Against PEP 240

[David C. Ullrich]

On Thu, 31 May 2001 02:02:43 -0400, "Tim Peters"
<tim_one at email.msn.com> wrote:

>[Paul Prescod]
>> We're talking about Python. I don't think it will accurately represent
>> irrational numbers in my lifetime!
>
>Jurjen N.E. Bos contributed a Python implementation of the constructive
>reals several years ago.  In effect, that allows working with any computable
>real number (from 42 to e.g. tan(exp(log(pi)+sqrt(e))) in such a way that no
>finite amount of computation can distiguish it from its infinitely-precise
>value.

Really? You don't happen to recall where I can find this, do you?

(Um: or are you referring to the celebrated real.py? At least in the
version of real.py I have the reals have arbitrary but predetermined
finite precision. If that's what you meant never mind. It sounds
like you're talking about actual "infinite precision" lazy-decimal
reals - if that's what you mean I'd appreciate knowing where I
can find a copy. 'Pologies for asking you instead of him, but he's
a hard guy to find...)

>But the general truth is that no finite representation of reals is without
>surprises, and under the computable reals, e.g., equality is undecidable!

Those wacky reals. Makes you wonder how they got the name.


David C. Ullrich
*********************
"Sometimes you can have access violations all the 
time and the program still works." (Michael Caracena, 
comp.lang.pascal.delphi.misc 5/1/01)