On 4/25/07, Nadav Har'El <[EMAIL PROTECTED]> wrote:
Uri, this is becoming (or was always) extremely off-topic.

Well, it is off topic.

Georg Cantor's
beautiful proof that there are more real numbers than natural numbers
(involving the diagonal of the real number's list) is one of the most
striking - and self-evident - proofs that I've ever seen (my father showed
it to me when I was a kid). Wikipedia (which I'm glad you're using as a
source - I thought you thought it was the devil :-)) also has an article
about this: http://en.wikipedia.org/wiki/Cantor%27s_diagonal_argument

I read this article before writing you.  Look what it says about the
constructivist approach.  Whether or not there are "more" real numbers
than rational numbers, depends on how you define "more".  Every
language contains ambiguities.  Proving that there are things which
can't be expressed by any specific language is in itself inconsistent.
For example, is "the smallest natural number that can't be expressed
in less than 20 words" a number?  Is there no such number?

Anyway, proving that a specific problem is "hard" for any algorithm
means something like "we can express more problems than we can express
solutions in the language we are using".  I'm arguing that if a
function is proved to be not computable, then it's not well defined.
The definition is not completely deterministic.

I changed my mind about Wikipedia.  It's a very good source for
information on many subjects, and I use it a lot (especially the
English version of it).  But there are some cases where it is
controversial, and I hardly ever write there at all.  If I write
anything, in most cases it will be deleted.  So it is both good and
bad, both one and zero, both god and the devil at the same time.  I'm
allowing myself to be inconsistent.

Uri.

=================================================================
To unsubscribe, send mail to [EMAIL PROTECTED] with
the word "unsubscribe" in the message body, e.g., run the command
echo unsubscribe | mail [EMAIL PROTECTED]

Reply via email to