I have read the report but couldn't get any answer to a basic question I was asking myself from the beginning: why MCAS was not designed to disengage on AOA disagree. I don't see any way such a basic security lock could interfere with the certification process especially as both the MCAS and AOA disagree alert were hidden anyway. This remains a mystery for me, I can hardly believe it's just a simple omission. I guess it was rather because the whole MCAS development process was covered with such a heavy veil of secrecy that nobody would know what others were doing about it.
Seeing how there were a significant number of engineers seemingly *against* the idea of disengaging the MCAS in such an event, I'm also pretty sure this wasn't mere omission, and tend to think this was part of the spec. The whole "we don't really want people to know/notice about MCAS" was probably a big part of it, but there may have been other reasons. Maybe they were pretty sure (again as some people even on here seemed to be) that NOT disengaging it was the best approach in terms of the risks involved. Maybe, or most likely, they were also largely underestimating the potential consequences.
This is addressed directly on pages 107-108, and it's mentioned in context throughout the report, though there doesn't seem to be a clear answer other than that MCAS was considered a non-flight-critical system, with the same failure consequences as a trim system fault, and therefore didn't *require* redundancy, but it isn't clear why Boeing chose not to include a cross-check anyway. I assume that it simply goes back to the Speed Trim system Boeing decided to 'integrate' it with, which is a legacy system that also only used one AOA sensor. This would not be true redundancy, though, just a fail-safe. They didn't include a 3rd AOA sensor for obvious cost reasons, it wouldn't have made a lot of sense to add it for *just* MCAS and not integrate it with the other systems (not that that would stop them), and this likely also would have required some small amount of additional training which they were desperate to minimize.
The report also mentions that Boeing's first foray with MCAS with the KC-46 did include an AOA crosscheck for MCAS, so either their engineers disagreed on the safest action in case of an MCAS fault, or there were different pressures on the MAX that lead to its exclusion.
Might also have been intended, but if it was based on the AOA Disagree alert from the flight control computers, that was faulty too.
Another maybe: IIRC, there was actually a way of disengaging the MCAS manually that would be similar to disengaging auto-trimming (correct me if I'm wrong), so maybe the people in charge of the specs thought that was good enough in case something went wrong, and again maybe they assumed this would actually be part of the pilots training - whereas Boeing's upper management decided to make this feature largely hidden and there was (AFAIK at least) absolutely no significant training or even basic information on how the MCAS worked and what to do in case there would be a failure.
Not just no formal training, Boeing had not even told the airlines or the pilots about MCAS *at all* until the Lion Air crash, and even after it, minimal information was provided.
Boeing's fundamental assumption was that a fault with MCAS would be treated by crews as a runaway stabilizer, which had an associated checklist that would have had the crew engage the stabilizer trim cutout switches - effectively disabling electric trim entirely. However, the two crews did not identify it as such quickly, since MCAS repeatedly activates the stabilizer briefly rather than continuously (and also due to the other alarms in the cockpit due to the AOA disagree [stick shaker, stall horn, IAS unreliable etc]). It's fairly apparent that the pilot's training and expectations were that an electric trim failure would result in continuous activation. Because MCAS was activating repeatedly (also wasn't supposed to happen), the aircraft was allowed to get so out of trim that once they did cut out the electric stabilizer, the elevators alone didn't have enough pitch authority for level flight and the stabilizer was too [aerodynamically] loaded to move with the manual trim.
There were multiple ways the crews *could* have recovered at various points in the scenario, if they had understood the full picture of what was happening on their aircraft, but with limited information both crews ended up missing the path to recovery because they didn't have enough information to understand the failure. Simply enabling autopilot probably would have saved both crews, since this would deactivate MCAS - but crews are generally trained to fly manually when there are indications things are going sideways.