Re: P.S. Re: Model Theory

At 7:57 am -0500 7/1/02, Peter F. Patel-Schneider wrote:
>From: Enrico Motta <>
>Subject: Re: P.S. Re: Model Theory
>Date: Mon, 7 Jan 2002 11:37:05 +0000
>>  2) Some 'KR programming cliches' are useful in practice but are
>>  tricky to formalise in a model theory (e.g., reification, defaults,
>>  etc..). Now, taking a 'pragmatic' approach means that we do not
>>  <<automatically>> discard these constructs, simply because we do not
>>  know how to write the model theory.  Maybe an alternative approach to
>>  specification can be used, or maybe we simply state algorithmically
>>  how intepreters will deal with these constructs.
>If you (or someone else) can come up with a workable specification for
>these constructs, then I would be very happy.  The problem lies in
>determining what constructs mean, whether it be via a model theory, via a
>proof theory, via axiomatization, or even via some more-operational
>mechanism.  In the absence of a workable specification for a construct,
>e.g, reification, and, moreover, one that has good connection to the
>intended meaning of the construct, I am unwilling to include the construct
>in a representation language.

I'd be unwilling too!  I am not arguing for 'voodoo KR'. My point is 
that we should consider also the possibility to include constructs 
which have some other kind of workable specification. So, it seems to 
me we are in agreement on this.



>PS:  Actually defaults do have several decent specifications.  However, the
>specifications have very bad computational properties, which make them
>problematic in a *useful* representation language.

True, however there are in practice lots of frame-based 
representations which employ limited (but 'clean'!) mechanisms for 
defaults which can be implemented very efficiently, e.g.,  "Robert 
Nado, Richard Fikes: Semantically Sound Inheritance for a Formally 
Defined Frame Language with Defaults. AAAI 1987: 443-448"

Received on Monday, 7 January 2002 10:22:34 UTC