[Pydotorg-redesign] Re: Moving python.org to Zope

[Aahz]

[switching to pydotorg-redesign]

On Thu, Aug 07, 2003, Trevor Toenjes wrote:
> Aahz:
>> 
>> In addition, it's my belief (based on conversations with a number of
>> expert Zope people, including my cousin (Michael Bernstein, co-author of
>> _The Zope Bible_))
> 
> no Way! Michael-webmaven Bernstein is your cousin?  I guess you got all the
> good looks in the family.  :)

<smirk>  Don't let him hear that -- he's the vain one.

>>  that Zope3 represents an quantum leap forward in
>> Zope's power *and* usability over Zope2.  But I'm adamantly against
>> converting to Zope3 until it goes production, and I think the pain of
>> doing a double conversion to Zope2 followed by Zope3 just isn't worth it.
>>
>> Can you provide any convincing arguments against my position?
> 
> hmmm. I agree with your concerns and the opinions you stated.
> I agree that Zope3 will be a quantum leap.  But it is a ways off.
> 
> But the final release of Zope3 is expected to have a reasonable
> migration path for up to date Zope2 code.  I dont think python.org is
> going to have any challenging objects that would break, which I think
> is the technical concern.  It's mostly text/HTML based documentation
> which will be trivial to convert.  My understanding is the conversion
> issues with Zope2 to Zope3 are about deeper custom-Zope objects
> called Products and External methods in Zope2, which I dont suspect
> are really needed for python.org.  I think if we stay with stock Zope
> objects, we would be ok.  I could be wrong....

That's what I'm concerned about.  The only iron-clad reason for using
Zope as opposed to some other system is Zope's support for full-blown
CMS.  I've been assuming that this will require extensive customization
to use appropriately, and that migrating Zope2 to Zope3 will require a
fair amount of work.  Less work than the original Zope2 converstion,
probably, and even likely less work than going directly to Zope3 from
the current site.  

But I'm still wondering just how much work the double conversion is
likely to be, given that the initial Zope2 conversion almost certainly
will take more work than going straight to Zope3.

> AND Zope.org is finalizing a long painful migration, and they went
> with Zope2/CMF+Plone.  If it's good enough for the Zope community, I
> think that says something.

Not really.  After all, these are people who already know and love Zope.
Perhaps you're right that some form of test/case-study would be
appropriate, but evaluating it is going to take a lot of work, too.

> I only suggested it to make python.org easier to maintain, which it
> would once you got over the hump.  So convert it now, and get 2 years
> of making life easier, and then look at Zope3 which might have its
> first maintenance release by then.

As I think I've said before, unless it's clear that we can reach the
crossover point in productivity in less than a month, I don't think we
can afford it.

> And a zope generated site can still be mirrored, because it can be
> designed as is...very static.  It's just that now you will have easier
> publishing and delegation of maintenance responsibilities.

That I know about.  Essentially, we'd have a separate server that runs
Zope and it'd push to the current server.
-- 
Aahz (aahz at pythoncraft.com)           <*>         http://www.pythoncraft.com/

This is Python.  We don't care much about theory, except where it intersects 
with useful practice.  --Aahz