[SciPy-dev] Timetable for 0.5.3 release?

Bill Baxter wbaxter at gmail.com
Mon May 28 17:30:43 EDT 2007


On 5/29/07, Travis Oliphant <oliphant.travis at ieee.org> wrote:
> David Cournapeau wrote:
> >> Perhaps we should move to a timetable-based release schedule for scipy?
> >> Every three months, release the current svn version. Also making a
> >> release after a numpy release is a good idea, so that the current
> >> versions work with each other.
> >>
> > As the scipy community seems to be growing, and as Travis wanted to have
> > a release manager for scipy, what about adopting a scheme similar to
> > bzr, which seems to work fine for them: having a different release
> > manager for each release ? Not that this is against having a timetable,
> >
>
> Here, Here.  The releases are slow in coming only because it seems to be
> entirely relying on my finding time for them.
>
> I would like to see more code move from the sandbox into the scipy
> namespace.  I am currently working on the interpolation module to
> enhance the number of ways in which you can do interpolation using
> B-splines (the basic functionality is in fitpack, but sometimes I can't
> make sense of what it is doing with the knot points --- there is a lot
> less flexibility then there could be).

I would like to hear comments on a possible change to a 3-stage
deployment of sandbox modules:

Stage I - "pre-Alpha" like current sandbox - module not distributed
with scipy builds.  Have to build from source if you really want to
try it.  At this stage it's probably just one or two folks tinkering
around.  Module may or may not work.

Stage II - "Alpha" - publicly release prebuilt module, but still in
sandbox namespace. At this stage module seems to work, for the authors
at least.

Stage III - "Beta" - move module out of sandbox namespace.  Module API
has settled down and most likely there won't be any more significant
(incompatible) changes to it.

The middle stage is the new thing.  Make a sandbox module visible for
a while first.  Then move it out after people have had a release or
two to hammer on it.

--bb



More information about the SciPy-Dev mailing list