From: Jim Cochrane <allergic-to-spam@...>
Date: 2006-11-26T06:05:06+09:00
Subject: Re: Bug in ruby?

On 2006-11-22, Paul Lutus <nospam@nosite.zzz> wrote:
> Jim Cochrane wrote:
>
> / ...
>
>>> The bug that cause Mars Polar Explorer to prematurely shut down its
>>> engine and fall the last 100 meters to the surface of Mars, never to be
>>> heard from again.
>> 
>> Due to a missing or incorrect unit conversion, right?!
>
> No, I left that one off my list because it was simple human error, not
> really software-related. One outfit was using English units, the other one
> Metric, and they failed to realize the discrepancy until the spacecraft was
> of course.
>
> The error I am speaking of caused a Mars spacecraft to shut down its descent
> engine prematurely due to the misreading of a sensor. At least that's the
> best theory so far for a spacecraft that didn't have any terminal phase
> telemetry.

Thanks for the clarification.  I had forgotten that there were two
different major Mars-mission errors.

>
>> Or how about the Ariane 5 (
>> http://www.around.com/ariane.html
>> ).
>
> Yep. Another bad one.
>
>> I don't remember ever hearing a description of the defect (or 
>> defects) that caused the DIA baggage sysgtem to fail (
>> http://www.cis.gsu.edu/~mmoore/CIS3300/handouts/SciAmSept1994.html
>> ).
>
> I think that one was mostly bad mechanical design. They were trying to do
> away with human baggage handlers entirely.
>

I did a bit of googling/reading on it when I posted the link.  It sounded
like the main cause was bad management and planning - poorly managed
requirements creep, not allocating enough time for thorough analysis of a
complext system, etc.  I think there were several system-design flaws,
some of which were mechanical, but I believe I read that software
defects also contributed to the problem.  Just a general, overall
foobar-ification.  (I've not heard how much, if any, of the original
DIA system is still in use these days; it'd be interesting to find out.)

--