Re: Question for the experts
I won't claim to be an expert but I have a little experience with sailplane design. Its very rare that just one component causes a crash, usually its a series of failures and shortcomings that are obvious in retrospect but weren't anticipated by the designers. One thing that concerns me about the MAX is that the combination of full down trim and full up elevator would lead to a stabilizer that has the potential to stall (its not just the main wing that can stall -- the stabilizer is also a wing and if that stalls it can make the plane uncontrollable). This is the kind of situation that would never happen in real life -- a pilot is unlikely to wind the trim one way while giving opposite command to the elevator -- but a piece of software isn't that smart, it only sees the big picture if it was designed to do that. So its easy to envisage a situation where the pilots got too much power (for some reason), the plane pitched up, the MCAS tried to compensate with the pilots fighting to keep control of the plane. This may explain why nobody's buying the quick software fix; the accident investigators need to look at the entire picture in order to figure out exactly what happened before the sequence of events can be unwound to find remedies that are true fixes. A patch to the MCAS software might be part of the solution but its unlikely to be the entire solution.
This highlights the difference between 'apps' development and real time systems development. If this Tuesday's patch doesn't work then the worst that can happen is you've got a bunch of annoyed customers. In a real time systems world if your software doesn't work you're quite likely to end up with a bunch of dead customers. There is no such thing as "Rapid Application Development" in safety critical systems.
Biting the hand that feeds IT
