Let me play devil's advocate for just a moment Let me run around for a while like a headless chicken. Please don't shoot me, I'm actually quite in love with Zope 3. But I can see myself having this argument many times in the future, so I'd like some good answers.

As a developer, I have a lot of respect for the fearlessness with which you refactor the software. Seriously fundamental things, like ZCML and the ZODB and ZAPI are up for discussion just as freely as simpler parts. Perhaps that is just a testament to how flexible and powerful Zope 3 is and how easily interchangeable its components are.

But it also scares the crap out of me. I read a whole book about Zope 3 that came with two chapters marked as deprecated by the time it hit the printing presses. I read what scant documentation there is (apart from the excellent interface documentation and internal documentation like doctests) and see that each tutorial makes fundamentally different assumptions. You have an admirably well-controlled process for deprecation, but the cycle is what - a year? By the time I've finished my application, I'll have a migration headache if I want to take advantage of the latest and greatest features or even non-backported bug fixes; by the time I've made my cool new add-on module, the API it depends upon spews out deprecation warnings.

My question is: will it ever slow down? Are you trying to evolve the most aesthetically perfect framework or the most useful and reliable one? How do people that have taken the plunge and developed on Zope 3 deal with these changes? Should I really view it as one evolving framework, or as a series of partially disconnected attempts and abandon hope for forward compatability?

Now: Please prove me all wrong. :-)



Zope3-dev mailing list
Unsub: http://mail.zope.org/mailman/options/zope3-dev/archive%40mail-archive.com

Reply via email to