Why Spirit stopped working...

Another great story about how hard 'complete' testing is and how important exception handling is in software.

http://www.eetimes.com/sys/news/OEG20040220S0046

Kentasy
Friday, February 27, 2004

What I find remarkable is not that they had a bug, but they anticipated the fact that they might have problems and had systems in place to recover.

pdq
Friday, February 27, 2004

So the Spirit was willing, but the flash was weak. :)

sgf
Friday, February 27, 2004

Most complex embedded systems have recovery mechanisms for when it really goes wrong.

I am still somewhat stunned that they
made such a bad mistake. It's not like it's a difficult
scenario to construct.

son of parnas
Friday, February 27, 2004

Picking the winning lottery numbers the day after the lottery is easy too. 

Of the 5,308,134 things to test, something is always going to be left out.  Otherwise the project would take 50 years, and cost $100,000,000,000,000. 

Sure it was an easy scenario to setup, it was just thinking would actually be a problem that was hard.

AnonAnonAnon
Friday, February 27, 2004

Pish. This isn't monday morning 20-20 hindsight.
I work on embedded systems and it should have
been thought of.

son of parnas
Friday, February 27, 2004

Here's what really happened to Spirit:

http://images.spaceref.com/news/2004/rover.armspin.mov

(Via the Chris Sells blog.)

Robert Jacobson
Friday, February 27, 2004

Yeah parnas let's see your space probe. Sheesh.

Rocket Scientist
Saturday, February 28, 2004

sgf:  OUCH.  (Nice pun.)

Sam Livingston-Gray
Saturday, February 28, 2004

Rocket scientist, i would love to create a space probe.
I would make mistakes, but hopefully they would be
a little less obvious ones.

son of parnas
Saturday, February 28, 2004

Hehe, the arm-spin animation is my work :).
www.maasdigital.com

Dan Maas
Sunday, February 29, 2004

No offense - the longer video is beautiful, but I like the shorter one more. :)

Philo

Philo
Monday, March 1, 2004

  Recent Topics

  Fog Creek Home