Re: defaults

From: Guus Schreiber <schreiber@swi.psy.uva.nl>
Subject: defaults
Date: Mon, 21 Jan 2002 17:06:24 +0100

> Peter,
> 
> During the f2f you mentioned that some limited representation of
> defaults could be possible in OWL. Could you elaborate on that? That would
> be very helpful. 
>
[...]
>
> Thanks, Guus

There is a very common idiom, which I like to call ``input completion'',
that works as follows.  I'll use frame terminology to describe input
completion for two reasons: 1/ the processing is much clearer in frame
terminology, and 2/ I hope that not using RDF terminology will help prevent
misunderstandings. 


An ``input completion'' directive is something like
	the normal value for slot S in class C is V
You process the definition of an object O in the input stream.  The next
thing that you do is to see if O belongs to C.  If it does not, then the
directive does not apply.  Then you check to see if O has a value for S.
If it does, then the directive does not apply.  If, however, O belongs to C
and also does not have a value for S, then give it the value V for S.  This
value is not asserted by default, i.e., information discovered later can
not override it. 


A coherent story for input completion is much harder to make for RDF, as
RDF does not have definitions of objects, nor does it have ordered inputs.
A coherent story for input completion is also difficult to make for
description logics, and thus would be hard to make for OWL.

peter

Received on Tuesday, 22 January 2002 06:45:15 UTC