Re: decline of Ada?

[adam@irvine.com (Adam Beneschan)]

JBisaillon <jbisaill@brainiac.com> writes:
 
 >I am not sure what the exact thread is here, I have a sense of what
 >it might be.  Let me shed some light on a few realities.
 
Uh oh, the "R" word.  Usually a sign that what follows is
opinion-heavy and fact-light.
 
 >1) If you are an Ada programmer, you had better pray that you don't
 >get downsized.  All that training, years of experience, are virtually
 >worthless unless you can find another DoD Contractor
 >doing Ada.
 
I kind of doubt that.  I've worked in Ada for 8+ years, and in that
time I've learned a ton of stuff that has nothing specifically to do
with Ada (software engineering techniques being just one of the major
things I've gained experience in).  Surely I would be a much better C
programmer now than I would be in my previous C programming job eight
years ago, or a better COBOL programmer than in my previous COBOL
programming job, etc.  So the statement that my experience is
virtually worthless is pretty silly.
 
 >2) DoD Contractors will not build in Ada because you can't find
 >trained coders to work for 24K to 32K per year.  But you can find "C"
 >coders.  Therefore, if you want the contract, you cut the cost and
 >program it in "C".  It really has very little to do with the language,
 >its advantages, or disadvantages.  It has to do with dollars.

Management would have to be pretty short-sighted to think like that.
Anyone who thinks that you can cut your project cost in half by hiring
programmers at $24K instead of $48K needs to write, "You get what you
pay for", 100 times on the blackboard.  (Of course, not all $48K
programmers are worth twice as much as all $24K programmers.)  I
suppose that hiring cheaper "trained coders" (why does this remind me
of "trained chimps"?) might allow a contractor to submit a lower bid,
and put off until later the concerns about the cost overruns that will
occur when the project doesn't work and your programmers, who only
have enough experience to be making $24K and are working in a language
that makes it more difficult to track down bugs, take much longer than
expected to get the bugs out.

Now that I've gotten all this off my chest, there's no _a priori_ way
to know whether it's actually cheaper to write a system in Ada (and to
maintain it, and to port it to other systems where necessary, all of
which need to be considered as part of the cost).  I seem to recall
that there have been studies showing that some Ada systems are less
expensive to produce than other languages despite the increased time
for initial coding.  But I don't remember any details, and my memory
may be faulty.

                                -- Adam