <hendler@cs.umd.edu>
>Subject: owl:All(something) ??
>Date: Sun, 11 May 2003 22:36:40 -0400
>>We have received a couple of comments that we should have some other 
>>constructs similar to allDifferent
>>  one comment suggest AllSame
>>another asks for AllDisjoint
>>we need an answer as to why we included the first and not the latter 
>>two, or we need to decide our current mechanism could be extended to 
>>allow it to be used on other properties  (which would probably not 
>>require another LC if we make it clear this is just a piece of 
>>syntactic sugar since we have the appropriate semantics already).
>>Any thoughts?
>One problem with this line of thinking is that we do not have the
>appropriate semantics already.  Every change to the syntactic constructs of
>OWL requires a change to the underlying semantics for OWL, because each
>such construct also has to exist in interpretations and have its own
>special semantic constraints.  Each construct must have its own
>section in all the correspondence proofs, as well.  Changes to the
>syntax of OWL cannot simply be syntactic sugar.
>The reason that AllDifferent was added to OWL was to provide a mechanism to
>easily handle a local unique-names-assertion, which would otherwise take
>O(n^2) differentIndividuals assertions.  No similar case can be made for
>the other constructs.
>If, however, it is deemed more important to be consistent in whether such
>constructs are permitted it would be better, in my view, to remove
>Peter F. Patel-Schneider
>Bell Labs Research
>Lucent Technologies
Received on Monday, 12 May 2003 12:41:56 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:04:44 UTC