Re: Identifying reals

From: Mark Baker <distobj@acm.org>
Subject: Re: Identifying reals
Date: Fri, 4 Jan 2002 14:37:01 -0500 (EST)

> So after all that, what *is* the relationship between "identifiable" and
> "countable".  Are they equivalent (it would seem so)?  

This depends on your processing model.  Some would argue that humans (in
particular mathematicians) can do better than ``countable''.  Some would
argue that, because the universe is finite, nothing can do better than a
finite state automata.

> It would just be
> interesting (and maybe useful at some point) to understand some of the
> limits of a system based on identity, such as the Web.

Well, in some sense, it is easy.  You pick the appropriate computational
model, either finite state automata or Turing machines, and proceed from
there.  In another sense, it is philosophy.  You pick your favourite
philosopher, and, depending on how you view philosophy, you either agree
with everything he/she says or you disagree with everything he/she says.

There are lots of other confounding factors, of course.  Do you want to be
able to understand the system?  Do you want to declaratively describe the
system?  Do you want the system to be realizable?  Do you want ....?

In the absence of real answers, we are more-or-less stuck with trying
something that we know how to do, getting some mileage out of that, seeing
where it breaks down, and trying to do better.  This is how engineering
works, after all.  The ``trying to do better'' usually involves some
mathematics or computer science, which may or may not be tied to the
details of the problem.

The above breaks down in the presence of hype.  Inflated claims result in
people thinking that unsolved problems are really solved.  Insufficient
effort is then made to solve the unsolved problems, resulting in lack of
progress. 

> MB

peter

Received on Friday, 4 January 2002 15:27:01 UTC