This reminds me of the story of when Lisp was still used at JPL.[1]
"Debugging a program running on a $100M piece of hardware that is 100 million miles away is an interesting experience. Having a read-eval-print loop running on the spacecraft proved invaluable in finding and fixing the problem."
The article frustratingly doesn't say whether this is a simple and correctable RAM one-bit error, or a permanent defect in a storage device. It's possible they don't know yet, but ideally, the A computer can be brought back to full functionality and serve as a backup to computer B.
The real irony will be if the defect was created by a stray neutron originating in its own radioisotope power supply.
In general, single-event upsets (SEUs), which are a single change to usually a handful of bits, from a cosmic ray, are far more common than permanent damage.
It's endearing, but I've always found that a bit facile. The idea that even a self-aware rover would want to go back to Earth is blasé anthropomorphism. The rovers weren't created to live in our gravity, or our atmosphere, or our ecosystem; they were created to live on Mars.
I find it more touching to think that we sent them home.
Sure it is anthropomorphism. That is just one of the ways in which a bunch of loser tree-monkeys make sense of a world in which they can send a robot to a different planet.
Humans are suited to our gravity, our atmosphere, our ecosystem, but we still want to go leave Earth.
If you're going to anthropomorphize the martian robots, maybe their ultimate goal isn't to see humans again, but to leave Mars and explore the rest of the universe.
I recommend reading Henry S. F. Cooper's _The Evening Star_, about the Magellan probe used to RADAR map Venus. They diagnosed bad RAM and a race condition in the OS that was causing it "safe" the orbiter. It's a good narrative.
Yes, it can be cosmic rays, or maybe just a task that didn't exit properly
One thing I find different about Mars Curiosity is how results are slow to come. In other Mars missions it seemed they would expedite the experiments (but of course, other robots had a much shorter 'due date')
"Debugging a program running on a $100M piece of hardware that is 100 million miles away is an interesting experience. Having a read-eval-print loop running on the spacecraft proved invaluable in finding and fixing the problem."
[1] - http://www.flownet.com/gat/jpl-lisp.html