[Zope3-dev] Re: What does python 3000 mean for zope?

[Philipp von Weitershausen]

David Pratt wrote:
> Hi. I am concerned about the announcement of python 3000 today that will 
> break backwards compatibility. Zope and twisted are my favorite 
> frameworks. The code base for both frameworks are not small. I haven't 
> evaluated the changes but I can say this is a not great day for the 
> python community either. I can see this dividing folks between present 
> and future.
> 
> Particularly, I'm thinking about incompatibilities developing around 
> packages and dependencies through some sort of drawn out transition by 
> the python community that may take years. Has anyone thoughts or 
> comments about python 3000 implications for zope? Unfortunately, my 
> first thoughts are that Python 3000 feels like Y2K for python :-(. Many 
> thanks.

We're currently struggling to get to Python 2.5 (which isn't exactly 
fresh out of the oven) mostly due to incompatibilities that it 
introduced compared to Python 2.4. So when Guido says Py3k will allow 
incompatible changes for the "first time", it'll be hard to imagine how 
big the implications really are. It's especially hard to imagine because 
Py3k isn't done yet. Will the stdlib be reorganized? Who knows. I sure 
would like to see this '2to3' tool tackle the Zope codebase. C 
extensions, anyone?

Zope is depending on a lot of third party code. This is a good thing, it 
helps us focusing on what do best: write web apps. But before those 
dependencies aren't ported yet, we might just as well sit back and enjoy 
Python 2.x. The flip side of the coin is, however, that with our many 
individual libraries, it could be easier to tackle (provided we manage 
to down on those interdependencies).


I'm not saying Zope never will or should or can run on Py3k, but I trust 
it'll be on Python 2.x for a long time still. And that can't be too bad, 
it has been for many years now after all.


-- 
http://worldcookery.com -- Professional Zope documentation and training