Air Force shows how meaningless Ada waiver process is

[srctran@world.std.com (Rhoda Metzger)]

    A recent Air Force article shows how meaningless the Ada Mandate is in
terms of providing an insurmountable obstacle to using anything other than Ada
for fielded systems: the Ada Mandate is always easily worked around by
experimental-to-fielded source code creep.

    The September 5 issue of Government Computer News has an article about an
Air Force developed system for airfield analysis called the Combat Readiness
Infrastructure Support Information System (CRISIS).  CRISIS was developed at
the Air Force Academy in the 1980s by some of the civil engineers there as a
way to make maps of the service's air fields and bases.

    The system is based on AutoCAD (a nice COTS system vis-a-vis Perry's
memo), with the MSDOS version using AutoLisp and Basic to provide the extra
functionality needed to support Air Force needs, and the Unix version being
extended with AutoLisp, C and C++ code, with links to external database
management systems.

    For peace time use, I suppose this violation of the Ada Mandate is
probably tolerable, but then the creep sets in.

    During Desert Storm, the Tactical Air Command used a beta version of
CRISIS to deploy forces in desrts of the Persian Gulf. By importing satellite
images of the area into AutoCAD, engineers were able to model airfields in
only a matter of minutes.  "When we started developing CRISIS, there was no
off-the-shelf software that could do what we needed.  Even now, there is
nothing as customized as CRISIS for the way the Air Force does things".

    Thus a rationale for getting around the Mandate. Develop something off the
waiver radars for a non-fielded application and then gradually introduce to
users in the field, which assuming the application has benefits, will slip
into fielded use with little concern about any non-Ada dependencies.  This
is not an isolated case.  SIGNAL magazine had an article a few months ago
about a similar creeping Air Force program now used in NORAD operations.
It's a shame no one writes clear memos for Air Force people that actually
mention Ada for advanced applications.

    What's worse is that these creeping success stories seem to figure out
how to do what Ada success stories can't - how to get published.

    Once again we see the myopia of DISA in continuing to reject the idea of
doing a comprehensive programming language use surver inside and outside the
DoD.  With such a survey, many of these creeping efforts can be nipped in the
bud and transitioned to Ada before it is too late.  Why DISA continues to
refuse to do such a survey defies Ada-supporter analysis.  And nothing in the
DualUse Plan addresses this detection issue.  If I can do it with no money,
they should be able to do it with all of the millions they have to spend.

Greg Aharonian