infinite precision (was Re: Against PEP 240)

[Alex Martelli]

<aahz at panix.com> wrote in message
news:200106011658.MAA03564 at panix3.panix.com...
> [posted & e-mailed]
>
> First of all, let me apologize.  My original comment was intended to be
> teasing, and I'm sorry this has descended into heavy-handedness.

Please let me apologize in return for my own contribution to the
heavy-ing (now THAT is a doubtful word indeed!-) of the exhange.

Originally I had thought it just funny that a widely used dictionary
happened to use identical words in its first definitions for both
"unbounded" and "infinite" (reflecting the words' history, btw:
"infinite" means, exactly and literally, "without boundaries" in
Latin:-), but when it turned to "standard definitions used in
CS" it started to look as if readers might get the wrong idea
about what (at least several) computer scientists actually use.

Every reader of this newsgroup knows I make enough mistakes
that I can't _afford_ to get in a huff just because somebody
calls me on one... indeed, I'm grateful!  But exactly because
I make so many true mistakes, I'd rather not be considered
in error in one of the cases where [I think] I'm not:-).


> program, but I can't argue with the weight of publications you found.
> It still doesn't feel right to me, though.

I don't claim it's "right".  It MIGHT be useful to have different
terms to distinguish (at least for non-integral numbers, I
guess) "really infinite" cases (constructive reals, I guess) from
the common "without _predefined_ limits" current usage of
"infinite" (or "unbounded") "precision".  But then, who ever
thought natural language was based on what is "right"?
(Ah, NOW we can have a really heated debate on descriptive
versus prescriptive schools of natural language study...:-).


Alex