Hacker News new | past | comments | ask | show | jobs | submit login

When you're building machines that need to work in alien environments, you run into a lot of "what-ifs", without clear answers. So you "over"engineer your solution, because it'll probably be cheaper than re-doing the whole mission.

Meanwhile, there's a academic research politics side going on, where a lot of people want their experiment to be done by your probe, the negotiations for which happen oftentimes before the probe has a solid design. You want to be able to tell some subset of people "as long as we don't screw up terribly, you should be safe betting on this", and tell the others "you'll be in the "nice-to-have" category, don't get your hopes up". So a mission lifetime gets decided, pretty much independently of engineering considerations.

They could, in theory, update their lifetime estimates after the project is finished, but there is little to gain at that point: The researchers with experiments in the "nice-to-have" set have already hedged their bets, the engineers would rather have a 50x success than have announced an over-estimate, and nobody really complains when the Opportunity is still truckin' around the Martian surface years later.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: