[Zope] [Newbie] Source configuration management with zope

[Gérard Oogle]

Hi list,


I recently discovered Zope, took the tutorial, read a couple of things,
and I am now in the process of writing my first little app.
So far there's something that is bugging me because I can't relate it to
my previous habits.
Ant it relates to the configuration management of what the developper I
am produces.


In my life before Zope:

    * I was developing in a sandbox composed of:
          o a "source sandbox" that was the mirror of a -cvs- repository
          o data sets for test purpose (files, database, ...)
    * There was a set of tools -compiler, sqlj, ...- that were able to
      give life to the "code" part of that sandbox by producing
      something that could be deployed
    * Sometimes there was another set of tools -ejbc, pl/sql script,
      ...- that were able to deploy the preceeding same artifacts into
      *different* environments
      The main ones beign
          o The QA environment
          o Real live production system
    * When we, developers, were 'creative'/'careless', we would release
      in supplement of the usual deployable artifacts 'migration scripts'

You've probably recognized the regular Java/J2EE mantras.
It is a little bit far now but I think I remember that things were not
*that* different in Smalltalk:

    * I was developing in a sandbox composed of:
          o a "source sandbox" that was an image of an Envy repository
          o data sets for test purpose: it happened that our data were
            also stored in an Envy repository but it was an instance
            different from the one in which we were maintaining our code.
    * There was a set of tools -packager/configuration manager (my
      memories vanish)- that were allowing us to deploy the resulting
      same artifacts into *different* environments
    * I never had to write 'migration scripts' between 2 Envy repository


If I mention Envy, it is because when I try to understand Zodb, I can't
help making parallels between those 2 "stores".
What puzzles me is that I confusely have the impression that:

    * when I am developing my little prototype, the code and the data it
      manipulates end up in the same Data.fs
    * I don't know where to code my second iteration of this prototype
      as soon as I will have shown it to business expert I working with
      and that he will start populating it with his data:
          o should I use Zope version management features and work on
            the *same* Data.fs than my business guy?
          o Does that mean that I cannot work locally on my machine and
            that I have to be connected to the server us 2 guys have
            access to?
          o if so, how can I tag each versions or better how can I
            package these versions to be able to deploy them in
            different Data.fs?

To put it in a nutschell, I guess I am wondering how logic and data can
be separated in Zope.
I guess also that I am not the first dinosaur the zope community is
welcoming.
I would be grateful if someone could:

    * either shoot me a couple of links explaining how to do that
    * or help me refactor my views on the world (and I guess that a
      couple of links would do it also).


Last thing, in that same frame of mind, since zope.org itself was
recently upgraded, could somebody please detail us the various steps
followed for that migration?

Thanks in advance,

-- 
Gérard