# [Python-Dev] stats.py (was 'summing a bunch of numbers ')

**Raymond Hettinger
**
python@rcn.com

*Mon, 21 Apr 2003 20:29:20 -0400*

[GvR]
>* > Since it already exists as a 3rd party package, we should definitely
*>* > not try to duplicate the effort. Then the question is, is it enough
*>* > to point to the 3rd party package or does it deserve to be
*>* > incorporated into the core? We can't go and incorporate every useful
*>* > 3rd party package into the core
*
Why not? From a users point of view, that is the best place for it.
Of course, not *every* useful third-party package is a candidate,
but if it applies to several different categories of users, then maybe.
For instance, the DNA seach packages are somewhat tightly targeted,
but basic statistics come up in many different types of work.
[Chad]
>* True. I just happen to be of the opinion that a statistics package is
*>* the single most practical and useful mathematics pakage that can be
*>* added to a language, after basic linear algebra (which isn't in the
*>* core... Hmmm)
*
I've maintained a pure python linear algebra package for several years.
It gives the basics plus QR decomposition, complex matrices, and eigenvalues.
Still, I haven't felt the slightest need to request that it be put it the core,
nor have any of my users requested it.
>* I would actually favor keeping it fairly lightweight
*>* (although t-tests and even ANOVA should go in).
*>* Heavyweight users could
*>* always download a separate add on package.
*
To keep it lightweight, it should be kept in pure python.
Heavyweight users can download the binaries when needed.
reinventing-the-wheel-is-fun-educational-and-non-productive-ly yours,
Raymond Hettinger