[SciPy-Dev] scipy.stats

[Ralf Gommers]

On Mon, May 31, 2010 at 11:14 PM, <josef.pktd at gmail.com> wrote:

>
> My approach has been to block any untested, non-trivial changes, and
> to review and test changes by others, so we don't increase the amount
> of "technological debt".
>

This is an approach I would like to see, not only for stats but for all of
numpy and scipy. And with the enthusiasm about moving to github it can be
done easily and quickly.

I hope you decide to stick around Josef.

Cheers,
Ralf


> The second part is testing and verification of existing functions
> (finish the statistical review which has been lingering for several
> years), which is slow going because I'm almost the only one doing it.
>
> The third part is develop enhancements and prepare for a consistent
> refactoring of "statistics in python" in a sandbox outside of scipy.
> (Here, I admit I have a small conflict of interest getting things
> quickly into scipy, because I'm more interested in "Statistics in
> Python" than statistics in scipy.)
>
> I have been a bit slow working on scipy trunk in the last year because
> I worked more on statsmodels, and I didn't have always the time (or
> didn't feel like) keeping up with recompiling numpy/scipy.
>
> (My alternative is to develop pythonstats completely outside of
> scipy.stats, maintaining a smaller set of verified functions and just
> ignore scipy.stats for now.)
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/scipy-dev/attachments/20100531/8a27fd2a/attachment.html>