[Numpy-discussion] What is consensus anyway

Charles R Harris charlesr.harris@gmail....
Mon Apr 23 18:39:43 CDT 2012

On Mon, Apr 23, 2012 at 5:02 PM, Travis Oliphant <travis@continuum.io>wrote:

> That is an excellent thought.
> We could make the odd numbered releases "experimental" and the
> even-numbered as stable.
> That makes some sense.    What do others think?
I'm starting to think that a fork might be the best solution to the present
problem. There is plenty of precedent for forks in FOSS, for example GCC,
EGCS, Redhat 1.97, LLVM and emacs, xemacs. There are several semi-official
forks of linux (Android, the real time Kernel, etc.) Zeromq just forked,
OpenOffice forked, there was XFree86 forked to Xorg, etc. Linus encourages
forks, so there is even authority for that ;) Of course, the further the
fork diverges from the original the harder reintegration becomes, witness
Android and wake-locks. But a fork would cure a lot of contention.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.scipy.org/pipermail/numpy-discussion/attachments/20120423/fe09ede2/attachment.html 

More information about the NumPy-Discussion mailing list