Nats, perhaps not a misunderstanding, but different points of view.
The details of AF447 are elsewhere.
The A330 technology did not fail; it behaved exactly as designed and certificated. If there are issues in these areas – yet to be established, then perhaps the problems are in the design / certification process, or the continued airworthiness process for an ‘in-service’ aircraft. The key to understanding AF447 is in the differences with the previous events and what exactly we (the industry) knew or assumed before the event.
Occasionally the industry encounters situations beyond the limits of certification; these are where the industry depends, either consciously or not, on the human rescuing the situation. We celebrate many notable successes. Unfortunately we suffer failures because some situations are beyond human capability; this hurts our pride, beliefs, and our professional standards.
The skill of a designer/certification engineer is to foresee these extreme situations and interactions. The greatest difficulty is in the assumptions to be made about human performance.
Aircraft certification is probabilistic, which to date has served the industry very well. Modern complex technology is now stretching probabilistic assessment and testing to its limits, but it’s still workable.
For the human however, this method has never been successful - excepting the probability of ‘1’ – Murphy ’s Law. Humans are not rational, we suffer bias and the effects of environment; yet within these constraints most designs can produce a satisfactory interface between technology, the human, and the situation.
Our training in the use of this interface has to consider all aspects; not just practicing pushing the buttons, but exercising the thinking processes, which in part forms and understands the situation, and thus selects action options to maintain control of that situation.