[mod_python] global persistence

On Wed, Jun 25, 2003 at 06:58:45AM -0700, Jonathan Gardner wrote:
> This is equally valid. My problem is that every time I get to start using one 
> of these, it ends up with so much data that I should've gone with the 
> database from the very beginning. If you *know* your data set is never going 
> to get very large, then using files or Berkeley DB is a very, very good idea.

There's no reason not to use either one of those just because your
dataset is going to be large. Berkeley DB in particular is quite good.
You're not using SQL for the backend implementation, you're using it for
the front end, the query language which abstracts away the back end.