[SciPy-dev] Scipy workflow (and not tools).

Matthew Brett matthew.brett@gmail....
Wed Feb 25 13:32:47 CST 2009


> Scipy is not in the same group, but I think a review process before
> commit, if it attracts more users, will make it more likely to catch any
> problems. There are many good statistical tools in scipy, however
> until recently I wasn't sure what I would use in a "serious" application
> since there are too many, possibly incorrect results.

Actually, watching your (Josef's) fixes to stats was an eye-opener for
me.  My impression was as you've said, that it was badly enough broken
that you wouldn't use it - before you went through it.

We can't afford to carry on shipping code that is that broken,
otherwise we'll lose momentum and users.   And at least - for stats -
you (Josef) came along and went deep into it.

Doing that is much harder without documentation and tests, making it
more likely that bad code with major bugs will carry on making Scipy
limp in the world of numerical libraries like matlab and R.

I think we do - clearly - have a problem, and that we do - clearly -
need a change to fix it.


More information about the Scipy-dev mailing list