[Committers] my platform

[Martijn Faassen]

Rocky Burt wrote:
> On Tue, 2006-13-06 at 19:18 +0200, Martijn Faassen wrote:
>> Zope 2 and Zope 3
>> =================
>>
>> We've had big discussions over how to manage the whole Zope 2 and Zope 3 
>> story, with suggestions for renaming the platforms, the way we present 
>> the platform, etc. This has strong ties to developer marketing, but is 
>> also broader than that. We need to be careful we don't confuse people 
>> and the road we're on now with Five has served us quite well in finding 
>> direction.
>>
>> That said, a course change that involves some rebranding seems required 
>> at some point. We need to carefully manage how we do this, given the 
>> limited resources we have.
> 
> Dunno if any of this was intended to be replied to, but I will anyhow ;)
> 
> Of course all of the items you mention interest me to some extent, but
> your comments on Zope 2 and Zope 3 interest me the most.  What I would
> like to know is if any (all) of the candidates have ideas for where they
> would like to see the Zope 2 and Zope 3 convergence happen (of if they
> see it happening at all).

I don't really know the complete answer to that yet. It's a combination 
of where we set the bar for convergence and how many resources we think 
we have to do it. The trick is to make it in people's own interests to 
work towards this conversion - by offering new features and better way 
to do things in Zope 2, and by allowing people interested in Zope 3 to 
benefit from the larger Zope 2 community.

I think we should go for a system that could run both Zope 2 
applications as well Zope 3 applications. We have competing pressures. 
On the one hand, people who want to run pure Zope 3 applications should 
not be bothered by Zope 2-isms and Zope 2 code. On the other hand, to 
ensure a smooth upgrade from Zope 2 we have to start with changing Zope 
2 code step by step - reverse fitting Zope 3 to run Zope 2 applications 
seems much harder. The Five project is doing pretty well - we are 
enabling Zope 3 technologies to work in Zope 2, and we're replacing Zope 
2's plumbing step by step by Zope 3 plumbing.

The unified platform is one goal. It would allow us to have a single 
paltform called Zope again and not to have to explain the details of 
Zope 2 and Zope 3 too much, at least to the outside world. This is 
valuable.

The other goal is the convergence of actual Zope *applications* towards 
this unified platform. This means a step by step refactoring of the 
applications to use more Zope 3 technologies. We've been on that route 
for a while now with Five, but it's clear that some hurdles are 
incredibly difficult to take - porting any of the big CMSes existing 
views to use Zope 3 views for instance. I'm still struggling for 
patterns to do this step by step too.

If we keep up the rate Zope 2 has been improving so far, we may end up 
with a Zope 2 platform that can run Zope 3 applications somewhere in the 
next year. We should be careful in the way we present this to the world. 
Not too soon before we're ready.

To developers we have to make clear that it's up to *them* to make this 
happen - it shouldn't be that we say: "we're going to do this" and 
everybody sits back and it never happens. That's the risk of roadmaps in 
open source. I also appreciate the vocabulary we have now (Zope 3, Zope 
2, Five). It's fairly well understood by the core developers, we use 
this to discuss this stuff, and we should take care not to lose that 
vocabulary too soon.

I'm not sure whether such a unified platform would make pure Zope 3 
developers happy yet. Maybe that's further of. It should be possible to 
make a profile of the unified platform that contains no Zope 2 code 
(that hasn't been ported to Zope 3) - more or less equivalent to Zope 3 
as it is now.

As to convergence for the applications - much further of. We have to 
consider strategies to make that happen, or strategies to live with the 
fact that it won't ever completely happen, or at least strategies to 
bridge the gap of years that might be ahead.

Sorry to be so vague. :)

Regards,

Martijn