Re: Non-defense Ada applications - answering several requests

[agate!howland.reston.ans.net!noc.near.net!inmet!spock!stt@ucbvax.Berkeley .EDU (Tucker Taft)]

In article <SRCTRAN.93Jun3233419@world.std.com> 
  srctran@world.std.com (Gregory Aharonian) writes:

>Mike,
>	Each time you post this, I keep on thinking "Yea, well mine's
>longer".  If I posted a list of similar C/C++ projects, it would be
>50 times larger for projects with a similar range of sizes. 

Actually, I think it would be doing the public a service if
someone would publish a similar list, but restrict it to C++.

My suspicion is that if one looks at large, long-lived, reliable
systems, one will find that Ada is being chosen for such systems
at least as often as other languages.  Almost all air-traffic
control systems are being implemented in Ada these days.
Certainly a lot of telecommunications is in C/C++, largely because
of the historical dominance of AT&T and Unix in that realm, but 
even in telecommunications Ada is being chosen (e.g. by NTT in Japan).

The fact is that when reliability, maintainability, and overall quality are
prime factors, companies are less swayed by how inexpensive are
the PC compilers for a given language.  It would certainly be great
if Ada compilers were less expensive, but in a large, long-lived system,
the cost of the compilers is dwarfed by the cost of development
and maintenance.

I suspect that if someone tries to put together a list of large,
long-lived, reliable systems being built in C++, (s)he will find
it is hard to produce a list much longer than the correponding Ada
list, for much the same reasons -- there aren't that many such
systems around, many companies keep these things to themselves, and
older languages like Fortran, Cobol, and assembly language are still 
being heavily used.

It would be useful to see such a list because it will help people
make more objective assessments of which language to choose
when facing the development of similar systems.  The major goal
of Ada relative to other languages is the production of reliable
systems, and so any information on the development of reliable systems
in any language would be very appropriate for posting to comp.lang.ada.

S. Tucker Taft  stt@inmet.com
Intermetrics, Inc.
Cambridge, MA  02138