Re: Procedure Examples VR7.1

On 2/3/2017 11:54 AM, Duane Scott wrote:

I am curious as to what experiences led up to these rules.

ILE has many options, many choices. That's to enable very sophisticated
applications. I don't have a sophisticated application and suspect few
do. What's sophisticated? One example: Sharing open data paths where
one program does a CHAIN (and no UPDATE) and a different program does an
UPDATE (and no CHAIN). Now imagine isolating that so that these same
two programs are called twice; in separate activation groups. And you
don't want the CHAIN from AG(ACCOUNTING) to bleed into the UPDATE for
AG(SHIPPING).

And to what extent (or minutiae) are the procedures/functions developed. For example: while getting a name and address from a customer record is something that could be used in many programs (and SQLs) within a system (or multiple systems such as A/R and A/P accounting systems), getting just the name or phone number of the main contact for that customer record may not be necessary in but a few. Would a separate procedure be written just to grab the contact from customer record? How would that be written.

I've been very happy with my discovery of 'only write as much code as
you need right now'.

I would assume that both would be in the same binding directory.

In my system, both would be in the same service program. The binding
directory identifies the service programs to the calling RPG programs.
The calling RPG programs' source code does not know or identify the
service program, the only thing the source code has is the sub-procedure
name. Unless you are smarter than I am and name your sub-procedures
Accounting_getName() and Shipping_getName() :-)