From: JD3GTRCW.TRANSCOM@transcom.safb.af.mil (CONROY WILLIAM F)
Subject: Re: INFO-ADA Digest V93 #522
Date: Thu, 12 Aug 93 10:26 [thread overview]
Message-ID: <9308121533.AA24097@ajpo.sei.cmu.edu> (raw)
This forum is extremely interesting to listem to some of the threads. I
work in a large govt organization trying to live with the ADA mandate.
It is extremely interesting to hear the holy wars and the argument to
follow the mandate because and only because it is law. I'd like to give
you some comments for thought:
ADA is a tool, it is only a tool. It is extremely good for some problem
domains, very poor in others.
The problem in the DOD the ADA mandate was addressing was the total lack
of standards and the inability to perform software verification on
complex mission critical systems. A second extremely subordinate point
is the ability to reuse code and attempt to have platform independance.
Another aspect of the problem is the C++ vs ADA issue. ADA is a very
good language designed by academics to support the verification process.
It has some really lousy implementation aspects which have been covered
for days in this forum.
Final problem: The world is driven by $$$$. The ADA mandate was pushed
through in a period when the govt had the arrogant opinion that we could
enforce our standards by law, forget the cost. This created a situation
where a few companies produced OK compilers at high cost, extremely slow
compile times, with minimal optimizations. The ADA mandate completely
violates the concept of the free market system. The C++ world, and I
agree C++ is certainly not a silver bullet, is not driven by a mandate.
Rather it is driven by market forces which are much stronger. As a
result we are seeing cheap, robust compilers which survive by being
responsive to the needs of the customer in terms of speed and features.
Pick up a computer rag and see how many add on ADA packages exist which
actually solve problems? If you can find them, what is the cost? As a
program management office I am resposible for producing computer systems
which "fix" real world problems within a decreasing budget. ADA is just
a tool, a rather expensive tool. I am required to pay top dollar for
decent, not spectacular, programmers. I have to spend a lot for
compilers which are incredibly sloooowwww..... I find the source code
extremely portable as long as I stay with one vendors implimentation and
don't use any of the vendor supplied libraries for anything like graphic
i/o or other non standard capabilities. Finally, I put up with a process
which takes an absolutely ludicrious number of years for the language
committee to fix basic <read basic> language architecture problems.
Finally, AT&T in their switching circuits have the same requirements for
robust verifiable software that we have. With the country in significant
trouble economically, how can we support spending a significant premium
to use a niche language, give up access to languages being driven by the
same market forces driving us. If you SERIOUSLY want ADA to be the
language it could be, you need to emphasize places where ADA has actually
an advantage over other languages in fixing "real" problems.
Send flames to congress or /dev/null
PS. I do comply with the ADA mandate and it costs the US taxpayer a
significant premium with no significant benefits over other tools we
could have used.
Bill Conroy <conroyb@tiberius.safb.af.mil>
reply other threads:[~1993-08-12 10:26 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
replies disabled
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox