/sci/ - Science & Math

Shit is probably being over engineered.

It's like a lot of cars now, but using German ones as a frame of reference. Repairing them is a fucking chore, you can't do it yourself and it's made so you can't and at this point only the dumbest of the dumb believe excuses about it needing to be that hard when it really does not need to be at all.

It's more complex because Boeing and the Airliners are massive jews that wanted profit over making a workable plane.

>>10612515
Yes.

t. control systems engineer

>>10612515
>>10612525
No. It's under engineered in the case of Boeing planes.

The entire reason why the 737 MAX is complete shit is because they literally took the old 737 plane with 1960s technology. Which was hydraulically controlled.

Then put a modern fly-by-wire digital controlling system on top of it.

But it wasn't even impelented well. The hydraulics system was still on board so that they could re-use the 737 design and manufacturing plants so they had some weird system where some imputs would be digital and others would be hydraulic. Complete mess of shitty engineering to save every little cent (Boeing trademark)

Airbus uses a complete fly-by-wire system where every input is completely digital.

The main difference between Airbus and Boeing is their philosophy about "who controls the plane".

Boeing thinks pilots should always be able to overrule the computer and be in control of the plane even if they have irrational input into the controls. They get warnings but the computer isn't going to take over.

Airbus is the complete opposite and thinks that most of the error is within human behavior (pilots) and that the engineering is always to be trusted over human behavior. This causes Airbus to take away control from pilots when they detect irrational input.

So what happened with the 737 Max is that some sensors had errors. And since the control system was a hybrid system the pilots moved in the hydraulics directions a lot more efficient. The sensors were giving off errors causing the pilot to overcompensate in only one direction which made the plane lose its lift and crash.

t. Aeronautics Engineer

>>10612542
Nope. In the case of Boeing it's the lack of complexity of their control systems what is the cause. Either build your fly-by-wire from the ground up to be a complete system and let the computer take over. Or have a full hydraulics system. This weird some axis hydraulic some axis fly-by-wire is the most ridiculous shit i've ever seen in aviation.

No, the crash of the 737 Max was caused by the greed of the company. They wanted to save as much money as possible, even at the expense of human life. This is about human behavior, not garbage software.

Their 60's plane design was never made for those big engines, it fucked the ground clearance (hence the engines look sightly oval, they push it to the limit), it also fucked the way the plane behaves, creating a tendency for the nose to pitch up. They fixed these problem with software.

Designing a new plane was too costly for those faggots, when they can just add some shitty software and sell the plane as just another 737, the airlines love that because they don't have to spend money training pilots.

Why don't they just let the 'pilot' fly the fuckin plane!

>>10612578
It's not about being too costly. It is that their competitors made a plane just like their old one but with bigger engines and the selling point for my airlines was that they wouldn't even need to invest in training, which is billions of dollars saved, because it was essentially the same plane.

Boeing was left in a position where if they actually innovated they would get left in the dust because their competitor was selling a much cheaper alternative. If you look at the economics of the issue, Boeing actually did what they had to do. And people had to die. Because regardless of people dying, Boeing got those contracts so Boeing still lived to fight another day. Now they'll spend a shit ton of money, but they got the contracts.

>>10612515

>>10612757
Cringe

They are. Old tech is far better.

based and redpilled

space force when?

>>10613035
Man believes he is the master instead of the servant. We need to initiate the Cult Mechanicus and begin worshipping the machine spirits, or this will get worse

Overconfidence in tech caused the crashes.
They don't even train the pilots because they assume the computer will do the job for them.


Overconfidence in tech has caused a lot of problems and will continue to do so. People don't understand how it works (nor do they care) so they treat it like infallible "magic." Zoomer generation will feel this the hardest because they've grown up with tablets in their baby hands and most couldn't even point out a CPU or explain what it does. I swear as I get older I feel like I'm becoming more and more a Luddite, but I now think my views haven't changed at all. It's the rest of the world that's less and less scared of technology, ignoring the potential dangers and problems they possess.
https://www.youtube.com/watch?v=8yyAgk9der4
sorry for vid, couldn't help myself

>>10613150
im a zoomer and i totally agree with you

Im fairly certain we have less crashes per flight hour and km covered now than before though

>>10612578
>companies want to save money
Wow, what a woke take my dude. Make sure to collect your woke points at the exit.

>>10613181
He isn't being woke he's just saying what the reason behind the 737 MAX failure was.

Boeing should have innovated in hull design by using more composites to improve fuel economy. They should also have put more time into designing the controls since they could have scrapped a fuckton of unnecessary hydraulics which would cut down weight and improve fuel economy even further.

>>10613150
Fuck off if anything Boeing has too little confidence in tech and always lets the pilot overrule the computer. Want to know why Airbus has a lower crash rate than Boeing? It doesn't allow pilots to overrule the superior computer.

Do you really think some random gut feeling pilot acting on whims is better than the engineers working out these problems? Boeing is just using the "Bro just override the computer" excuse to not put a lot of effort into their systems.

Meanwhile Airbus KNOWS everything will depend on the computer so their fly-by-wire system is tested rigorously and as a result the planes fly themselves 99.9% of the time with the pilot being a dead weight. It'a a superior system by far.

>>10612757
GD shill assblasted as fuck that daddy wont buy their broken piece of shit electric slingshot

>>10613200
>brainlet unironically believes that Boeing planes do not "fly themselves" 99.9% of the time
Disabling manual control is a sure-fire way to cause disasters: the auto-pilot systems in place are not nearly as "well trained" (e.g. programmed to handle a very large amount of possible scenarios) as you seem to believe.

What 10 000 lines of code does, a human can do by pure instinct.

>>10612515
Pilots were already doing nothing 95% of the time and just banging flight attendants.
So they had to make them work hrder.

>>10612515

no. female engineers and pilots, yes.

Le'ts not underestimate a pilot's workload.
Especially on approach and landing.
It's not that it's particularly complex when you understand the ropes.
It's just happening very fast.

>>10612757
95% of electricity power is generated through steam turbines to this day, brainlet.
Severe cringe meme.
Please consider getting your wrists hobbled so you can no longer use a computer.

>>10613181
Simply epic reply XD
Epic troll my guy, can I upvote?

>>10613181
Every single plane used to have 3 operators. Radio, Pilot an co-pilot.
Radio operators don't exist anymore, so the two people left have to mess with the radio on top of landing 150+ people.

>>10613401
Let me add:
Think about it as talking on the phone while driving.
That's exactly what all pilots do nowadays.
And it's dangerous for the exact same reasons.

There is trend to manufacture everything as Apple products are to extract most aftermarket value. Environmentalism used to care about this.

>>10613392
Sure, go ahead :-) !

>>10613451
I guess they don't want to save lives then?
What prevents them from retiring an airplane by having all seats being paid by customers, have it crash, and blame software?

>>10613205
>>10613389
Look, having nuclear — my uncle was a great professor and scientist and engineer, Dr. John Trump at MIT; good genes, very good genes, OK, very smart, the Wharton School of Finance, very good, very smart — you know, if you’re a conservative Republican, if I were a liberal, if, like, OK, if I ran as a liberal Democrat, they would say I’m one of the smartest people anywhere in the world — it’s true! — but when you’re a conservative Republican they try — oh, do they do a number — that’s why I always start off: Went to Wharton, was a good student, went there, went there, did this, built a fortune — you know I have to give my like credentials all the time, because we’re a little disadvantaged — but you look at the nuclear deal, the thing that really bothers me — it would have been so easy, and it’s not as important as these lives are — nuclear is powerful; my uncle explained that to me many, many years ago, the power and that was 35 years ago; he would explain the power of what’s going to happen and he was right, who would have thought?

>>10612515
No they are cause by multimonth shutdowns

>>10612515
no.

do we have a retarded hateful moron as a president? yes

>>10612515
if we make a.i. planes then in the future we will have break dancing muslim robots that suicide bomb places.

>>10613480
He's not wrong, you know.

>>10613926
this is now normal

>>10613937
Yikes!

>>10613480
>>10613926
>>10613937
>>10613940
>mfw it's true
https://www.youtube.com/watch?v=Elhyo-_fR0E

>>10612515
Trump is a retard with Alzheimer’s

>>10612757
Literal dementia.

>>10613200
I don't trust a team of engineers to think out every single possible scenario either. I don't trust sensors to work 100% of the time. I don't trust "chinese boxes" (the philosophic thought experiment, not the crap plastic products) to truly interpret real time events and input 100% accurately in every scenario.
In case you haven't realized we still have a few kinks in complex automation. Even simple manufacturing bots still make mistakes.

Our best super computers are still crap compared to a human brain. That's why I still trust a well trained pilot over a completely automated system made by people who've never even flown a commercial jet.

>>10613480
>Takes college class to learn ebonics
>Can't even try to figure out a Trump speech on their own
Typical liberal

>>10613974
Ebonics is similar enough that you can learn it by just talking to black people.

G-d I hate computer dorks. I really should have given wedgies and taken lunch money when I had the chance

>>10613964
The eternal French vs Anglo debate - do you trust intellectual, designed systems or individuals? We know which one wins in the end.

Can also be reframed as the virgin engineer vs the chad pilot.

>>10613212
>What 10 000 lines of code does, a human can do by pure instinct.
Interesting. I've heard the same line from players of various games (chess, go, starcraft, dota). Have you noticed how fast computers are beating this "human instinct" when you give them some watts and a couple guys who get paid to program it?
And guess what, verification of neural networks isn't unfeasible, since some CS tards actually paid attention in their anal class and we now have them exploiting Lipschitz continuity of the networks to prove at least some properties. And there's more to come, although i really doubt networks will be incorporated into autopilots any time soon.
>>10613964
>That's why I still trust a well trained pilot over a completely automated system made by people who've never even flown a commercial jet.
Ironically, human error is what brings most planes down, not software malfunction. It would seem that those people who've never even flown a commercial jet know more about it than a well trained pilot in the end.

maybe

software complexity was a major cause of delays in the f-35 program

>>10614043
That was a self-inflicted wound. C++ is like a suicide for safety-critical project. Especially with coding standard as vague as lockheed had (for some wtf reason).

>>10612515
That's a fair summary of the recent Boeing Max scandal. The Max had an automated system which sometimes made wrong decisions, but the situation could be saved by a human pilot. We know this because highly trained American pilots reported multiple incidents. However, other countries, especially in the developing world, do not train their pilots for as long as we do. So humanity's not perfect. Also, had Boeing used a different system in the Max than Airbus, and Airbus didn't have these crashes.

>>10613200
Airbus doesn't have fewer crashes because they forbid pilots from overruling the computer, it's that the specific system they use is better than the one Boeing opted for in the 380 Max (Boeing planes with the same better system don't crash either). The crashes that happened were from pilots accepting or failing to properly correct the bad computer. There are American pilots who reported the same error but, given superior training, they were able to recognize that the system was failing and correct it. Give an Airbus with its superfluous pilots the inferior 380 Max system and you'll see September Eleventh all over Europe.
The original problem here is not that pilots are obsolete but that Boeing cut corners with an inferior system when a proven reliable alternative was available. I preduct no one will be punished and the executive who made the call will be awarded more money than I will ever see in my lifetime. I hate this society, I wish al-Qaeda was real.

>>10612515
Modern Jetplanes basically fly autonomous, the pilot is just there as an overseer, literally. The Eurofighter runs on a FreeDOS cluster which uses somekind of correction software for flight patterns. Whenever a Jetpilot hits a specific speed and therefore gets exposed to specific forces, the plane takes control and does what the Pilot literally can't and wouldn't even be able to due the applied forces. This is done throughout every possible kind of plane out there, they are highly automatized and pilots just become administrators

>>10614095
But I thought Americans weren't allowed to be good at anything.

>>10612525
1) u r dumb
2) "overengineered" - sounds like you don't understand the Boeing issue at hand
3) computer systems make car problems super-easy to diagnose - including, and proabably especially, German cars.

>>10614076
this. pajeet pilots are the actual problem

>>10614037
you could also argue it's also a matter of responsibility

When a human pilot is in control that pilot knows he's 100% responsible for all the lives onboard. You have an autopilot to 95% of the job and all of a sudden that pilot is only 5% responsible for keeping people alive. The other 95% falls on the shoulders of some ambiguous and abstract idea of "the design team." And you can be sure they're not going to punish a computer programmer for fucking up the code. So, who's responsible for the crash then? Where does the buck stop?

The airline takes a hit to stock prices but they can easily absolve all responsibility by blaming the design team, and the design team can absolve responsibility by blaming the engineer and so on and so forth. Ultimately, nobody gets punished, there is no justice for the negligance that caused lost lives.

In short, nobody is personally responsible for keeping passengers alive anymore.

>>10612515

Nah, it's more like companies cutting corners to save money is the most likely cause of planes crashing

>>10614275
>The other 95% falls on the shoulders of some ambiguous and abstract idea of "the design team." And you can be sure they're not going to punish a computer programmer for fucking up the code.
Idk how it is where you work, but here in yuropoor, if my mistake was to get through the ludicrous amounts of verification (not neccessarily formal, unfortunately) and then caused a crash or malfunction, i would be fired and unemployable in safety-critical.
Am i legally liable? No. Do i get punished hard for fucking up? You can be sure.
But for Boeing it was a matter of life and death to come out with the shitshow they did. Life of a company takes priority over life of few passengers, that's true for every avionics company.
You can be sure their next plane will be properly engineered and not this hybrid abomination.
>In short, nobody is personally responsible for keeping passengers alive anymore.
It might look like that from the outsider perspective, but i've seen people fired for mistakes that were caught early in the verification chain, even before they got anywhere near reality.
And i'm all for making employees (especially programmers) personally liable for damages, not just in avionics. That would filter out the brainlets for sure.

>>10612577
So in your mind, a hybrid digital and hydrolic system where the computer can override some inputs but not others, is somehow less complex than a purely digital system?

>>10612515
Kind of. Generally, no, they're mostly caused by incompetence by the pilots and occasionally by freak accidents of nature. But in this case, the 737 was an overly complicated abortion so Boeing could reuse the 737 again in a new plane. The changes they had to make for efficiency caused the airframe to become unstable and unflyable without electronic assistance or rigorous training. Since training pilots is expensive and would lessen the appeal of the plane, Boeing opted to throw an electronic control system between the pilot and plane so that the pilot could fly the plane like a traditional 737. The system then fucked up.

>>10614037
The problem with AI fights is radio communication. It'd be pretty trivial, especially with how proceduralized flying is made to be to reduce cognitize load on pilots, to have a computer scan instruments, make a decision, and give the correct inputs, since the big problem they have with self-driving cars isn't there; all the information can be easily digitized. But you'd have to completely redo ATC to allow digital communication with AI planes, since trying to get a computer to understand what the fuck ATC is saying between static, slang, accents, and everything else would be a fucking nightmare.

>>10614369
Yeah ATC should have been automatised first, but instead it's like looking at a tribe on New Guinea from your USS Enterprise stationed on the orbit.
Anyway, components to solve that particular problem are already in place, but instead of ATC controllers shouting their noisy stuff into microphone, they would have to type it into a computer.
Now just to find a solution to those damn regulations.

>>10612515
You seem to assume that airplane crashes have become more frequent as complexity of the planes has increased. Do you have data to support that? I'm not sure it is true.

in looking into that, it wold be important to control for the increasing numbers of planes,increasing miles flown, etc.

>>10612515
no it's caused by the incompetence of Boeing