Re: Need help with PowerPC/Ada and realtime tasking

[bobduff@world.std.com (Robert A Duff)]

In article <31AC0712.29DF@lmtas.lmco.com>,
Ken Garlington  <garlingtonke@lmtas.lmco.com> wrote:
>...At least
>in my environment, the software engineer provides feedback to the domain
>engineer, so I suppose it is a software engineering job to get requirements
>changed, suggest additional safety features, etc.

A very good point.  If a brick-layer notices that laying bricks the way
the architect said is impossible or stupid or even just questionable,
then the brick-layer ought to notify the architect that something is,
perhaps, screwed up.

>It sounds like the point has already been made, but it is also good to
>remmeber that, technically, correctness and safety don't have to be related.
>You can have correct software that is unsafe, and incorrect software that
>is safe.

I suppose it depends on your definition of "correct".  The
proof-of-correctness folks define "correct" to mean "correctly obeys the
formal specification".  To me, that's a bogus definition.  In plain
English, a "correct" program is one that does what it's supposed to do,
regardless of whether the specification is wrong.  If you show me a word
processor that deletes my file when I told it to italicize a word, to
me, that's incorrect, even if you can show me a (bogus) formal spec that
says that the "italicize word" function should delete files.  If you
show me a flight-control program that crashes airplanes, that's
incorrect IMHO, despite the fact that you might show me a formal spec
saying it did what the formal spec says.

- Bob