[SCL] Re: [KIF] (SCL) Seqvars revisited

[Bill Andersen]

On Wednesday, Dec 18, 2002, at 11:05 US/Eastern, pat hayes wrote:

> Heres a sketch. A relational atom is treated in CL basically as a 
> predication on tuples, right now, if you check the semantics, and 
> seqvars are actually a kind of schema for talking about tuples of 
> arbitrary length by implicitly quantifying over all finite tuples at 
> once. The key point here is that 'finite' which we kind of cheat with, 
> since we say its finite since its all in the syntax; but the seqvars 
> are using this to define a range of a *quantifier*. Hence all the 
> non-FO complexity. OK, so lets stop cheating and get explicit about 
> where that 'finite' comes from, which is a recursion on the length of 
> the tuples: you can add another widget to a finite tuple and it stays 
> finite, and we take the standard fixedpoint of the obvious recursion. 
> We don't even use the concept of 'finite' in the metatheory anywhere. 
> Tuples are then something awfully like LISP lists (cf. KIF) and RDF 
> collections (cf. OWL).

Not sure I understand all of the implications, but this sounds right & 
reasonable.  I will reply later with more questions as I assimilate 
what you wrote.

   .bill

--
Bill Andersen
Chief Scientist
Ontology Works, Inc.
1132 Annapolis Road, Suite 104
Odenton, Maryland 21113
United States
Office: 410-674-7600
Mobile: 443-858-6444