[Numpy-discussion] Going toward time-based release ?

[David Cournapeau]

On Sun, 2008-05-11 at 21:26 -0700, Jarrod Millman wrote:
> 
> The one caveat to this is that you may recall I have tried to start
> having a 3 month release cycle ever since I took over release
> management last summer.  I was almost able to do it for the first
> release of NumPy and SciPy.  But since November of last year, I have
> been struggling to get out NumPy 1.1, which was originally scheduled
> for early February.

My impression, but maybe I am missing something, is that the release
slipped because everybody added new code. If we have a strict policy to
say no new code, only bug fixes at least for 2 weeks before a release
(and even no changes at all for a period), the release process becomes
much easier, no ?

IMHO, the main advantage of time-based release is to be able to say no
to new code just before release, and to be able to say that an api
breaks cannot happen between two releases: any change needs at least N
releases in between with warnings, and we know what N means because N *
release period is the time you have to make changes if you want to stay
up to date.

I personally really did not like what happened with ma and matrix in the
1.1, and I would like to avoid this in the future. It is already pretty
bad to break code in a minor release, but to do it without warnings is
really something which should never happen again IMHO.

cheers,

David