design methodology question

[Jaco Smuts]

Hello there.

short version of my question:
I am looking for an article discussing methods used to code a layer 
between my rdbms and my application.


longer version:
I've just completed an application using wxPython and pysqlite (I've 
worked with mysql, sql server, oracle, db2, etc.... before. Opted for 
sqlite out  of curiosity and licensing considerations). As usual at the 
end of the exercise I am highly frustrated with all the code involved in 
maintaining the persistence of my objects in the rdbms I once again went 
down the road of investigating OODB's (including ZODB), and I've once 
again come to the conclusion that though it will speed up my initial 
development, I expect to run into a lot of work once I start writing 
reports, for one. I am not really interested in use an OODB as an answer, 
but if you have sound reasoning to back you up I  will gladly hear what 
you have to  say. I admit my bias, I come from a relational db background, 
and I'm comfortable for now that my data model and my object model 
(problem space/description) are not the same thing.

I've looked at http://modeling.sourceforge.net/ and other similar 
projects, but I'm not yet convinced. Would be interested in your 
experience with these type of bridging frameworks.

What I am really looking for is best practice advice when using a rdbms 
backing an application designed/written using OO concepts. (Something that 
includes or touches on GUI's would be even better).

thanks in advance
jaco







-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20040714/d885baf5/attachment.html>