Obviously they are not writing this stuff with provability in mind. Obviously emergent state occurs with numerous subsystems however it's not like the same issue isn't shared in other complex systems... culture of technical debt and commercial shortcuts must have overtaken a culture of testing for a very long time to let this happen. People were raising issues years before the latest 737 Max issues:
Oh yeah. I'm not flying on a 787 [...] I would definitely avoid flying on a 787. - Cynthia Cole, former 32 year Boeing employee and former President, Boeing Engineers' Union
Amusing(?) idea for outlandish conspiracy theory (in the spirit of 'virus all the things!'): bad actor within US bioweapon agency goes rogue, misrepresenting government as effective Boeing stockholder and major client releases COVID-19 to slow travel sector and mitigate impact on company and distribute pain and suffering to Airbus.
I believe that your point stands in the context of a large-capacity commercial airliner, but it is worth considering that in some fighters, this is actually by design. For example, the F-16 is not passively stable--even when the pilot lets go of the stick, the control surfaces are still working to keep the flight moving properly. However, there's much lower benefit when, you know, you're not designing a plane to dogfight with other planes in a life-or-death struggle thousands of feet in the air. Even then, I believe that the control systems are based on voting, where many operations occur in triplicate between various processors and then a consensus is reached for the action.
But building a system like that is difficult, unnecessary for a craft like the 737 MAX, and crucially didn't come close to meeting the same level of redundancy, since I believe the crashes were traced back to single-point failures in the angle-of-attack sensor. Boeing tried to have it both ways, and ended up killing hundreds of people as a result.
Anyway, I thought it was worth pointing out that sometimes fixing aerodynamics problems with software is actually used in airplane design to good effect. The problem is that no one should have thought about or signed off on doing so in this situation.
> Even then, I believe that the control systems are based on voting, where many operations occur in triplicate between various processors and then a consensus is reached for the action.
> But building a system like that is difficult, unnecessary for a craft like the 737 MAX
The 777 was a clean sheet design that incorporated fly-by-wire as it was designed in the late 1980s/early 1990s. Thusly it cannot be operated without the computers running, whereas the 737 can and has.
Its also 4 times larger than the 737, meaning that the more direct way control surfaces are moved by the controls in the 737 is not possible on the 777.
11 comments
[ 3.7 ms ] story [ 32.0 ms ] threadLoss of vertical stabilizer control is very, very bad...
Oh yeah. I'm not flying on a 787 [...] I would definitely avoid flying on a 787. - Cynthia Cole, former 32 year Boeing employee and former President, Boeing Engineers' Union
... via https://github.com/globalcitizen/taoup
Amusing(?) idea for outlandish conspiracy theory (in the spirit of 'virus all the things!'): bad actor within US bioweapon agency goes rogue, misrepresenting government as effective Boeing stockholder and major client releases COVID-19 to slow travel sector and mitigate impact on company and distribute pain and suffering to Airbus.
https://www-theregister-co-uk.cdn.ampproject.org/c/s/www.the...
I have tried this approach and can report that "hope" doesn't accomplish this.
But building a system like that is difficult, unnecessary for a craft like the 737 MAX, and crucially didn't come close to meeting the same level of redundancy, since I believe the crashes were traced back to single-point failures in the angle-of-attack sensor. Boeing tried to have it both ways, and ended up killing hundreds of people as a result.
Anyway, I thought it was worth pointing out that sometimes fixing aerodynamics problems with software is actually used in airplane design to good effect. The problem is that no one should have thought about or signed off on doing so in this situation.
Perhaps a broader problem is that a culture was created that facilitated that.
> But building a system like that is difficult, unnecessary for a craft like the 737 MAX
Why? The 777 has it.
http://www.citemaster.net/get/3c501aaa-39d5-11e4-9cb6-00163e...
Its also 4 times larger than the 737, meaning that the more direct way control surfaces are moved by the controls in the 737 is not possible on the 777.