Re: += in ada

[Marin David Condic <nobody@noplace.com>]

Sometimes it is neccessary - either you do a hand optimization or you 
drop the project. (Been there. Done that. Same processor as the Ariane 
5, BTW)

But the point ought to be that in the overwhelming bulk of applications, 
this is not necessary. Most of the time, what the compiler does is 
*better* and one should leave it alone. Once in a great while, you are 
so bound by time, you *must* turn off all checks, struggle for every 
optimization, etc. This is just extremely rare and should not be 
undertaken lightly.

Also, it was never clear to me that leaving in the checks would have 
saved the Ariane 5. After all, the software did exactly what it was 
designed to do. It detected a fault, presumed it was a hardware failure 
and shut down the channel. Had the overflow been trapped by an exception 
handler instead of the hardware interrupt, the decision would have been 
the same. The FDA was correct for the Ariane 4. Just that in the Ariane 
5, it wasn't a "Failure" so the "Detection" and "Accommodation" was the 
wrong thing.

MDC



Hyman Rosen wrote:
> 
> Then the Ariane 4 people counted the cycles.
> You were claiming that Ada people believe that this sort
> of micro-efficiency is in the domain of the compiler, and
> I am pointing you at a prominent real-life project where
> this was not the case.
> 


-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: m c o n d i c @ a c m . o r g

     "All reformers, however strict their social conscience,
      live in houses just as big as they can pay for."

          --Logan Pearsall Smith
======================================================================