[H-GEN] Dismal performance of Mars mission software

Greg Black gjb at gbch.net
Wed Feb 18 00:13:13 EST 2004


The background to this rant[1] can be found in the story at

    http://www.securityfocus.com/news/8054

The story above is a bit of a breathless news report written by
somebody who listens to the spin doctors at the software house
and does not quite understand the material he is describing, but
it is easy enough to see the story behind the descriptive prose.

What we have here is a computer system whose basic hardware and
software has been deployed in countless instances over many
years; the list of deployments is rather scary when you consider
how bad it is.

Essentially, it seems that one of the Mars rovers fell into a
catatonic state because it frittered away all its rather small
amount of memory on storing useless files that were added to its
records day by day.  Eventually, it wasted so much memory with
crap that it was not able to reboot.  Then its batteries failed
and it came up in some "fail-safe" mode when they recovered, and
this mode avoided loading the rubbish.

The company representatives are busy congratulating themselves
on their cleverness in getting it to re-start in this way, when
they should be falling on their swords for the incredible
stupidity that allowed the system to fail in the first place --
and, by failing, to cause a lot of grief and wasted effort on
the part of the mission control people back on Earth.

The restricted memory of the system could not possibly have been
a surprise to the designers; the squandering of this limited
resource without any thought about its management is the height
of stupidity.  I really don't understand how it is that people
with so few clues can ever get to be in charge of stuff that is
so hard to fix when it's in the field.

Many of us here have spent time excoriating Microsoft for their
woeful software -- and that's fine, for their software is really
bad and its failures cost people and organisations huge amounts
of time, money, grief, etc.  But here we have an embedded
systems company that prides itself on making vastly superior
products, and somehow convinces customers of its superiority,
making schoolboy mistakes.

Sure, they got away with this one -- but how many other "lost"
space missions fell prey to some similar stupidity?

Cheers, Greg

[1] This footnote is for those who have noticed that this might
    have been a blog entry if I had a blog ...




More information about the General mailing list