RE: when to use a function or file for information

I'm lucky in that I am the only IBM i programmer here so keeping track of changes is relatively easy (except when the senility slips in). In the case of a multi-programmer shop it can be difficult, even with programmer work libraries, to keep things in sync without some kind of CMS (even a homegrown one) where the objects are checked out. Otherwise, you can spend a lot of time manually coordinating changes ("Hey, I need to add a field to the Customer Master. Anybody else working on that one?").

I like Loyd's approach of adding fields to a secondary file. In fact, I'm in the process of adding two fields to our sales history. Rather than change a ton of queries and programs, I created an adjunct file with the same keys as the primary. With a join of the two files (using DDS since SQL creates views which are non-keyed) I can access everything with one CHAIN, or I can access just the file that I need. I've used the approach in the past for our Item Master and Customer Master tables, and it made life a lot easier. Usually the adjunct tables and joins are used in fewer processes so adding a field to one of them is less of a maintenance pain.

Jerry C. Adams
IBM System i Programmer/Analyst
--
B&W Wholesale
office: 615-995-7024
email: jerry@xxxxxxxxxxxxxxx


-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Chris Bipes
Sent: Tuesday, March 23, 2010 10:23 AM
To: Midrange Systems Technical Discussion
Subject: RE: when to use a function or file for information

We have this issue all the time. What I am working towards is a single development file lib that is identical to production. Now programmer put changed files into their own work library and all the objects they are changing. Their development library is ahead of the general development library until they are ready to install. Then they install to the general Development library and test. From there they promote to production or QA depending on the scope of the project.

Chris Bipes
Director of Information Services
CrossCheck, Inc.

-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of Jonathan Mason
Sent: Tuesday, March 23, 2010 8:01 AM
To: Midrange Systems Technical Discussion
Subject: Re: when to use a function or file for information

Hi David

Surely if the value could be identified by referring to the CustomerType
then you wouldn't need the extra field anyway.

Really what it comes down to is basic change control for your system.
Anybody making a modification to a file should be running some kind of
impact analysis to identify which other objects (programs, queries, etc
) make use of the file so any changes or recompilations could be
actioned.

There are many change management tools out there and I'm sure if you
asked our host, Mr Gibbs, he could recommend a very good one that we
also use.