🚀 go-pugleaf

Hello,
I'm looking for suggestions on how I might "trim down" my Application 
Server.  As of now, our Application Server is basically a bunch of RDMs 
with a few methods to exec queries, handle auto-gen'd keys, etc.  But as 
we begin looking at adding all of the "meat & potatoes" code, we're 
trying to figure out where this code should go.  Some of this code 
involves some very complicated processes, and I'm thinking that it would 
be nice to be able to separate some of it out into DLLs (or maybe 
packages?) -- so that it's only loaded into memory "as needed".  Some of 
our users will NOT be running the App Server on a dedicated box -- it 
will be running on each individual workstation alongside the client app -
- therefore, we must keep memory requirements in mind (we're assuming a 
32mb workstation).

Here are the two main issues/questions:

1)  Could we indeed move some of these processes out into DLL(s) or 
packages?  If so, which would be better?  Furthur, how could we access 
these processes via the App Server?  Would we create "shell" interface 
methods in the app server which dynamically load the DLL(s) and do 
whatever it needs to do?

2)  What if we need to access data in the DLL process(es)?  Since all of 
my RDMs are already set up as a part of the App Server (along with 
validation rules and such), do I have to set up separate data modules for 
the DLL to use or can I somehow use those RDMs from the DLL? (That sounds 
scary to me!! <g>) 

I apologize if these questions sound confusing, but we're learning this 
multi-tier stuff as we go.  We just want our app to be architected in the 
proper fashion from the beginning -- and NOT have to redo everything 3 
months from now!!

thanks in advance...

Kevin Talley (ktalley-at-synergysoftware.com)
(replace "-at-" with "@")