737 problem explained

The regulators should be sufficiently competent to be able to ask real questions about emperor's new cloths,

In self certification, the authorities just check that the manufacturer has properly processed those issues identified by the manufacturer.

People too closely related to a product has hard to think out of the box, thus it is a good idea to have some outside person looking at the situation with fresh eyes.

In order to limit the amount of information leaking to the competitors, the regulators are just the right position to do this analysis.

No, I expect a computer to do it because it doesn't care about G's.

Actually he would have to do coordinate substitution to compute the non-rotating system.

It is sufficient because all the program has to do in the presence of conflicting data is -not- put the plane into a dive and sound a stall alarm instead. Even less well-trained pilots have had to react to a stall alarm before.

I believe the purpose of the MCAS system was not so much to prevent a stall as to compensate for the handling characteristics of the plane, so it operates in the background. There are already stall warning systems on the plane.

Any mistakes you attribute to the market. This is not an example of letting the market decide. It's letting the engineers decide.

IMNSHO, if C++ is the answer, then you ought to change the question. If anybody doubts that, have a look at the C++ FQA. I'm particularly amused by the section on const correctness

But C++ !=> OOP, and OOP !=> C++.

There are several /good/ OOP languages with simple well-understood syntax and semantics. You /can/ understand what they are going to do :)

ll as to compensate for the handling characteristics of the plane, so it op erates in the background. There are already stall warning systems on the p lane.

yep, it has stall warning and stick pusher, MCAS was just there to make it behave more like a 737NG so they could say it was a 737 and the pilots could fly it with a 737 type rating

I don't disagree, but in practice that means the regulators would have to be at least as competent as the manufacturers.

Proving a specification is "correct" is more difficult than implementing a specification.

An additional issue is problems with the autothrottle. Greg's Airplanes and Automobiles has just posted a discussion of faults with the Boeing autothrottle system that helps explain why the airplanes crashed with such high velocity. This made the manual trim wheels so hard to move. So it was not just one problem, it was multiple flaws that killed so many people.

"737 MAX it's NOT just the MCAS"

On Dec 21, 2019, Tom Gardner wrote (in article ):

Actually, it?s undecidable, and somewhere between intractable and impossible. This is the Halting Problem in Computer Science:

The Halting Problem is also NP-Hard, which means that in practice it?s impossible to solve.

Joe Gwinn

Not necessary unless two separate and independent designs are attempted.

It often useful to have some outside person pop up into a design review meeting and ask "dumb" questions to the designers. Some of the questions are of course ignorant but in some cases such person will

about. Is this person a regulator or someone else is not so critical, it is important to have broad understanding of issues but not too specialized knowledge,

There have been crashes in which both the captain and FO has selected the same faulty source. This is a problem especially at night or in a cloud, During the day and below the clouds, many conflicting data issues can be checked by simply looking out of the window,

I don't think they've recouped the total development and production costs on the 787 sales, yet. 15 years of development and 15 years to pay off is a tough way to earn a reliable income these days, a lot can change in that time.

The goal was as I understand to allow quicker but fuel-efficient point-to-point flights bypassing hubs. Maybe a good goal for the industry of the late 90s, nowadays fuel is relatively cheap compared to airline revenue which has increased, and as for bypassing hubs the airlines have found passengers are willing to suffer about any abuse if you can keep the fares down.

Yup, that's the theory.

The question is how to achieve something in practice :)

That doesn't follow, if you think about it.

Agreed - with an important proviso. It requires that both parties have the same overall objective, and that push /doesn't/ come to shove.

Problems include whether marketing or engineering is more important, and which design/process/strategy is best.

All that means that, sooner or later, push /ought/ to come to shove.

Que?

The engineers didn't decide; they were overruled by the marketing considerations.

Apart from that, I don't attribute all failures to the market, and more than I regard the market as wise or infallible.

The market, like other mechanisms, has strengths and weaknesses, and will optimise *something*/ The question is whether that something is the most important thing to be optimised.

Safety vs shareholder profit is a key problem area. I presume you do understand the Ford Pinto saga.

Two problems: - see Rick C's response indicating you are fundamentally wrong about MCAS's purpose and operation - sensor failure can cause the plane to think it is climbing when it is diving, or vice versa. Famously such sensor failures in human brains have killed many pilots :)

Yup.

In the Kegworth 737 crash the pilots cut the power to the functioning engine and pumped more fuel into the malfunctioning engine.

Another problem was that the trim cutout switches could only cutout ALL electric movement of the trim wheels (although there are 2 cutout switches, they essentially perform the same function). So the only option after using the cutout is to use manual cranking of the wheels, and when the automatic system has already moved them a lot by that time, the loading on the tail surface is so high that the manual movement of the wheels is no longer possible. That is the problem the Ethiopia Airlines pilots faced.

There should have been two separate cutout switches: one that cuts out all automatic trim (including MCAS), and the other that cuts out the electric trim as controlled from the yoke. They could have cut out only the automatic trim and then corrected the situation using manual electric trim.

And indeed there are two switches, but apparently this detailed function has been removed long ago (before the introduction of the MAX and MCAS)

The AoA sensors are notoriously failure-prone. In the first accident the sensor was actually replaced the day before.