[avr-gcc-list] OT - in defence of MISRA (was Feature request ...)

[Graham Davies]

"Douglas Dotson" wrote [in part]

> ... These MISRA guidelines seem to me to be just
> another in a very long line of attempts to make
> programs better by enforcing rules. My experience
> is that success using this approach is questionable
> at best.

I basically agree with this when you are considering any individual,
qualified and experienced software engineer.  To cut the babble about my
personal experience, I think rules have a place in environments where you
have teams of indifferently qualified engineers who have not yet had time to
develop personal insight into what is good and what may bring trouble.

Example follows:

I don't consider myself particularly inexperienced.  Yet, when I brought up
my problem with a compiler generating multiple stores to a volatile variable
as a result of one assignment in the source, what I took away from the
discussion was that, whether or not the compiler violated ANSI C, I should
not have trusted it in this case.  A "rule" figured out by you guys in
committee would have saved me some trouble.

Graham.