Time and again people fail to realize that Zope 3 wants to create
the low level framework right first, and only after that add high
level simplifications and shortcuts to have less configuration and
provide fastest developer exeperience.
Of course people aren't attracted to Zope 3 yet because they feel
it's complex. And there's a reason: it *is* complex. For now.
So you don't think anyone but hard core Pythonistas should use Zope 3
Does anyone think that code generation in the form of a "starter
package" could help here?
It would be a great experiment. Zope 3 currently inherits the Zope 2
tradition of starting with a full application, including a management
UI, and developers add to that existing application. It would be
valuable to explore an approach where developers start with a simple
application (with no management UI) that they are expected to replace
rather expand. I expect that developers would have a great experience
initially that way.
Most web frameworks have developers start with a minimal, replaceable
application, but application developers write code in ways that hinder
integration with other web applications even though they use the same
framework. The theory behind Zope's configuration system is that
application developers *will* be able to integrate after all, at least
for a longer time than they can now. A starter package generator would
test that theory.
Zope3-users mailing list