At 06:42 PM 7/31/02 +0200, Stephen D. B. Wolthusen wrote:

. . .

>So yes, getting students to do Ada is the issue. I've been toying on and
>off with offering a graduate-level Ada / Formal Methods / Security
>Engineering course at Darmstadt University since just bit**ing about the
>problem doesn't help -- even though that would properly be a task for the
>regular faculty, not a hobbyist like myself. Maybe I'll find the time to
>prepare something in time for next fall.

Actually, I came to the conclusion, long ago, that getting "students"
(especially C.S. students) to learn Ada or Software Engineering, is not the
correct, nor even a viable, solution to the problem.  In fact, that is
precisely one of the main reasons so much crummy software is being written,
because we (our industry) have taken it for granted that the problem can
only be fixed by fixing C.S. education.  We (especially we the ACM) have
been working on this principle for at least 35 years.  And look at what has
happened.  The number of C.S. grads rose steadily during most of that time,
and may still be rising, though it might be more in steady state now (Mike,
do you have any recent figures?).  If we had a good way to plot the quality
of software over that period of time I conjecture the result would very
nearly match the inversion of the rate of increase in C.S. grads.

Well over a year ago I shared my alternative solution in this forum, but it
was not well received, so I won't bring it up again.

Just every one please consider that we have to at least recognize that our
"traditional" solution obviously does not work.

S. Ron Oliver, the U.S. representative for Top Graph'X, developers of high
quality software components, using Ada, including OrbRiver the
multi-language ORB.  A single distributed programming environment for all
developers.  Supports Ada95, Java, and C++.

For more information, check out

Semi-retired professor of Computer Science and Computer