Hacker News new | ask | show | jobs
by squirrel6 894 days ago
True, I’ve heard and read that many pilots prefer the cockpit config over the 737. Boeing needs to just start over.
2 comments

> Boeing needs to just start over

Would that even be possible? How long time would it take them to complete a completely new design? And then how long time from the design is finished until they can start delivering planes?

The whole reason the 737 Max series exists is to avoid having to start over. By maintaining "backwards compatibility" type approval with the regular 737 then the existing massive pool of 737 pilots can jump straight in it and start flying.
I always thought that was the most idiotic part that the FAA accepted. Like, I get that there has to be leeway for similar configurations of the same plane so they don't have to go through the process for each. But it's been so clearly abused that it's ridiculous now.

This: https://upload.wikimedia.org/wikipedia/commons/9/94/EM_N323A...

Is so obviously not this: https://upload.wikimedia.org/wikipedia/commons/3/32/Alaska_7...

I'm not sure a simple picture is enough.

The Airbus A330 and A350 share the same type rating. (At least some variants; I'm not an expert.)

  The Airbus A330 and A350 share the same type rating.
In fact, they do not in the United States.

https://registry.faa.gov/TypeRatings/

In Europe, they do:

https://www.easa.europa.eu/en/downloads/11737/en

However even within the same type rating you may have additional training or restrictions. For instance the FAA prohibited Southwest from using the same pilots across three generations of 737 (Classic, NG, MAX), so Southwest ditched the Classics when they bought the MAX.

A big advantage Airbus has is their more modern designs are full fly-by-wire, so they're able to more practically compensate for the handling characteristics in software.

As far as I know (not a pilot or aeronautics engineer) MCAS was more or less an attempt to do the same to a non-fly-by-wire plane, and we see how that turned out...

The picture is just to show how far the plane has deviated from its original design on a colinear (e.g. same price/size target among the new generation) scale. If you want more depth into the pains Boeing goes to not trigger a "new design" recertification, there are plenty of articles on it released during the initial Boeing MAX MCAS forced crashed saga, such as ArsTechnica's series of articles.
How bad can retraining possibly be?

Is there something in that process we can address instead?

Is not starting over even possible for Boeing a this point?
Since the MAX program was created specifically to avoid going through a full certification process, I'd say they're going to triple down. On the spectrum of MAX shenanigans, a hole blowing open in the fuselage barely ranks. The plane didn't even crash this time! And it is not yet clear that Boeing intentionally deceived the regulators; they might have paid Spirit to do it.

It's great that the FAA is looking into this. Who is looking into FAA, though? Is this another case of institutional capture?

Using a trusted and proven design (eg: 737 NG) is fine and even beneficial. In fact, reusing older designs in and of itself is fine.

The problem is Boeing clearly doesn't know how to manufacture good aircraft anymore. Whether it's a rehash or clean sheet design, the result won't change until Boeing is run through the washing machine.

Except it’s been so heavily altered that they had to use software (MCAS) to correct for the flight characteristics induced by the larger engines. And we all know how that worked out.
Yes but this is not a huge problem with the airframe itself. The tendencies without MCAS are fine. The pilots just need to be trained for them.

The problem was that airlines want to skimp on the training.

The same thing happened with the engine management. Boeing doesn't want to introduce an engine warning system because it would mean pilots have to be retrained. A lot of these barriers aren't part of the physical design but the industry as a whole being extremely wary of training. Probably as a result of cheap low-cost carriers emerging.

Not saying Boeing is a great manufacturer but it's not the only issue at play.

> The tendencies without MCAS are fine

FAA seems to disagree..

> ([0], scroll to bottom) Following publication, an FAA spokesman on January 11 provided the additional statement regarding Administrator Dickson’s comments around Boeing’s selected path for meeting Federal Aviation Regulations for the original implementation of MCAS on the 737 Max. The Air Current sought and received comment from the FAA in advance of publication and that is reflected in the story above. We provide the latest response in full:

> The fact of the matter is that the FAA — and the other authorities — determined during a 20-month review that MCAS was a necessary part of the flight control system. We all reviewed and approved the changes to the system as part of the recent certification.

> This is directly addressed in the 100-plus page report [1] that was filed with the Airworthiness Directive. We made it clear that the aircraft would not have been compliant with the stick force and G requirements and higher angle-of-attack without MCAS or some other type of mechanism. The FAA does not tell applicants how to design planes, so the choice to develop MCAS was an engineering decision on the part of Boeing. The FAA’s role is to evaluate and approve proposed designs against the regulatory requirements.

[0] https://theaircurrent.com/aircraft-development/mcas-may-not-...

[1] https://www.faa.gov/foia/electronic_reading_room/boeing_read...

I’ve just finished reading Flying Blind by Peter Robison (recommended elsewhere on HN recently), which had engineers arguing that the plane’s flight characteristics in extremis were not acceptable, and that the correct fix was by physical redesign, not software.

Another point the book made was that since nearly every other modern aircraft has far more advanced avionics, it causes the issue of pilots who are used to computers handling a lot of tasks suddenly having to do that themselves. This is probably not as much an issue with carriers who solely use 737s, like Southwest, but I can definitely see the issue otherwise.

> The problem was that airlines want to skimp on the training.

The airlines weren't entirely at fault; Boeing misrepresented the necessity of serious pilot training.

The FAA - not Boeing - eventually mandated simulator-based training for pilots transitioning onto MCAS-equipped aircraft.

It clearly hasn't been fine and beneficial.

You can tweak and upgrade an aircraft to a certain extent, but they've pushed the 737 design too far.

737 MAX flies just fine so long they are manufactured properly and we stop pretending it's a 737 NG with pilots trained as such.

Y'all think 737 MAX as an airframe can't fly safely, and that's straight retarded. I sincerely believe Boeing should face bankruptcy from their many recent failings, including 737 MAX, but even I will still say 737 MAX are fine birds so long as common sense is applied.

787 was announced in 2003, and the first shipment was in 2012. Obviously we have no idea what the extent of work was prior to the announcement.

A220 was officially launched in 2008 (with like 3-5 more years of some degree of development prior to that) and first delivery was in 2016.

A350 was officially launched in 2004 and first delivery was 2014.

Expecting 8-12 years would be reasonable.

It’s definitely “possible”: Clean out the outsource-advocating management from C-suite to middle, and replace them with engineering managers.

This was always a leadership issue.

Whether the will is there? The fear of not dying while flying 4 miles high is a powerful incentive, and the public/market had spoken

What other (poor) outsourcing can we disincentivize?

The 747 was designed in 65 and doing its first commercial flights 5 years later.

https://en.m.wikipedia.org/wiki/Boeing_747

Yes, because redoing something from scratch usually reduces the amount of problems. Do you really think that?
eerrr...

I can only speak from experience with software, but several times I've been in projects trapped with implementations that take months to incrementally update and then seen clean sheet implementations blow them out of the water in a matter of weeks.

In one case I saw an implementation of a radio system that met 100% of the technical requirement that a staff engineer wrote as a side project in six weeks. The corporate implementation couldn't meet the spec, had taken 2 years and was being developed by scores of engineers.

Interestingly the corporate implementation was supplied to the customer, caused a huge reputational dent and was ulitimately scrapped... while the staff engineer was dismissed!

>I've been in projects trapped with implementations that take months to incrementally update and then seen clean sheet implementations blow them out of the water in a matter of weeks.

Firefox vs. Chrome in a nutshell.

Not the example you are looking for. KHTML -> WebKit (for a while, that was the engine of Chrome!) -> Blink.
737 might be at point where redoing it from scratch might be sensible. You can only iterate from same starting point that is over 50 year old design so long.

At certain point starting from new will make more sense. And it might be now.

The problem is that 737 design requires significant redoing to make it inherently safer (either structural redesign, or switch to Fly-By-Wire), with all options involving significant certification hurdles.

At some point the question becomes if there aren't bigger wins possible with new airframe.

Or at least resurrecting 757 and maybe making a shortened version.