[Zope-dev] The bleak Future of Zope?

[Stephan Richter]

On Wednesday 21 April 2004 03:58, Martin Kretschmar wrote:
> Maik Jablonski of the german speaking Zope Users Group
> DZUG issued a pretty bleak outlook for the future of
> Zope. What are your oppinions?

To not make the previous mail too long, here my general opinion.

1. Maik likes to do things the quick and dirty way. See Epoz and Mailboxer.

That works well for small and personal projects, but is not the answer for 
large projects. If Zope 2 or 3 would have been built this way, they would 
have already fallen apart. Abstract thinking is a required for framework 
development.

Epoz has been totally redesigned (Kupu) in a more abstract way and works very 
well for end users in Silva...and it is easily adjustable and extensible. For 
Mailboxer I can only say that he should have leveraged the development power 
behind Mailman and develop a nice UI on top of it as I had demonstrated with 
some code a year earlier. This suggests to me he is either (1) not a team 
player or (2) technically not good enough to integrate. It is much, much 
harder to play nice with other projects than starting your own. I have done 
this mistake myself often enough (back then I was not technically good 
enough ;-).

2. Maik is is frustrated with the releases of both Zope 2 and Zope 3, 
including their merging.

First off, I do not develop Zope 2 and I am not involved there, so I have no 
qualified opinion. However, it is always easy to complain about ZC and push 
all the responsibility to them. I bet you that ZC would allow a 3rd party to 
do releases, if they show interest, knowledge and wisdom. However, people 
just keep complaining and do nothing. 

The situation is even more obvious with the Zope book. All the community has 
to do is to give a particular part/chapter/section to a couple of people for 
maintenance. But oh wait, that would need someone to manage this effort and 
*that* would be just too much work.

For Zope 3 however, I can give a very well-informed opinion. Philipp privately 
pointed out to me that people exected Zope 3 technologies to arrive earlier 
in Zope 2, such as the CA and principals maybe. This was not desirable in 
several ways. First, the API was not stable and Zope 2 as a mature software 
would have suffered from the ever changing API. Next, there was still a lot 
of restructuring going on that would have caused interruptions in Zope 2. 
Third, none of the code was optimized and dog slow, nothing someone wanted to 
use for a large site. Finally, we just had no bandwidth for it! Who was to 
support the Zope 3 in Zope 2? At the end it would have been Jim and it 
distract him from finishing Zope 3.

Concerning the release schedule, ZC has little to do with that for Zope 3. In 
fact, I have been release manager since this summer and I am responsible for 
the release schedule and packages. However, I decided not to release often, 
since again we do not have bandwidth to support the milestones. Since the CVS 
is as stable as any milestone release (we have tests for everything), 
releases are less important and it is much easier and less time consuming to 
support the current HEAD, which you can just download via the Web. However, 
we are getting the first alpha out by the end of the month. Hopefully, by end 
of May we will have finished the X3.0 to-do list and will release the beta. 
At this point the API will freeze and application developers are encouraged 
to have look at it.

I have more to say, but I the E-mail would become too long. Overall, I think 
Maik's predictions and scepticism is fairly uninformed from a Zope 3 
perspective. He has never seriously participated in writing 
code/documentation and/or contributing to discussions.

Regards,
Stephan
-- 
Stephan Richter
CBU Physics & Chemistry (B.S.) / Tufts Physics (Ph.D. student)
Web2k - Web Software Design, Development and Training