RE: How can I Improve this procedure...

Brad,

>>If we were really *that* worried about a small performance gain, 
>>well, websphere salesmen would be flipping burgers.

This is, at best, specious reasoning. 
First, please define what flavor of the "WebSphere" umbrella you're referring
to.

Second, "small performance gain"?  A "small change" doesn't equate to a "small
performance gain". So you're saying its better to spend hundreds of thousands of
dollars fixing something with hardware, than to spend a few minutes/hours/days
getting your own code to run faster? Or better yet, spend the money rather than
ask IBM to improve the performance for everyone who uses lengthy return values? 

Obviously you're not spending our own money, but if the people you're working
with have a blank check, or don't care if something takes 10 minutes to run
instead of 4 or 5, then you're in a better position that I am.

For me, I have the energy and drive to make my code work faster and better. 
If everyone writes crap code, then all code will be the same quality. There will
be nothing compelling about your code or my code vs code written in the 2006
incarnation of a sweatshop.

-Bob