[Zope] Philosophical question - Zope Objects vs Entities in a RDBMS

[Rowan Hick]

Okay so I'm getting into this whole Zope deal, and from what I have seen
so far it looks pretty darned good, I like the OO concept, but with it
it poses a few design issues for websites / web applications

Traditionally all data is stored in some form of relational database,
and then the scripts pull it back out again. Now we all know and hate
relational databases (well I do anyway as I'm a child of the information
age - learning in OO concepts and then having to go back to a relational
database) but they are remarkably good for what they do - storing large
amounts of data and allowing you to retrieve it in a quick and
relatively painless manner, if you get your head around SQL.

Now Zope comes along with the ability to create and persist objects
without the need for a relational database. Wooohooo... all my prayers
have been answered, or have they?.

Q1. Has anyone tried to build webapplications with large (maybe into the
10's of thousands) numbers of Zope objects?
Q2. How much more or less efficient is Zope for storing a piece of data
over a traditional relational database ?
Q3 (I'm a newbie here) Can a python external method access Zope objects
?  Eg if I create a task object from a ZClass can I then do things with
it from a python script?
Q4. I think I remember seeing a post about the Zope database being a
single large file posing problems with the Linux filesystem.. or was
that the windows filesystem.. if so then that could limit Zope's
potential in a commercial production enviroment, surely ?? If this is
true what have people done to try and get around this, ie archiving
objects (however that could occur) etc.. ?

Thanks in advance for any responses,


Rowan Hick

Business Analyst
Soft Tech NZ LTD,
Auckland, New Zealand.