Re: how to determine cost savings based on resource utilization

On 12/3/2014 9:47 AM, Tommy Holden wrote:

I'm trying to convince the powers that be to allow us to modernize (or at
least update) our custom code. We still have a ton of old RPG 3 (written
more like RPG 2) and way too many CLs using OPNQRYF, etc. and the question
is how is it possible to quantify this in terms of $$$. I know that I can
come up with a semi-decent guess as to man hours saved on program
maintenance but I can't even hazard a guess as to how I can calculate the
reduction in resource utilization, response times, etc. does anyone have
any ideas or methods they have used to try to at least get some inkling of
the savings of modernizing?

If changing from left hand indicators to DO or IF blocks reduced my
resource utilisation, I sure couldn't notice it. This might be because
RPG is a compiled language.

We are currently migrating to a web based
financials (with the added bonus I get to at least do some cleanup on the
interfaces for some of this jalopy code). Since I have a small bit of
leverage during this process and we are in a holding pattern with the
vendor I'm trying to get a consultant to work on updating the existing
code to use modern techniques (i.e. Service programs, SQL, etc).

Anyway if you manage to decipher that long-winded diatribe & have ideas or
suggestions it would be MOST definitely appreciated!!!

Modern coding reduces the load on programmers, not machines, and that's
a very good thing. I've never convinced a manager that rewriting my
code using modern techniques, but doing the same thing as it always did,
has value. On the other hand, I don't ask my manager if I can modernise
code that I'm already making changes to. Other programmers call this
refactoring, and it means 'improve the code I am touching without
breaking anything'. Of course, other programmers tend to have bite
sized functions and test suites; neither of which we are going to have
with old fogey code. This means that I refactor as I do maintenance,
and build test cases on the fly. Since I'm going to do an all up test
on the changed code anyway, that's as good an opportunity as any to
spiff up the look.