Hacker News Reader

Comment by c7DJTLrn | Hacker News Reader

This is an interesting engineering problem and I'm not sure what the best approach is. Fail safe and stop the world, or keep running and risk danger? I imagine critical systems like trading/aerospace have this worked out to some degree.

View on HN

There isn't and cannot be a preference to either one. It always depends on what the system is doing and what the consequences would be... Pacemaker cannot "fail safe" for example, under no circumstances. It's meaningless to consider such cases. But if escalation to a human operator is possible, then it will also depend on how the system is meant to be used. In some cases it's absolutely necessary that the system doesn't try to handle errors (eg. if say a patient is in a CT machine -- you always want to stop to, at least, prevent more radiation), but in the situation like the one with the flight control -- my guess is that you want the system to keep trying while alerting the human operator.

But then it can also depend on what's in the contract and who will get the blame for the system functioning incorrectly. My guess here is that failing w/o attempting to recover was, while an overkill, a safer strategy than to let eg. two airplanes be scheduled for the same path (and potentially collide).

Absolutely no idea on what is correct, but I love to reference this article on software practices at NASA[0], They Write the Right Stuff.

[0] https://www.fastcompany.com/28121/they-write-right-stuff

The best approach is to simply print the error to the screen, rather than burying it in a “low level log” which only the software vendor has access to.

They had a four hour buffer until the world stopped, but most of that was pissed away because no one knew what the problem was.

Reply on news.ycombinator.com