Re: [KIF] Re: SUO: tuples
Thanks for your clear comments, Mark.
On Wed, Jul 10, 2002 at 03:49:25AM -0700, Mark Stickel wrote:
> I also have reservations about sequence variables.
> ...
> While sequence variables may be useful and the proposed treatment of
> them elegant, they do have some drawbacks. Sequence variables are
> largely intended for metalogical reasoning,
But isn't that the point, Mark? Given that the scope of CL is broader
than automated reasoning, isn't the appropriate response to your
observations simply to carve out an appropriate sublanguage of CL that
doesn't contain sequence variables, or at most only allows them in tail
position?
> but Prolog's use of lists appears to be adequate in practice for much
> metalogical reasoning.
Maybe so, but this raises the earlier objection that a theory of
sequences doesn't belong in a logic standard.
Also, I don't recall the semantics of Prolog very well, but isn't the
domain in an interpretation a fixed-point wherein everything is an
n-tuple? I really don't think that is what we want as the basic notion
of an intepretation, given the goals of CL.
-chris