Boeing 737 crashes

[heironymouspasparagus]

Again in a position of having to start a new thread.  Why hasn't someone - maybe one of the airplane people - started one on the 737 crashes yet?

I don't know airplanes, but I do know (embedded) software and this preliminary report shows the exact point where a software problem reared it's ugly head!  Page 26 shows a line called AOA divergence.  Where the angle of attack value suddenly changed in less than a second to something that totally blew the mind of the computer.  IF that were real, there would have been an accompanying change in acceleration to maybe 2 or 3 g's (more??).  Which did not happen.  So the gaping hole right there was that AOA was allowed to operate independently of any knowledge of what would physically occur during an actual event.  And why didn't stick shaker right do something, too?   Let both the people in the front seats know something was going on?

Goes to the overall "robustness" of the software - this was a huge problem with that section of code.  What it seems to show me is that there is some dependence on 'artificial intelligence' type code - code that can't actually be exhaustively tested because no one really knows exactly what it is 'doing'.  Same kind of thing we see once in a while with the self-driving cars.   Until better methods are developed for creating this code, there had better be a very good way to disable those kind of systems without fail.   Which has become even more difficult due to the decoupling of control from mechanical to total electronic systems.  It's a dilemma.  Will not doubt get better, but there will also be a lot of "growing pains" in the process.   I just don't wanna be on the plane that is still in adolescence.  Or anywhere near the self-driving vehicle at the same spot...

https://www.flightradar24.com/blog/wp-content/uploads/2019/04/Preliminary-Report-ET302.pdf

[patric]

Taking pilot decisions out of the hands of pilots, has consequences.

https://youtu.be/H2tuKiiznsY