W3C home > Mailing lists > Public > www-rdf-logic@w3.org > May 2002

Re: rdf inclusion

From: patrick hayes <phayes@ai.uwf.edu>
Date: Wed, 29 May 2002 10:02:46 -0500
Message-Id: <p0511171bb91a9d983b8a@[]>
To: "R.V.Guha" <guha@guha.com>
Cc: www-rdf-logic@w3.org

>   As far as I understand, contexts/microtheories and argumentation 
>are two examples of CycL features that add something to the 
>semantics and are not just efficiency hacks.

OK, OK, I forgot about contexts. I don't think much of contexts in 
any case, as you know. The argumentation stuff seems to be an 
inference strategy rather than a different logic, though: it *uses* 
first-order reasoning, doesn't it?

But to get back to the point, what makes CYC 'extensible' , whatever 
its logic is ??


>patrick hayes wrote:
>>>pat hayes wrote:
>>>>What does that mean? CYCL is a notational variety of FOL, right? 
>>>>In what way is that 'extensible' ? It doesn't enable one to 
>>>>extend the model theory.
>>>Come on Pat. You know better! CycL is way more than that ... Its 
>>>the kitchen sink of a large scale cafeteria plus
>>>many other things thrown into it, discussion of which is probably 
>>>neither apropos for this mailing list or for polite company ;-)
>>As far as I know, it is not *way* more than that. It has many 
>>pieces of specialized inference machinery, but most of them perform 
>>first-order valid inferences, I believe, albeit perhaps by 
>>nonstandard (and very fast) methods. I realize it has some 
>>technically non-first-order things built into it involving fixpoint 
>>semantics, but I bet they could all be transcribed into KIF using 
>>row quantifiers. And none of it is 'extensible' in any useful 
>>sense, as far as I know.  Please enlighten me if I am wrong.

IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
Received on Wednesday, 29 May 2002 11:02:38 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 2 March 2016 11:10:38 UTC