So, I'm reading Ian Abramson's Beginners guide to Oracle Database 10g (coming from using mysql), and it mentions PL/SQL and gives all these great reasons to use it, however I'm doing this in preparation for a rewrite of several applications my company uses.
My head programmer argues that if PL/SQL is no more efficient at database access, all parsing and queries should be done on client side programs to spread the load out from the Oracle server. I argue that we should compartmentalize our code and attempt to avoid code duplication, and PL/SQL may help us do that.
The problem is, we don't actually know (And this book doesn't tell us), how efficient database access is in PL/SQL over an application running on a different machine that makes 3-5 (largish) queries per operation. I would think it's more likely to run faster because (to my knowlege) pl/sql functions store and save their execution plans, and the traditional network overhead would be cut by a factor of 3-5 (depending on the number of queries)