Garbage Collecting

I'm a postgresql newbie that's inherited eight production servers running
Postgresql 8.2.5 as the backend. I have many questions covering topics such
as administration of the database (upgrading, maintaining conf files, etc),
improving the schema of the system (many tables don't currently have primary
keys; to do anything useful you must join at least 5 tables), optimizing
poor performing queries that can take hours, and knowing where functionality
of the system should reside (curenly as PL/SQL functions, as external c
code, external php code, and external perl code).

Please refer me to appropriate documentation/FAQs/books. I've read
http://www.postgresql.org/docs/faqs.FAQ.html If anyone knows of writeups for
newbies that touches upon the things I mentioned, that would probably be
really helpful for me.

I have one specific question about "Garbage Collecting" within the database.
The database system I'm working with has data that is no longer needed after
a period of time. For example: transaction records only need to be kept
around for the last 31 days; php web sessions that don't need to persist
longer than a day. Could I create some function in the database that would
act a bit like a daily cron job that deletes old records from tables (and
then performs the appropriate VACUUM to regain the space)?

If yes, how does one impliment something like that? As a trigger function
written in PL/SQL? Can I hook the function into something that executes once
per day?
If no, why? Should the external scripts/code that puts the data into the
database be responsible for removing the old data?

Thanks in advance for any/all pointers!

-Joshua

--
Joshua Berry
Software Engineer
Opentech, S.A.
+(595 21) 282557 Work
+(595) 981 330 701 Mobile