The demise of the I-CASE program says a lot about what's going on in the Defense Department when it comes to software. I-CASE--for integrated computer-aided software engineering--was designed to impose a uniform methodology and toolset for all DOD software development. It took five years from conception to award in April 1994. Yet last month DOD brass reduced the ambitious program to merely an indefinite-delivery/indefinite-quantity contract for software products in the I-CASE catalog.
The demise of the I-CASE program says a lot about what's going on in the Defense
Department when it comes to software. I-CASE--for integrated computer-aided software
engineering--was designed to impose a uniform methodology and toolset for all DOD software
development. It took five years from conception to award in April 1994. Yet last month DOD
brass reduced the ambitious program to merely an indefinite-delivery/indefinite-quantity
contract for software products in the I-CASE catalog.
I-CASE wasn't a bad procurement per se, nor did it offer bad products. In calling for
its reduced role, Assistant Defense Secretary Emmett Paige cited product prices. But he
was nodding to the reality that even after I-CASE got rolling, few in the DOD trenches had
bought in.
The failure of I-CASE as a program must therefore be seen in context, since
other DOD software initiatives are equally troubled.
The Ada mandate may be alive and well in the minds of upper brass, but it's pretty
tattered at the outposts where development takes place.
And, under Defense Secretary William Perry, support for the migration systems effort is
waning. It turns out that simply declaring this financial system or that personnel system
to be the prototype doesn't mean anyone actually migrates there.
What happens to all these plans?
Certain fundamentals stand out. And there are lessons in them for all of government,
not just DOD.
The first is that top-down mandates without real support from those on the line rarely
take root in large organizations. In the case of I-CASE, many of the original proponents
have moved on, making the original mandates all the more hollow.
Equally important, the goals of a program get obscured if the focus narrows too much on
details of execution. The real aim of I-CASE was not--or at least should not have
been--for everyone to use the same tools. Its purpose was to help bring badly needed
predictibility and interoperability to software development.
There are signs of hope. One is DOD's new emphasis on upgrading its software centers'
capability maturity ratings. Another is that as migration systems fade, there is increased
emphasis on standard data elements--the one absolute requirement for eventual
interoperability.
I-CASE may have failed, but its original goal still is worth shooting for.