Sorry I agree with your manager

If its working don't try to fix it.
IF (big IF) you need to change it at a later date thats the time to think about changing the architecture imho.
But for the sake of argument lets assume I agree with you that it needs to be changed. Justifications I would use are:
1) EJBs are simple to add now with modern IDEs and EJB 2.0 local interfaces. If you use JDeveloper 10g you can have a basic Entity bean collection for a simple DB up in about 5 minutes, the session bean structure (without business logic) in another 5. It will be fine grained and some purists will go mental, but there ya go. **nontech talk :SIMPLE, LOW DEVELPMENT TIME
2) They produce an elegant OO interface on top of the database, with a nice session facade and good use of resource locator
pattern any programmer with a basic
java knowledge can use. He can forget
JDBC, prepared statements, SQL etc etc. **nontech talk :EASY TO UNDERSTAND, IMPROVE MAINTAINANCE COSTS / TIME
3) EJBs are not PL/SQL stored procedures. PL/SQL is ugly. It is not sexy and we hate it lots. We don't care if its efficient and powerful. Its not Java and therefore we hate it. **nontech talk : PLEAAAAASEEEEE LET US DO IT!
4) Even oracle are getting Java barmy so perhaps we need to get ahead of the rest and get comfortable with the new thing

**nontech talk :ADVANCED, J2EE, MODERN
Thats what id try, but ill lay odds your boss will go: Anyone paying us to change it? Anyone (and by anyone I mean anyone paying us cash) complained about the existing system? Anyone except you give a rats ass about elegant code? Ok, go bring me
coffee!
morgan
[ July 20, 2004: Message edited by: Morgan Bath ]