Hacker News Reader

Comment by exabrial | Hacker News Reader

> As various warnings went off in the cockpit, they never reached the conclusion to use the runaway stabilizer procedure.

We continue to put the microscope on Boeing but this is why there is a person in the cockpit: to oversee the planes automated systems. There were multiple failures, and while we pick apart Boeing, we have not done the same to the pilots.

View on HN

> There were multiple failures, and while we pick apart Boeing, we have not done the same to the pilots.

At the time of the crash, it appears that the established procedures were insufficient. In particular, the MCAS failure creates a qualitatively different 'runaway stabilizer' than the previously-imagined failure mode of a stuck-on motor.

In turn, that seems to have created a situation where the pilots did accurately follow the runaway stabilizer checklist, but the trim was so far off neutral that they could not manually control it.

That's why the article here notes that the older-spec second switch, which deactivated just autopilot inputs into the electric trim system, could have helped. However, that is still far from a final answer, since the troubleshooting checklist would still need modification to try deactivating just auto-trim rather than both switches.

Checklist amendments would require far less retraining than creating a new type.

The core of the discussion precisely revolves around Boeing wilfully making it more difficult for the person in the cockpit to oversee the planes' automated systems.

It would not be invalid to observe that the first Lion Air crew dealt successfully with the malfunction, and ask why the second crew did not. Every aspect of these accidents should be investigated.

The first crew had three members.

If I recall correctly, the first crew encountered the issue during cruise, at which point which you have both (a) time, and (b) sufficient altitude that you can afford to push the stick forward in order to reduce the aerodynamic force on the horizontal stabiliser to the point where it will be physically possible to correct the trim without powered assistance.

Pilots encountering the issue during takeoff have neither of those things, with results that we've already observed.

This is probably it, but even if not, there's a large amount of randomness in these kinds of situations, so it's not clear you'd be able to derive anything meaningful at all. Disasters are a highly complex, multi-factor, chaotic phenomenon.

That's because this is HN. Playing the blame game is what we get to do. The NTSB will analyze all factors and if the pilots had any contribution to the crashes that will be noted. This is why aviation has become so safe. If the investigators just looked for blame like we do, planes wouldn't be nearly as safe as they are today.

If a hazard or potential failure mode can be eliminated in the design of a system, it's better to do that than to rely on every single operator to mitigate it individually.

Exactly. Engineering the hazard out of the system is a vastly superior way of mitigating it than relying on administrative/operational procedures for mitigation

The pilots are dead, how do you propose fixing their failures? Or do you mean we should also focus the standard checklists and procedures? Are we not doing that also? I think the article sufficiently tried to cover those as well.

The aviation industry does that all the time. If the pilots have any fault, the NTSB will say so. We should not shy away from that conclusion, either, the point of the investigation is not to find blame, it is to prevent future accidents.

Read up on the extensive reporting on both accidents. They never reached that conclusion because it was absolutely unexpected, as a result of Boeing’s stealth changes to the aircraft.

There was an article recently where they put very experienced pilots in a simulator under the same conditions and they all failed.

But at least one aircrew in the real world actually succeeded.

True, but this seems to sound like, "if we just hire the right people none of these safety procedures would be necessary". If we only had ethical people we wouldn't need laws.

Its more important to focus on the processes rather than the individual people, especially because even great people can have off days.

My point was that we should figure out why one aircrew worked it out, and one did not. If we want to fix the process, we have to consider the humans. I get the impression from this whole debacle that one significant contributor to the accidents was that Boeing's idea of what the procedures should be and what pilots actually do is different enough to be catastrophic. Putting everything else aside for a moment, this has to be a lesson about how procedures are designed and documented. It does no good to decree how something should work if the operators won't, as a practical matter, ever follow those instructions.

Edit: This seems to be unpopular. I am curious how you guys would do it differently? Just document the controls and provide no procedural recommendations? Interesting.

It’s like seeing someone make a half court shot in basketball at the buzzer and wondering why other players don’t do the same. You could adjust the probability to it being a 3 pt shot, dunk, or free throw.

The failures in the Boeing 737 introduced a significant probability of unrecoverable failure. The fact that a few pilots managed to recover does not mean that the ones who didn’t recover did something wrong.

On the other hand, the fact that 2 crews with over decades of safely flying planes were unable to recover from the same plane indicates it had a lot to do with the issues in the plane.

Sorry to confuse, I am not really interested in assigning blame. I am more interested to see what can be learned from this so flying continues to get safer. That will include looking at how the pilots responded to the training material in practice, and figuring out how that plays into designing future procedures, documentation, and training. I get that nobody here seems to have much appreciation for human factors, but I fly a lot. I do care :)

The entire discussion does revolve around human factors; how decisions were made, how the changes were designed to avoid the need for retraining, and in this specific article, designed to hide a change under the guise of preserving standard procedures for the people flying the plane. I’m not sure what other angle you’re after.

If you're point was to consider human factors as part of the overall process, I agree

Debugging flight systems while the lives of 189 people are hanging in the balance is not something that should be encouraged.

Depends on your definition of “debugging”. That’s actually the entire job of a pilot in the plane during flight, to be ready to respond when something goes wrong.

> Read up on the extensive reporting on both accidents. They never reached that conclusion because it was absolutely unexpected

Apparently you haven’t read it either, otherwise you wouldn’t have said what you did. I’m not defending Boeing in the slightest, they should be tarred and feathered over this, but it’s getting really old seeing folks on HN acting like experts and not even knowing the basic facts, especially when the article they are commenting on refutes their own statements.

Reply on news.ycombinator.com