Re: Ariane

[Dan Nagle <dannagle@verizon.net>]

Hello,

I'm cross-posting this to comp.lang.ada,
so we can get some more input.

To c.l.a readers:  sorry to barge in like this,
but this discussion of the Ariane 5 failure
has been going on in comp.lang.fortran
for far too long.  I reckon some c.l.a folks
can clarify the issues.

In short, robin argues that there is an inviolable rule
that all exceptions must always be handled, irregardless
of other engineering constraints, such as processor load.

dnagle has been arguing that there were a combination of decisions,
which taken together caused the Ariane 5 failure.  In my view,
robin is simply incorrect: engineering often requires
calculated trade-offs among competing constraints.

robin claims he has read the relevant ESA reports,
for example,
http://sunnyday.mit.edu/accidents/Ariane5accidentreport.html
and/or the PDF version
http://ravel.esrin.esa.it/docs/esa-x-1819eng.pdf.

robin wrote:
> Dan Nagle wrote in message ...

>>Having the computer shut down, after putting it's error code
>>on the data bus (and not on a command or error reporting bus,
>>for example).

> That alone meant that it was essential that ABSOLUTELY ALL
> exceptions be trapped.

But the shutdown was a separate engineering decision
from the decision to save the cycles needed to keep the processor
under 80% utilization!

I've argued all along that there wasn't _one_ decision
causing the failure, but rather a combination.

Your hindsight has you simply fixated by the operand error issue.

-- 
Cheers!

Dan Nagle
Purple Sage Computing Solutions, Inc.