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

[Gael Varoquaux]

On Tue, Feb 24, 2009 at 01:48:47PM -0800, Mike Hansen wrote:
> It doesn't seem your current process is conducive to have a
> "maintainer" for each package.  What is the maintainer supposed to do?
>  Monitor all the relevant SVN commits hoping that some broken,
> untested change doesn't go in behind his or her back?

> With a review process, maintainers emerge since code doesn't get
> included if they don't.  You also get a lot more people knowledgeable
> about more areas of the codebase.

The problem is simply that you are lacking people. No more, no less. I
say you because my contribution to scipy code has been nothing, although
I try to be supportive of the project.

I wonder how we go ahead and fix this problem. Tough question. What
puzzles me is that there are plenty of people writing numerical code. In
external packages, or in their in-house libraries. Why are these people
not interested in putting their effort in something bigger, and thus
probably longer-lived, puzzles me.

Bah,....

Gaël