numpy, overflow, inf, ieee, and rich comparison

[Kevin Jacobs]

Tim Peters <tim_one at email.msn.com> wrote:
> So, by the time you guys stop stabbing each other in the back <0.9 wink>,
> perhaps C99 will be widely implemented enough that Python can actually do
> something useful for you here.


Most platforms that I use have come a long way to supporting correct IEEE
arithmetic.  The major problems I've had come with libraries, third-party
compilers and interpreters (Python!) that try too hard to correct for past
sins.  Why not make a "IEEE-clean" option for Python 2.1 that strips out all
the vestigial numerical helper code and adds standard methods for getting
and testing for +-inf, nan, and floating point classes for platforms that
are close enough to C99?  I know it will require a bunch of compile-time
tweaks for every platform, but they're generally easy to write and test.  Of
course, we won't be able to control the quality of the underlying C library,
but lets focus on fixing what we can.

Thoughts?

-Kevin

-- 
----------->  Kevin Jacobs  <-----------|------->  (216) 778-8211  <--------
Informatics Consultant                  | Department of Epidemiology
Primary mail:   jacobs at theopalgroup.com |   & Biostatistics
Alternate mail: jacobs at darwin.cwru.edu  | Case Western Reserve University
----------------------------------------------------------------------------