Hacker News new | ask | show | jobs
by foxglacier 470 days ago
The first one tipped over because a sensor failed. I suppose we don't know why this one did yet but why do these sorts of failures bring out the caveman in everybody suggesting completely giving up on the whole concept and doing something "dumb" that doesn't require control systems? Just because control systems feel scary and you might not personally know how to design them yourself doesn't mean they aren't great when they work. Falcon 9 lands upright pretty reliably but even in the early days of that when it wasn't working, people were saying they should give up and use a giant net or towers or something for it to dumbly fall into. It's like seeing a car crash and saying "Why don't we just have giant balloons around cars to absorb the impact when they crash or guide rails along the roads so they won't go off course if the driver falls asleep?". Yea we could but it's both cheaper and possible to do it smarter.

If you were writing software and it had a bug, you wouldn't throw out the whole thing and replace it with a spreadsheet, you'd fix the bug.
1 comments
KennyBlanken 470 days ago
> I suppose we don't know why this one did yet but why do these sorts of failures bring out the caveman in everybody suggesting completely giving up on the whole concept and doing something "dumb" that doesn't require control systems? Just because control systems feel scary and you might not personally know how to design them yourself doesn't mean they aren't great when they work.

Because $150M was/is at stake and "bouncy ball that rolls to a stop and then unfurls" has been proven repeatedly to work?

link
foxglacier 466 days ago
Legged landers have also been proven to work on the moon many times. I suspect the reason for this thinking is that engineers like to find technical solutions to problems but when the only information you have is "it fell over", the only apparent technical solutions are to use different concepts that can't fall over. Once you know more specifics of why it failed, you'd be thinking of ways to prevent those, like maybe correct the software bug or implement redundant sensors or whatever. The troubles with these landers won't be something as difficult as "Reliably balancing a slender object upright is beyond the cutting edge of technology" since that problem has been thoroughly solved in other contexts like Falcon 9 and Segway.
link
foxglacier 469 days ago
But at some point we're going to want to do it more efficiently or upright, especially if it's got people on it. Why not work that out now?
link