W3C home > Mailing lists > Public > www-rdf-rules@w3.org > July 2005

Re: Web Rule Language - WRL vs SWRL

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Fri, 1 Jul 2005 02:26:27 -0400 (EDT)
To: Michael Kifer <kifer@cs.sunysb.edu>
Cc: www-rdf-rules@w3.org
Message-Id: <20050701022627.CRU00209@isrmail.isr.umd.edu>

---- Original message ----
>Date: Fri, 01 Jul 2005 00:44:38 -0400
>From: kifer@cs.sunysb.edu (Michael Kifer)  
>Subject: Re: Web Rule Language - WRL vs SWRL  
>To: bparsia@isr.umd.edu
>Cc: www-rdf-rules@w3.org

>Bijan Parsia wrote:
>>
>> Drew McDermott wrote:
>> >
>> > One could develop a pretty formal account of how
>> > this thing worked, including an account of what inferences were
>> > licensed under what circumstances.  But it would have nothing to do
>> > with the _semantics_, which could be specified in advance,
>> > independently of the details of the inferential mechanisms.
>> 
>> I'll be interested to know if Michael thinks this is a defense of him :)
>
>I think that here by "semantics" Drew means Tarskian-like model theory --
>definitions of what is and is not a model. By choosing a subset of
>"intended" models one can define entailment relations of various degrees of
>hairiness.
>
>This is a standard story of semantics for nonmon inference.
>I don't see what you think should upset me here. 

I *think* (and my thinkings of what people would say have been notably 
unsuccessful) that Drew would say that you punned on 'semantics' above (as 
indicated by your use of scare quotes in the first one). I took Drew to be denying 
the second story of semantics, which, if I had been right, would, I believe, put 
you in opposition.

>I may be misunderstanding either or both of you. :-)

Well, consensus seems to be landing the misunderstanding on me, even if 
everyone is willing to claim it possible that it is theirs :) Though it still looks to 
me that you are working semantically (defining an entailment relation) where 
Drew denies that one is ("""But it would have nothing to do with the _semantics_, 
which could be specified in advance, independently of the details of the 
inferential mechanisms."""). Of course, a different set of sanctioned entailments 
will have different sound and complete inference mechanisms, and I guess I 
didn't read the "in advance" enough (though, it does seem that it is in 
advance....or as in advance as it has been historically for FOL).

But I now have little clue as to what's going on :) I would like to join the 
indifferentists and say that these details are negligible, but alas, that hasn't 
been my experience. The layering story in OWL has been complex enough to 
work with and explain that I'm rather gunshy at adding more variance. I also 
think you underestimate how difficult one's life can be made in a working group 
by *not* having such issues resolved. The time we've spent in this thread can be 
easily dwarfed in a working group :) (Well, the SWSL experience may give you a 
good idea!)

Cheers,
Bijan.
Received on Friday, 1 July 2005 06:26:48 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:53:12 GMT