Lonewolf, re #709
Has the system, in trying to reduce workload, cluttered up task management in a critical phase of flight?
Yes this is a central issue in recent safety debates.
Many of the implementations of automation are, with hindsight, now seen as 'flawed'. We, the industry, design, cert, operators were unable or unwilling to consider the consequences of automation.
See
'Behind Human Error: Cognitive Systems, Computers, and Hindsight', although 'heavy' reading it does cover most to the human, machine, environment, integration issues; and remind us that trying to fix today's problems with yesterday's solutions is unlikely to be effective.
Wirb, re # 710 as above.
However, complexity is not a simple man-machine interface, but includes the operational environment and social norms (need for instant answers, Google / Wiki mentality).
E.g. is the industry accepting more tail wind approaches, higher wind speeds, and thus higher descent rates, less time to react; whilst larger aircraft may have lower control margin, control response, or tail strike margin, due to inertia or geometry, all of which which could increase the overall exposure to the risk of an unstable approach or bounced landing, requiring a GA.
SOPs are useful for standardisation, but like automation, their effectiveness depends on how they are implemented; as a guideline, information, or as a rule.
Similarly for FMA; information and confirmation, but not for control. You cannot determine what the aircraft is doing from the FMA. The information is only feedback as to what the autopilot/aircraft should be doing, but this can only be understood by first assessing what the aircraft is actually doing, - is this what is required for the situation, and is it consistent with what you selected the auto-flight system to do.
Many FMA/SOP implementations focus on reading and calls, inferring safety, yet overlooking the hazards of 'rote' behaviour, - that a mode has been engaged or thrust level achieved but wasn't, or positive climb said but not achieved.