Re: Revised Axioms for Cardinality Restrictions

> I still don't think you need to change the definition of
> NoRepeatsList to have a second argument, the length of the list.

I agree.  Your proposal to have an axiom saying that if ?r is a
NoRepeatsList and ?x is not an element of ?r then "(cons ?x ?r)" is a
NoRepeatsList is a more elegant way of taking care of the list existence
problem.  In KIF, that axiom would be:

  (=> (and (NoRepeatsList ?r) (not (item ?r ?x)))
      (NoRepeatsList (cons ?x ?r)))

Richard

Received on Friday, 14 September 2001 17:44:40 UTC