[Pythonmac-SIG] The versioning question...

has hengist.podd at virgin.net
Mon Dec 27 23:29:02 CET 2004


Chris Barker wrote:

>>That doesn't fix the multiple versions problem.
>
>This is a big issue that the core Pythonistas don't seem to be 
>interested in addressing. It's odd, because I think it's a 
>no-brainer that python modules need to be versioned, and there needs 
>to be a way to have multiple versions co-existing and user (that is 
>code) selectable.

Agree completely. I think the reason folk are reluctant to address it 
isn't that it's an inherently hard problem, but that to solve it now 
requires either 1. hard political decisions to achieve an easy 
technical solution (i.e. a solution that's not backwards-compatible 
with previous versions of Python and would require big changes to the 
way that module-related tools and community operate), or 2. difficult 
and costly technical solutions to obtain a politically expedient 
answer.

Alas, that's weird and wobbly and completely paradoxical world of 
software for you. Most wonderfully pliable medium in the universe, 
yet constricted and bound by the inexorable weight of its own 
history. I suspect the only way to remain free is to be completely 
unsuccessful - at least without users there's nobody to upset but 
yourself. Oh well, maybe in Python 3000... ;)

Cheers,

has

p.s. FWIW, I do remember proposing a versioning system in another 
such conversation - can't recall where, but could probably dig it up 
if you were really interested. I think it was quite a promising 
concept worth further investigation: simple and very flexible to use 
(most schemes are either simple and completely rigid, or supposedly 
flexible and undesireably complex and over-engineered) - sufficient 
to cover the majority of user needs out of the box, with the raw 
underlying APIs still available to anyone who might occasionally need 
to arrange something special.
-- 
http://freespace.virgin.net/hamish.sanderson/


More information about the Pythonmac-SIG mailing list