RE: Service Programs VS normal Programs

 >> They have to spend time again in rewriting, redesigning their
applications with a ROI of $0.00.

I have to disagree with this statement because I have seen so many cases
where it was not true.

There can be a _very_ significant ROI in a number of areas.

First of all in maintenance - any program no matter how well written
originally tends to get harder and harder to modify as time goes by.

Second and perhaps more importantly - the "hidden opportunity" area.
Several customers that I have worked with have benefited greatly from being
able to use the more modular "business object" approach that they
implemented as part of the rewrite in other areas.  One customer told me
that 50% of the usage of their new components is in areas that they didn't
even know would exist when they started out.  Their ability to respond to
change has gone up by an order of magnitude.

In some ways the RPG compiler is an object lesson.  We get more new features
with _each_ release of RPG IV than we got in the _whole_lifetime_ of RPG400.
Admittedly some of this is due to the fact that the team developing it are
among some of the best programmers it has ever been my privilege to know.
But some of the benefit also comes from using a highly modularized
construction that can be modified far more simply and faster than the
previous compiler could ever have been.

Should one re-engineer for the sake of it?  No.  But the "don't fix it if it
ain't broken" is often a fallacious argument since "broken" is too narrowly
perceived.  if you modify the phrase to "don't fix it if it never needs
enhancements or maintenance" then I might be inclined to agree.  Otherwise
all programs reach a point where a rewrite/redesign will yield benefits in
the long term.  And perhaps that is the problem - we're all too focused on
the short term.

Jon Paris
Partner400
www.Partner400.com