[Distutils] Attempted summary of version number thread
Sat, 12 Dec 1998 12:01:24 +1100
> So: do any of you object to the above characterizations? I'd be
> interested to hear the opinions of anyone outside the above six,
> especially if you feel there're more sides to this debate
Well, you asked :-)
Im definately on the side of the Anarchists. IMO, the "deliverables"
from this SIG should be unobtrusive as possible, and should conform to
_my_ wishes, rather than the other way around.
For example, my Windows extensions use a simple "build number", and
this is what I want to continue to use. My reasoning is simply that I
make many incremental releases - its not a product per se. There is
never a clear distinction between "versions" - what would "version 1"
of my extensions be, or "version 2"? Further, it is never clear to me
when my releases are "beta" - new features may not work completely as
advertised, but core features are typically more stable than previous
versions. This is almost certainly an artifcat of the volume of
things in my single package - but that is the point - the way _I_ do
things dont necessarily fit with the other version branding models
Of course, other things I have in mind _do_ fit better with some of
the proposed schemes. So one size doesnt fit all, even just for me!
If the deliverables require me to rework too many things, or shoehorn
into a scheme that doesnt fit for me, I may simply not bother.
Without presuming to speak for Marc and Greg, IMO, the message is
clear - if some people on this SIG dont see it working for them, then
the chances of it becoming truly general purpose to the wider
community are slight.
Just MHO (Mark Hammond's Opinion :-), of course!