print longs
Tim Peters
tim_one at email.msn.com
Wed Oct 6 01:02:58 EDT 1999
[Tim]
> ...
> These "different rules for different numbers", like tagging long
> ints with "L", don't make much sense to people.
[William "Billy" Tanksley]
> I'd like that too. Only keep floats out of it if possible -- Lisp had
> the right idea there. I think 1/2 should return 1/2 unless otherwise
> coerced.
>
> ABC worked that way, IIRC.
Yes, and that's why Python doesn't <0.1 wink>. A simple numeric loop that
doubles its storage and processing requirements on each iteration is a
common outcome of using unbounded rationals under the covers. That's not
user-friendly either; rationals are as much a tool for experts as
floating-point *should* be, but for different reasons.
The only scheme I've seen that "real people" pick up at once is REXX's --
because Cowlishaw deliberately implemented an arithmetic that mimics what
people do with pencil and paper (fixed-but-unbounded precision decimal
floating point, with a massive exponent range). In a lucky break, despite
that people no longer know how to use pencils, much the same scheme is used
in pocket calculators.
I think Python2 is going to need a lot of ways to spell numeric literals
<wink>.
> 'bots have sisters?
Humanist.
all-god's-children-have-sisters-ly y'rs - tim
More information about the Python-list
mailing list