Hacker News new | ask | show | jobs
by nostromo 2657 days ago
Seems like anti-stall software shouldn't be killing more people than actual stalls.

When was the last major airline downed by a pilot error stall anyway?

3 comments

In german we even have a word for this: Verschlimmbessern [1]

Making something worse in an attempt to make it better. As an engineer I'm often concerned about committing such a fallacy.

[1] https://en.wiktionary.org/wiki/verschlimmbessern

Anecdotally, I was speaking to a friend who has a pilot's license, and his thought was that a well-trained pilot should know how to adjust due to the failure, but the system seemed counter-intuitive. I think it does at the very least beg the question of whether this feature has already caused more problems than it would have likely prevented. Would commercial pilots have to report when a plane stalls/would the airline keep track of such data?
The pilots of both Air France 447 and Colgan Air 3407 managed to stall perfectly serviceable aircraft in 2009, killing 277 combined.
Exactly my point -- we've just seen 346 people killed by (presumably) an anti-stall system -- when it's been a decade since we've had the problem this system is meant to prevent.
If the 737 MAX anti-stall system has a flaw of course it should be fixed. That's hardly an argument against anti-stall systems in general. The accidents prevented by anti-stall systems don't make the news.
Yeah, it is tough to assess without knowing the likeliness of a stall, and the likeliness a stall causes a crash. Otherwise, we run into the problem of having such a small sample size.