Originally Posted by ridefar

I am with you all the way CC. Which is why I asked " In which case the pilot training should not only be for how to shut it off when it activates incorrectly, but how to fly the plane when it didn't activate but should have?"

I get that it isn't a binary condition. The plane can be flown without it. However, my point (perhaps poorly phrased) was that it is off and needed, and if it is more likely to be off and needed than prior to the fix, then shouldn't training also be covering the off and needed scenario. In addition to the on and trying to force the plane to crash scenario.

If we really wanted to be precise, it would be interesting to know what percentage of problem scenarios are addressed by the "off and needed"? Because if it is only 1%, and the other 99% are "activated and not needed (and trying to cause a crash)" then I would ask: is it worth the risk? What is the likelihood that even with good training it will still result in a problem? OTOH, if "off and needed" is 95% of the scenarios, doesn't it seem like a good hardware fix is a better idea than just a software fix? (And yes I know that strays in the Boeing vs. Airbus control philosophy...)