Print

Print


[log in to unmask] wrote on 1997-11-07:
>AdaWorks <[log in to unmask]> 11/07/97
>05:46pm wrote:
>
>>I wonder if Mr. Paige has any idea the
>>damage that has done with his recent
>>directive.  Here in Silicon Valley, my
>>colleagues in other software environments,
>>even those at some of the defense
>>contractors, have interpreted the directive
>>as a clear statement that Ada is no longer
>>important as a DoD language.  I was in the
>>office of the CEO of one well-known DoD
>>contractor recently who said, "Get with it
>>Richard. Ada is dead, dead, dead."  His
>>company has pulled out of the Ada
>>marketplace entirely.
>
>Although I smell similar foul breezes here,
>my company (Boeing) does have a booth for
>Tri-Ada '97 here in St. Louis which may be
>some small antidote to your sad story
>(plug, plug ;-).
>
>>Some others are scrambling figure out how
>>they can transition their Ada 83 to C++ in
>>the next round of software upgrades.
>
>Amazing.  Transitioning from C to Ada was
>"too costly", but transitioning from Ada
>(83 no less) to C++ is a top priority.  I
>sniff a hint of these foul breezes around
>here too, though nothing concrete yet.  It
>would certainly lend credibility to the
>notion that the reason why Ada is losing is
>because "it's no fun to program in Ada.
>You can't do cool stuff."
>
>>Meanwhile, C++, a dreadful gargoyle of a
>>language, continues to enjoy a place in the
>>sun, or Java a place in Sun.  The more I
>>look at C++ the more I realize how
>>dangerous it is.  Nevertheless, I expect
>>that more and more of our warfighters are
>>going to be risking their lives on the
>>quality of our C++ code.  Consequently, our
>>responsibility is to do everything we can
>>to ensure that, as Ada fades from
>>prominence in our weapon systems, we
>>encourage the C++ developers to use that
>>language with extreme care.
>
>Quite frankly, that is the area in which I
>see the opportunity for me to excel, since
>I do not have language-choice
>responsibilities.  I advocate when I can
>(and watch the eyes roll back in the
>sockets - a depressing reaction to say the
>least), but 99% of the time, that is not my
>place (for all I know, if it were, I'd be
>horrible at it anyway).
>--
>James Squire
>MDA^H^H^HBoeing St. Louis
      http://www.boeing.com

I participated in the 16th Digital Avionics
Systems Conference the week before last and
had the opportunity to encounter a few scary
things as well as have some apparent success
with Ada advocacy.  In fact, my presentation
was entitled "The Effects of the New Military
Procurement Practices on Software Development"
and was very well received as taking a reasoned,
balanced approach to several issues (language,
reuse, ...).  I ended up participating in a
panel discussion, along with Joyce Tokar (DDC-I
should be commended for having her participate;
I did not see any other compiler-vendor folks
in any form at the conference) and others.

My position is that if you have a small micro-
controller doing simple things, and code in C
has been used for many years to do those things
on that controller, then, by all means, reuse
the C code.  For complex new software on major
processors for which an Ada compiler is available
(and that should help drive the processor choice),
code in Ada.  I talked about some of the historic
baggage that Ada still carries from its early
heritage in the mid-1980s (cost, immaturity of
compilers, ...) and why that is wrong for Ada 95.
I also explained some of the major problems being
experienced by users of C++ and Java.  Some people
came in sympathetic to C++ and left saying they
were going to warn their people back home against
using C++.

The scary thing was the push for COTS.  The FAA is
picking up on this from DOD, without understanding
how Commercial Off-The-Shelf software is just a
special subset of Non-Developmental Items (NDI).
They seem to picture all the tools in boxes down
at Acme Computer Store they can buy, just like
Wile E. Coyote (Roadrunner cartoons), rather than
pay to build them.  They do not read the licensing/
(non-)warranty information on the side of the box
and correlate with the high reliability and safety
issues critical to warfighting and commercial-avionics
software.  Yes, reuse existing software with care
(emphasis on "with care", remembering Ariane 5),
but that is not the same as buying something off
the shelf.  I managed to put a scare into an FAA guy
there, although I doubt it will do much good in that
case, because the issues seem to be politically charged
and driven from the top.  There is nothing like having
technically unknowledgeable bureaucrats trying to look
like they are doing something and stepping in with
their solutions to technical problems, and all their
underlings are going to comply, come hell, high water,
or airplane crashes!

Howard W. LUDWIG
Working (but not speaking) for Lockheed Martin Electronics & Missiles Co.