When Linux goes bad at 10,000 Feet

[me-5yx05kfkO/aqeI1yJSURBw at public.gmane.org (Matthew Gillen)]

Robert La Ferla wrote:
> http://www.akihabaranews.com/en/news-17171-When+Linux+goes+bad+at+10%2C000+Feet.html

People that see a crash and make all sorts of extrapolations about "stability"
bug me...

I've "flown" linux on embedded systems that were at near-space altitudes, and
in my experience OS crashes are almost always hardware related.  So the
poster's implication that BSD might have done better in that situation is
highly questionable. (note: I've never used Windows or it's variants in that
environment)

It turns out that heat dissipation is a real problem for flying embedded
systems (even at altitudes where it's fairly cold), since you need to seal
them pretty good to keep moisture out, and you can't rely on the outside air
being cold all the time since there is a significant portion of the flight
when you're close to the (relatively hot) ground.  This might be less of an
issue on a passenger-jet than on a space-constrained UAV, but the point is
that no OS can compensate for an overheated processor or memory chips...

On the other hand, it could be that the on-board video system uses a
custom/proprietary driver, and that driver turned out to be flaky.  Not only
would a BSD-based system have the same problem, you could argue that the more
liberal licensing terms of BSD vs. the linux kernel would encourage
closed-source/proprietary drivers (and the associated flaky-ness).

Matt