zlacker

>>belter+(OP)
Would be very interested in any writeups on how NASA anticipates all the thousands of scenarios that can go wrong up-front and prepares for them. Sounds like there might be some useful thoughts there on how to write more resilient software

>>waihti+L5
I think it isn't about anticipating every possible scenario as much as designing a platform with enough redundancy and ability to measure, turn off/on, adjust, reprogram, etc. pretty much everything.

Part of this is just necessary for ability to learn for future missions. If something fails in space, you want to be able to figure out what happened so that you don't make the same mistake the next time. And you don't have a chance to send a second mission just to "replicate" the problem.

So you do things like build your test equipment into the probe so you can measure stuff while in operation. Or maybe make sure you have a switch for everything so that you can turn something on or off to see if the problem persists.