Floating point errors?
Grant Edwards
grante at visi.com
Sat May 12 10:50:58 EDT 2001
On Fri, 11 May 2001 23:16:21 GMT, Nick Perkins <nperkins7 at home.com> wrote:
>>Since 0% of the possible real numbers are representable in IEEE
>>754, then it's obviously useless and all floating-point support
>>should be dropped from Python.
>
>Same goes for integers, 0% of them are representable,
>so let's get rid of them too.
At least you can pick a range of MININT to MAXINT where a
significant percentage (one would hope for 100%) of integers
can be represented. Even within the range MINDOUBLE to
MAXDOUBLE, 0% of the real numbers can be represented. It's a
sad state of affairs. Analog computers are the only way to go
for non-discrete problems -- but then there's the whole problem
with the thermal noise floor.
--
Grant Edwards grante Yow! I'm EMOTIONAL
at now because I have
visi.com MERCHANDISING CLOUT!!
More information about the Python-list
mailing list