The memo from Emirates about not setting altitude in system to airport altitude is so weird. I really hope that's not the actual message Emirates is putting out there.
It doesn't matter was in system. You run your checklists and configure for your own flight / your own departure etc. Memo should be to training / recruitment - where did you find these pilots to fly a 777?
So they obviously didn't pre-flight the flight. And they don't seem to have used their takeoff checklist either? That should get them through flap retract and 2,000 feet or so.
Takeoff checklist might have rotate at 160 or so. Then you get to positive climb and V2. You then clean up the plane (flaps / gear etc), maybe hit 2,000 feet / 210kts?
This is all checklist stuff. I've no idea what's Emirates has as checklists, but this is like 101 entry level flying stuff. What was pilot monitoring doing during all this?
They still on ground at 215 kts? 262 kts at 175 feet?!
It's unbelievable that the pilot's didn't realize they were doing 215 knots while still ON THE GROUND and only decided to pull back on the yoke once they were nearing the end of the runway. They should have at least seen the end of the runway coming up and wondered why they weren't airborne yet. The ends are colored in amber and red as it gets closer to the end, it should be pretty obvious something is wrong. Pull back and get airborne, you're far past the point you are going to be able to stop.
Where's the basic pilotage here?
I'm a US CPL holder and there are numerous basic airmanship systems in place to prevent this. For one, you note your acceleration rate and how far down the runway you are. Then you have V1, V2 speeds where you rotate.
How you get to this level of incompetence flying a 777 is far beyond my understanding.
Pilots are getting far too comfortable flying what the computer is telling them instead of flying the actual aircraft. You don't fly the flight director command bars when they are telling you to stay on the ground and you're doing 215 knots. You ignore them and figure out what went wrong when safely in the air.
This applies to most catastrophic failures, not just in aviation but in any other place that puts reasonable effort into preventing them. Usually there are plenty of safeguards, which means that most of the times, when something goes wrong, one of the safeguards prevents a catastrophic outcome and you don't hear about it.
But when you do hear about it (due to the catastrophic outcome or a very near miss), it'll usually be a tale of many mistakes, very egregious misbehavior, multiple failures etc. - because it generally has to be to cause a bad outcome or even get close to it.
Keep in mind that despite this, nothing seriously bad actually happened here. It was a near miss that needs to be addressed, of course, but nobody died, there likely was no serious equipment damage, etc. Ultimately, the system worked.
>>Keep in mind that despite this, nothing seriously bad actually happened here.
This is the exact wrong way to think about it.
With substantial experience with managing risk, the most important thing I know is that you MUST UNDERSTAND the difference between managing risk with knowledge, technology, planning, and skill, vs rolling the dice and getting away with it.
With risk management by knowledge, technology, planning, and skill, you can make an ordinary practice of very risky activities very safely.
But when you are rolling the dice and getting away with it, you may get away with it for a long time, long enough to think you are safe, but it WILL bite you, and hard.
"Near misses" are very clear indications that your risk management has a failure mode you didn't account for - and you are now in the zone of rolling the dice. If you do not attend to that, you are about to get bitten.
So, just because no one was hurt or died — this time — does NOT mean you can be relieved.
As I said, I do acknowledge that a near miss still needs to be taken seriously, but the point is that in terms of actual harmful consequences (not "might be could have been" hypothetical ones) from this incident, there was either basically nothing, or absolutely nothing.
It definitely was a failure in the sense of "we need to do failure analysis and make sure we don't get this close again", but it also was a success in the sense of "ultimately, something prevented a catastrophic outcome" (and studying that, and potentially trying to replicate those positive aspects at other airports, is also something that should to happen).
>>"ultimately, something prevented a catastrophic outcome"
The most likely "something" was Pure Dumb Luck, just a random combination of factors that happened to work out favorably. Of course, on the chance that there was some particular factor/event/action that could be replicated and reliably prevent the disaster, then sure, we should have a look.
I'd definitely make a good survey for that, but the focus really needs to be further back in the chain of events to find out what thing(s) failed.
And of course to be in any way relaxed about this because they weren't scraping up parts from 300+ bodies this time is folly.
Crew Resource Management should have kicked in and the Pilot Monitoring should have taken over on rotating the moment the Pilot flying did not steer after V1.
That this did not happen, and nobody actually concentrated on flying the plane, while the other party went through the corresponding checklists and diagnosis - thats cultural rot and a major issue.
Yup, and that cultural rot is also a key failure to manage risk.
Just because you are [using knowledge, technology, skill, & planning] to manage risk once does NOT mean that the job is done. It needs to be constantly maintained.
If it is not constantly maintained, you are slipping back into the zone of [rolling the dice and getting away with it]. It will bite you.
What comes to mind is the last video from the very experienced skydiver which was normal until chute-opening-time, then got very chaotic and ended - he'd become so comfortable/complacent that he didn't do his checklist, jumped with only a knapsack instead of his chute. I'm sure he started out checking everything well. But getting casual and complacent killed him.
Flying was one of my most irrational fears until Westjet introduced me to a pilot. He went into a truly massive amount of detail on their systems (both manual and automated) that prevent things like this.
I’m not a pilot (I’m still very scared) and it’s been a long time so forgive me if my language isn’t correct. However, my understanding from that conversation is that this sort of thing is a system failure. Nobody got hurt and there may have been no damage, but they would consider this an overall system failure.
Several people made mistakes here. There was the pilot who made the near fatal entry. And many pilots failed to spot the error. They didn’t perform their pre-flight properly and would have had to forget a checklist.
They apparently didn’t even figure it out when the runway lights changed to signify “you done fucked up.” (Can a pilot weigh in?? Is that an aviation term?) :)
This system failed. It’s pure luck that everyone survived.
It definitely was a failure in the sense of "we need to do failure analysis and make sure we don't get this close again", but it also was a success in the sense of "ultimately, something prevented a catastrophic outcome" (and studying that, and potentially trying to replicate those positive aspects at other airports, is also something that should to happen).
The real problem here is that many overseas governments and airlines don't hire pilots based on competence, but on their coziness with the ruling class.
Events like this are not Boeing's fault and will continue until foreign airlines can get rid of corruption.
I think the OP was focusing on the consequence, not the probability. Since risk is often defined as consequence x probability, I think you are both talking about distinct but related things
I agree, that's covered by the probability part of the risk equation. Even if you got lucky and nothing bad happened, the negligence still increased the risk by increasing the probability that something may have happened.
> Pilots are getting far too comfortable flying what the computer is telling them instead of flying the actual aircraft.
I'm not a pilot but this stuff does concern me.
If you Google around for GPS related accidents with cars the results are pretty mind boggling. There's a number of cases of folks driving straight into a lake because GPS told them to go that way. This plane incident reminds me of that.
I've been wondering why we haven't really had any advances in our voice commands for nearly a decade - things like zooming and such are actions I've long wanted to be able to perform.
Even that rotate map.. when your car is shown at the bottom and map is hard locked to current heading, when navigation is about to do a sharp right you end up with zero visibility of the rest of the route as they still haven't bothered doing a projection and modifying rotation dynamically - they're not even trying.
All your examples are exactly the kinds of things that an intelligent system should be able to respond to, in order to be useful.
Getting that haircut ordered is the _last_ thing we need.
Because it takes a dedicated team reviewing commands all day, every day, and dev team implementing event handlers, coordinating with other product teams. It takes lots of dedication.
I used to run a chatbot that screened people. We thought we were asking simple enough questions, and yet two million chats in, we still discovered at least a Dozen unexpected/misdetected intents that required care. A day. And here is the problem: management will ask: "so if we add handling for 'rotate the map 30 deg', how many people will appreciate it?". The answer is usually "one person every 400k sessions". For an MbA trained manager that means there is no reason to do it. And its the same with every other similar discovery.
Running an NLP based tool requires staffing a 10-20 people full time, telling them to keep the quality high and leaving them the fck alone. Thats anathema for most companies that proclaim to be "metric based". Google is a prime example of this behavior.
> The answer is usually "one person every 400k sessions".
But you control that to some extent. If more people are aware of the command and believe it will work (I wouldn't) then you'll probably get higher incidence rates. I wouldn't believe it would work because like the person two parent comments above, less than half the novel crap I try works. So I try novel crap less and less frequently compared to things that I know will work.
I just disabled Google assistant and all of its functions, it's beyond useless and probably the most frustrating thing I have ever used on my phone. And I don't mean just the voice controls, but all the "helpful" tips and suggestions are just bad and I can't stand them.
When driving I don't bother to use Siri for anything other than "call <person>" or "text <person>".
The most hilarious example of how bad Siri is for me, is that literally 100% of the time that I'm playing something on Apple TV, if I say "Hey Siri, pause the TV" (or any variation, calling it Apple TV, including the name of the room, etc.) it always replies that there's nothing currently playing. And if I tell Siri on my phone to play on the Apple TV, all it can do is start Apple Music playing a random song on the TV. It can however respond OK to "resume on Apple TV" which restarts whatever I was watching that had been paused.
OK Google is awful for me, it repeats back exactly what I said but rarely interprets it, at all. "Don't know how to help with 'navigate home'", "'send a message'", "'find a supermarket'". Bleurgh. This is via Android Auto. Next minute it's reading me a text and asking if I want to reply by "SMS or WhatsApp".
None of these systems show you enough 'under the hood' to know why it's failing and so react. Clearly, here the voice recognition, tts are working well; some optional visual feedback would be great - like show on screen a list of responses:
- Keyword heard
- OK Google responding
- Voice1 recognised
- Input "navigate home"
- Input sent to maps
- Maps didn't understand
- Reporting failure
I guess Google know all this, at the time of the query, but as they don't know what I'm trying to do (though 'navigate home' is pretty clear, perhaps I was actually saying 'buy a gnome') ... doesn't seem like there's a proper feedback loop to help direct/correct the system?
"OK Google, navigate home" has never failed to work for me for a couple of years now. Maybe you had opened the app to a profile that doesn't have your home address configured? Maybe you had been logged out?
The sole thing I use it for is setting alarms/timers, or starting the process of a calendar appointment. Those are the two use-cases I've found useful.
Say "navigate to X" also can kinda work with less fumbling than typing half the time but it's a bit messy.
Last time I tried, I couldn't even figure out how to exit navigation using voice. I'm sure it's possible, even easy, but the combination of not knowing the right incantation, car noise, language issues (does it expect German or English commands?) and overall stress meant I just ended up muting the sound system, which is a physical button.
Would you like to see a map of the nearby roads with no lables on it? I fyou zoom in to the point where a road is half the width of the screen I'll deign to add a label on one of the little side roads you don't care about.
Never will I ever tell you what the arterial route you're currently on is called though. Sucker.
It's not just Google Maps though. Also Apple Maps, Music, Podcasts... I find it so frustrating to use voice commands on my phone in the car that I mostly just don't even try. I set things up before I head off as much as I can and mostly stop somewhere before I make changes.
I've just gone from a 2004 Corolla to a 2020 Serrato. It's a low end car but it has lane assist, braking detection when a car slows in front, cruise, speed limiter, etc. Combine all those settings with google maps and it easy for me to see how you could get complacent driving a new car and make mistakes.
All these features work well most of the time, but they all notably fail on, still frequent, edge cases.
Yes and as with Tesla "FSD", supposedly the driver is to devote full attention to monitoring the car whenever activated. But this is silly because as these tools get better, the driver will inevitably monitor less. But it won't any time soon (ever?) avoid random errors every now and then. But now these will be catastrophic failures.
Driver training should be updated to include what to do when self-driving modes are used (ie what do you actually do to supervise the vehicle, which checks to do, etc).
I don't really see how the needed monitoring would be more interesting than fully driving the car, maybe with a few convenience and safety assistances. Planes can use autopilots because there are not litteraly surounded by traffics 2m away. If a car needs monitoring, and they will probably need it for a very long time, you should as well drive it.
> If a car needs monitoring, and they will probably need it for a very long time, you should as well drive it.
This is the crux of the intractable problem of AI/self-driving marketing vaporware, and why until this nonsense is shut down, the situation will probably have to get scandalously worse before it gets better (better meaning regulating fake "self driving" off the road as a public safety travesty).
> But this is silly because as these tools get better, the driver will inevitably monitor less.
And for good reason. The main benefit to FSD, as I see it, is the ability to do something more productive with my time while getting from A to B. Most wealthy people end up paying for drivers not because they want an exceptional driver, but because once you reach a certain threshold of income it is more profitable to work (phone calls, laptop, w/e) or sleep, and pay to be driven everywhere than to drive yourself. FSD lowers that income threshold tremendously and could be a significant boon to overall human productivity. If it ever happens.
>is the ability to do something more productive with my time
Seems to miss my point, I may have not stated clearly. Yes I understand the sales pitch of self driving cars stated above...the "marketing" / "hype" of AI/self driving cars/ Tesla FSD evokes the idea of completely offloading the responsibility to the car and doing whatever you want with your time during commutes, as if it was public transit.
But my point was anything less than this is a complete waste of time and useless as a product because until you are completely hands and attention off, you have to be completely hands and attention on/ready...no matter how good or near actual FSD it gets.
Because 99.9, or 99.9999, or 99.999999% effective is effectively 0 percent effective given the catastrophic outcome of any error if you aren't devoting full attention to monitoring the car.
But aside from all of this theorizing, the actual product is NOWHERE near high 9's reliable. As it gets better, attention will drop but errors will still be pervasive. And with this as backdrop, the fact the marketing has ever even hinted at being general-AI type FSD (the name itself for example) is a scandal. The videos of people riding in the backseat, or any other instance of not monitoring the car fully because they think it can drive itself. I predict there's going to a be an endless stream of accidents from these cars then FSD will be shut down until fully proven, which won't be technically possible for a long long time if ever in our lifetimes.
> But my point was anything less than this is a complete waste of time and useless as a product because until you are completely hands and attention off, you have to be completely hands and attention on/ready...no matter how good or near actual FSD it gets.
Agreed. If I got a Telsa tomorrow I wouldn't do so with "FSD".
> Because 99.9, or 99.9999, or 99.999999% effective is effectively 0 percent effective given the catastrophic outcome of any error if you aren't devoting full attention to monitoring the car.
I'm not sure what effectiveness you're measuring. There will always be fatal accidents with heavy vehicles moving at high speeds. It's more of a regulatory issue where the accident rate of auto-pilot systems is sufficiently lower than human drivers (because I believe it is already lower in many situations, just not sufficiently lower or ubiquitously lower) and we decide its good enough to no longer require monitoring by the driver.
Edit: I guess its even more complicated than that. Using hypothetical numbers, if the accident-free rate of a human is 99%, and auto pilot is better at 99.9%, but combined (auto-pilot with human monitoring and rare intervention) is 99.999%, can we justify allowing no monitoring in that situation? Can we expect people to follow the law or will it turn in to more of a speed-limit type grey area.
Not all errors or accidents are created equal. As someone else mentioned in this thread the nature of human error is qualitatively different than the nature of seemingly random errors we see with self-driving. Code bugs can produce any scale of error. Humans do no in any given situation have open-ended likelihood of making any errors.
Search youtube videos of Teslas if you haven't seen them yet. The cars will randomly swerve into parked cars, inexplicably steer out of turns, aim directly at medians on the highway. Human errors on average do not equate to this if we want to compare human error "rates" to machine. What is the rate of a random catastrophic error? Machines are way way too high for the foreseeable future, and because of the nature of software bugs they arguably always be higher than humans. The scope of what AI has to be to achieve "general" driving on open-map roads with near infinite variables nearly equates to "general" AI itself.
You can't allow "halfway" monitoring the car. It's all or nothing. It's always monitored or it's not monitored because errors can happen at any time. But anyways, humans aren't robots. Their attention basically not focused if it's not focused completely... multi-tasking is false, and humans already fall asleep fairly easily driving manually. If they have to sit there "focused" but not driving this will lead to even more instances of sleeping or just not monitoring generally.
The numbers are a thought experiment what I mean by effectiveness or mistake or error is just the driver (man or machine) doing something they shouldn't have. Messing up. Humans have very low rates of these mistakes/bad actions/errors. They have very very low rates of catastrophic mistakes. I predict Tesla's won't be 99.anything for a long time. But think of how error-free they have to be to account for the huge domain of roads, signs, weather, terrains, other cars such that when they make their mistakes it doesn't result in disaster... or just "accidents" galore.
What is a Serrato? I've not heard of this kind of car. If it's a typo or something, apologies, I don't know what it could be. I'm not trying to be snarky, just genuinely curious what it is?
> Combine all those settings with google maps and it easy for me to see how you could get complacent driving a new car and make mistakes.
This along with the fact that in my 2020 Toyota Corolla those features can more than likely get you into an accident is the reason I have them turned off. I don't actually feel safe with them on after experiencing their performance. Also notice a stutter in acceleration due to fuel injection timing and response from the gas pedal that concerns me.
This is exactly why the mass delusion propagated by Elon that Tesla's have full self driving is terrifying, and why "self driving" in general is a sinister lie that endangers the public.
After working in the industry, I can safely say self-driving is a massively profitable lie. It can work in limited instances, but I don't think we'll have full self driving aka level 5 in urban environments until infrastructure changes. You simply need 5G/V2X for that. Just relying on cameras will not work. On that page, I have my concerns about EVs in general, who knows what the long-term health effects are for sitting on top of a HV battery and being surrounded by that electro-smog.
> On that page, I have my concerns about EVs in general, who knows what the long-term health effects are for sitting on top of a HV battery and being surrounded by that electro-smog.
Say what? Is this some kind of industry joke that I'm not getting?
All I'm saying is there's a possibility of discovering the power electronics affecting us in an adverse way. Similarly how we thought lead, asbesthos, radium were harmless at first.
> I think your examples support the thesis that we're actually decent at discovering harmful effects, but we don't mind them.
I think it's less that we don't mind them, and more that it unfortunately is a slow process to go from first research realising there's a problem to general scientific consensus to then one or both of public awareness becoming big enough that companies have to respond and stop choosing profit over safety, or government concern needs to become great enough to enforce it through regulation.
I think we're seeing it with plastics at the moment, it seems very likely that in 50 years we'll a) know much more about the harm micro-plastics do and b) look back at now and think "there was enough evidence, why didn't they ban it immediately?" Well, because it's really convenient and cheap, companies don't want to give up profits, people like
me don't want to give up convenient packaging etc, and governments don't want to piss off consumers and businesses at the same time by banning it too fast. Similar story for the climate crisis.
We already know that microplastics are dangerous (BPA, leeching of random compounds, plankton overpopulation, animals with stuffed stomachs), there's no need to wait 50 years more.
Otherwise I agree with what you say, and I'd say it agrees with what I said: we're pretty good at detecting harmful effects, and good at not taking them into account.
I don’t think you’ve followed completely the phases that need to be gone through: the fact that microplastics are still being produced is the most straightforward demonstration that can be had on the fact that companies are still to be forced by the general public to stop producing them
I’d be less worried about long term effects and more worried about the fact that lithium-ion is very flammable. It’s bad enough that airlines ban lithium of quantities no more than a single smartphone from entering the cargo bay of a passenger airliner.
And then you have an entire car’s worth just strapped to the bottom of your chassis.
Metal fires (class D) are much harder to fight than petroleum fires (class B). They can burn way hotter, reignite after seemingly going out, and of course react to water to produce hydrogen.
Current fire fighter directives for putting out BEVs is either dump a loader bucket of sand on it (which most don't have yet), or drag it some place safe and let it burn out.
Also gasoline just sits there until ignited. Lithium battery packs can have various failure modes which send them into thermal runaway. Chevy Bolt EVs are being recalled en masse and owners told not to charge overnight or in a garage, least it go off unattended.
I'm also very pro EV. Working in self-driving and hybrids/EVs driving around in prototype vehicles I often found the on-board GPS jammed. I just found it unnerving to drive around in a vehicle emitting such strong EMF even on-board devices go haywire. That made me think.
As the sibling post said, lithium fires are a completely different class and beast than a petroleum fire.
Material science improvements won’t overcome the fundamental fact that the third atomic element on the periodic table just really wants to light things on fire.
This line of reasoning makes littke sense. The fact that certain things were incorrectly labeled as harmless should definitely inform how we go about determining id something is harmful. But the leap that we should expect harmless-seeming things to be harmful is unhelpful.
It's not far removed from flatearthers/qanon/911-truthers/moonhoax people defending theie stance by naming conspiracies thought false at the time but later revealed as true. Such events, and your speculations, are merely examples of the fact that we're sometimes wrong. Not that we should live our lives expecting that every accepted fact is wrong.
If it’s good enough for us (two small cameras and two microphones all mounted right next to each other) why is it inherently infeasible for a computer?
Those cameras are backed by processing that's so far ahead of anything silicon can do it's not even funny. It's not inherently infeasible, but it seems computers will need a huge leap in certain capabilities before they will be able to act as a drop-in replacement for human drivers.
I find that argument being thrown around once in a while, but let me ask you this. Why are we spending billions of dollars in research that we can vaguely hope to reach human perception levels in one distant day. This whole thing is just not worth it. That's why I said especially for dense urban traffic driving, we need extra sensing, and need from the environment so it can help the car. Imagine if the car knew there is nothing coming from a junction we don't need to stop all the way etc.
I don’t object to providing extra data to self driving cars, I’m just a teeny bit confused (as I think many people outside of ML are) about why certain problems are much harder than others for AI.
I think there is more senses being used in driving then just eyes and ears. I am not the best driver in the world, but I use a lot of senses while driving, abnormal vibrations that are felt cause me to be more alert. Constant awareness of hills and road conditions are felt.
Saying it's just cameras and microphones vastly over simplifies what driving involves for a human.
It's interesting that you talk about using more senses while driving. I always crack the window to listen to the road noise as it helps me guage what the other cars are doing and has helped several times in avoiding collisions with other drivers that weren't paying attention since I could hear the change in noise within the standard noise of the road.
who knows what the long-term health effects are for sitting on top of a HV battery and being surrounded by that electro-smog.
A fellow named Planck knew, when he used the expression E=hv to describe the difference between ionizing and non-ionizing radiation. You just missed class that day, I imagine. Fortunately it's never too late to learn.
The premise behind self driving cars that can handle most situations themselves relies on the assumption that a human operator is paying attention and can take over in situations the computer can't handle. I'm not a pilot, but it's likely that assumption can hold while flying at cruising altitude, but not so much when driving in various inconsistent road and traffic conditions.
Think of how silly this is though. You offload driving to a car, then...demand passenger give absolute attention every second that car operates itself. For one, this is a pointless tool if you have to monitor it all the time. For another, it is impossible that humans will be able to focus their attentions fully when they aren't really doing anything to drive the car. The better self-driving gets (say, safe and effective 99.999% or any incremental level higher), the less the human monitor will actually be watching. Thus the more likely any error will be catastrophic.
My own experience is that I tend to snooze as a car passenger on a long trip while I do not when I am the one driving. No need for speculation, I know I would be incapable of taking over on a short notice if the car does something bad.
It doesn’t follow that errors will be catestrophic. It could be the car just fails to handle edge case situations and just stops… granted on a highway that’s bad but everywhere else, that’s a hassle to traffic at worst. Also highway full self driving is much closer to being a solved problem.
Even the failures you see in the beta of Tesla FSD, aren’t catestrophic. The worst I’ve seen is a curb strikes in terms of actual “accidents”
What you are describing is level 4. And if tesla were anywhere near that that would be freaking brilliant.
Tesla FSD has many deaths and countless serious accidents under it's belt. Which doesn't necessarily have to be that bad, but as expected a computer doesn't make the same kind of mistakes that a human does and the kind of mistakes that tesla does doesn't inspire confidence (but rather nightmare fuel that you just can't wake up from).
They should be banned until they have level 4 and take responsibility for their cars actions. Blaming it on a human driver and calling it a day is ludicrous.
> a computer doesn't make the same kind of mistakes that a human does
I think this is particularly problematic because with other humans, we can at least develop a sense of what types of mistakes they tend to make and try to adjust our driving accordingly. Things like not driving in another person's blind-spot, etc. but also more subtle stuff. I feel like there a sort of hive-mind that develops amongst drivers and probably makes things safer than they would otherwise be.
I just looked up a random video of someone testing FSD 10.8 and in the first few minutes, it failed to signal before changing lanes and it almost caused an accident. All the while the driver is saying how good it is and how comfortable he is.. The car drives not smoothly at all and does random braking when around pedestrians.
The car also features a yoke instead of a wheel. Which makes holding the yoke while autopilot is driving almost impossible.
This is because drivers are giving 100% attention to monitoring the car in these scenarios. So again, this means the tool is useless (right now it is), or if the driver stops paying attention and becomes an unfocused passenger (the actual point of self-driving) then the tool becomes absurdly deadly.
Everywhere along this spectrum, as attention declines, catastrophic risk goes up. Self driving is vapor ware at best, a scandal at worst. It's not going to end well.
You just keep adding nonsense to the discussion. Is self driving going to be limited to 50km/h? No it's not, so what are you talking about?
Is crashing your car at 50km/h not potentially deadly or dangerous? Of course it is. So, what are you talking about?
When planes typically fly in the air, how many other planes are nearby? 0
When cars drive on the road, how many other cars or pedestrians are nearby? Often many
Just because planes get completely obliterated at the speeds they travel doesn't mean cars going much slower in accidents come out unscathed. This is like elementary school stuff here.
Apologises for triggering you here. I’m not meaning to troll, or add nonsense.
I agree that there is a lot more going on with cars, as traffic and road conditions vary more than air traffic and sky conditions.
But my response is the failures aren’t that bad in the car world. Autopilot (ie lane keep and cruise control) have been probably safer than human drivers. Navigate on autopilot is very stable, but limited to highway… that’s why I referred to the lower speed, because high speed is closer to solved. The hard part of full self driving, is city streets, which is typically lower speeds. The failure mode of FSD is typically for it to just stop… I have never seen a single failure where it full speeds into another object. This did seem to happen sometimes in autopilot in the past, but not FSD
Agree that isn’t good. And definitely looked like it would collide. I guess if you agree to sign up to train the NN you need to accept the fact it’s going to do this stuff.
As an aside, it seems strange to me that the high beams switch Off as they approach the oncoming car. Seems the driver disabled them… auto disabled would have meant they detected the car, which then seems unlikely they’d move into it’s path. But very hard to judge off camera alone… they might not even be on FSD. Shrug.
I almost crashed into a tesla on the highway when I changed lane right behind a Tesla. The Tesla was going much faster than me, and when I changed to its lane a bit too close, it started breaking for no reason, forcing me to jump on the brakes as well. I do not believe it was the driver, but the automatic brake assist that did not like me coming in too close…
Problematic and completely unexpected. No one speeding past you is going to brake check you for moving into their lane, so this kind of behavior from a computer is inevitably going to lead to accidents.
You can’t know what is in front of the other car or why the Tesla braked. It could have been passing a slower car or truck (likely given the situation as described). If that car swerved a bit into the Tesla’s lane, braking could have been the right choice.
The car behind is always responsible for maintaining distance.
Realistically, if the tesla felt squeezed by a car it was passing, going a bit onto the shoulder would have been what I would have expected from a human driver. But no auto steering, tesla or otherwise, is going to do that.
> of course, in a large sense, the one that hits from behind is at fault
I don't know where you live, but where I live (Denmark) brake-checking is a serious traffic offense which can lead to loss of license. While the driver behind has to keep distance, you cannot brake for non-traffic reasons. I.e. you cannot simply brake-check someone on the freeway because you think the driver behind you is driving too close (also a serious offense).
I think that is not really related. Following a “guidance system” that’s using gps (and altitude in the planes case) is very different from having a neural net interpret the surroundings visually and make decisions based on that.
I don’t know where/why you are attributing sinister motives here… it seems a very understandable motive to make transport better.
I would hope I would not drive into a lake or any other obvious hazard, but I have driven the wrong way onto a one-way portion of a street when I was following Google Maps.
My gf once asked me why I wouldn't just park on the street, it was night, it was Delft (Netherlands), it was very still standing water in one of the canals with leaves on top.
It did look surprisingly like a street, but I knew I was in Delft.
We used to have a big pond in the frontyard. One day it was completely covered in duckweed (kroos in dutch) and one of my friends just walked straight into it, not aware of the water underneath.
that's just very bad design keeping water at same level with street, I was under impression channels in Netherlands are always below ground level for many obvious reasons
It wasn't designed that way. In other locations in the city, street level is at least 150 cm above the water. But because of natural subsidence, ground water extraction, construction and peat oxidation, some parts of city's peat soil have subsided in the last few decades to only 25 cm above the water.
Sometimes the water decides to do that on its own.
I've ridden through Chiang Mai after a heavy rain when the streets were all a foot deep, so you'd get a nice foot wash riding along and despair for your flip flops when stopped.
And if you strayed too far from the car in front or cut a corner, you'd find the 4 foot deep drainage ditches and 12 foot deep canals along the sides of the main roads.
> How you get to this level of incompetence flying a 777 is far beyond my understanding.
I don't fly, but what I hear from my commercial pilot friends is that many carriers are emphasizing pilots to run on auto-everything so their actual fly-the-plane skills deteriorate. And much worse, the newer generation of pilots never acquired those skills to begin with. due to relying on automation.
There is a well known video among commercial pilots called "Children of the Magenta Line".
The magenta line refers to a line on the display you can just "follow" around that is calculated by the computer.
The video is all about how young pilots are just following around the magenta line and not having the "big picture" of what is going on in their heads.
And this video is exactly what happened here from my limited understanding. Due the the desired altitude being set at zero, the magenta line the pilot flying was aiming for was... well, set to 0ft.
I can see a longer-than-usual roll-out, but not going rate positive when you are well past V2 and hitting amber runway lights is an unforgivable mistake.
> Pilots are getting far too comfortable flying what the computer is telling them instead of flying the actual aircraft.
Can confirm. I flew with someone recently who has a very fancy integrated glass cockpit, and shortly after takeoff he turned the controls over to me. The flight director was very clearly giving incorrect indications, turning us onto a track 90° from our intended course, so I ignored it and flew what I knew to be correct before turning my attention to getting the radios set up right. It took an unbelievable amount of arguing to convince the owner that the aircraft was configured incorrectly, and that I was in fact flying the airplane correctly. And all I had to do to notice this was look outside.
> How you get to this level of incompetence flying a 777 is far beyond my understanding.
Well, training that emphasizes automation rather than manual flying and thinking (an even bigger issue in Eastern countries, see Asiana on SFO) but that carries over to the top levels.
People that train on the subject because it "pays well" or because of fame rather than because they like it.
People that have some other favourable attributes to flying Emirates.
Lack of CRM on the cockpit.
But anyway, these are contributing factors. You could argue that checklist are an automation step and those guys missed it. Apparently 4 people were fired, and I really can't disagree with this.
I think the article was spot on about this - everyone was focused on managing the computers rather than actually flying. I wouldn't be surprised if no one was actually looking out of the window, but instead looking at the screens and wondering why the numbers aren't correct.
> The article said--it's the smallest plane Emirates flies.
What does that have to do with anything?
The 777 is a huge aircraft. It has a maximum passenger capacity of 550 people.
If you can have 550 people lives in your hands and are making mistakes like this, it doesn't matter if it's the smallest aircraft in their fleet.
And before anyone bothers to point it out, I know this particular airframe was not configured for 550. But being the smallest in the fleet means absolutely nothing.
I thought based on how pilots were paid, longer routes were more desirable? And shorter routes would be in lower demand, since airlines don't typically pay for time spent on the ground?
Even the least promoted pilot shouldn't be this incompetent in an airline like Emirates. I could perhaps understand it from some small airline in developing country where the pilots got their licences through bribery.
You shouldn't be flying a Cessna 152 if you can't follow a checklist. Although I would wait until the investigation is finished...it's hard to know what goes on in the cockpit until all the data is gathered and analysed.
Well, I wouldn't doubt one of the best airlines on the planet. However, the article claims that they hire pilots with lower flying hours as they cannot find experienced pilots like American or United does. I fail to understand this. Can't they just hire experienced pilots from other airlines instead of hiring the inexperienced ones ?
They can't, because their working conditions are terrible compared to those other airlines. Nobody with thousands of hours of jet time would ever switch from a US or EU legacy carrier to someone like Emirates.
This is pretty sad. Emirates, for all their glory doesn't pay their pilots well. I mean, what's the point in spending on stuff like fancy first class seats and all that without paying pilots well. True irony. Fanciest and the most comfortable airline for the passengers, the exact opposite for the pilots.
It's not just pay, it's also very demanding schedules (which is a safety concern, many incidents happen due to crew fatigue) and zero labor protection in UAE laws. I know a few ex-Emirates pilots that got fired during the pandemic, they were treated like trash almost literally thrown out of the country.
I've been a frequent flyer of Emirates for the past ~ six years. From my perspective, the airline suddenly changed about three years ago. The quality of their service dropped and still every time that I take an EK flight, I notice something else is missing from the service. Flying Emirates does not have any glory anymore.
They still have some of the brand association of luxury, but indeed it's a very normal airline just like all the others. With similar service, legroom and food. But with no direct flights, which is less comfortable.
It’s more a question of Emirates being a relative newcomer with no strong unionization, and therefore not having to align with the absurdly high salaries the legacy carriers are saddled with. Compared with other new airlines (EasyJet, Ryanair etc) Emirates pay quite well
Comes down to a few things. First, money. Second, location. Third, skill.
The Middle East doesn't have nearly the pool of pilots to start with compared to North America, and even NA is suffering from an extreme shortage of pilots. (Wonder why the airlines no longer fly to your small town? Probably because of a shortage of pilots)
Is there actually a pilot shortage, or is there a pay shortage? Most of the people I know who wanted to be pilots bounced off of the extremely low salary for a very long time. Unless someone does military and gets a lot of hours, and even then not always, they are stuck paying for their own license then making a bit above minimum wage flying between cleveland to nashville or something like that. If the airlines paid more money I think the shortage would probably disappear quickly.
Pilots for large airlines are paid very well (100k-200k+). The training is expensive and time-consuming, but not that expensive or that time-consuming, if we compare to e.g. medicine or science.
There is a low-paid and debt-heavy portion of the career. However, Medical students aren't paid and even the lowest-paid pilot is making more than the average phd candidate.
The fact that medical students have it hard doesn't justify it being bad for the pilots.
Also it's probably easier to get student loans for medicine than it is for piloting.
Maybe the regional carriers should pay more and have better conditions but funny how they describe it as a "pilot shortage". It's just supply and demand.
new commercial pilots who are flying small airplanes for large airlines in the usa (eg “united express” pilots) barely crack 50k usd salary. and this is after big pay rises the last year.
Some of both: An early-career pay shortage leads to an actual shortage of experienced pilots to fly big jets.
You can, of course, easily solve the problem by raising wages for regional jet pilots, but it takes a few years for the excess pilots to propagate to the major carriers.
The error we're talking about here is something that shouldn't happen even to a rookie straight out of flight school. Hell, even a kid flying flight simulator knows that something is wrong if you're still on the ground at 200kts on a 777.
The takeoff isn't directly automated, but there's a system called the "flight director" [1] that provides indicators on the pilot flying's display that correspond with a configured trajectory. In this case, it sounds like the flight director was badly misconfigured and the pilots not only omitted to reconfigure it, but followed the erroneous indication far past any point at which it was remotely reasonable (or remotely safe!) to do so.
I mean, you say that as if it were a joke, and yet I can't fathom any other explanation for this kind of event but that the pilot flying assumed the automation "knows where it's going". Unfortunately, without Rainn Wilson as the PNF...
Why does it not have basic sanity checking to make sure the configured flight path is not going unreasonably close to terrain? Is it solely controlling heading and altitude or does the flight director incorporate waypoints and topology?
It's not controlling anything, just providing steering cues for the pilots along whatever route it's set up for. Setting it up for the flight's intended route is a checklist item, which appears not to have been performed on this flight. Also on the checklist is identifying the correct airspeed for current conditions, called "Vr," at which to "rotate" and bring the aircraft off the ground during the takeoff roll.
Actually doing that is a memory item, intended to be prompted by a "Vr" call from the pilot not flying. As far as I know it's not yet clear whether the PNF gave that call, but the PF certainly waited to rotate until well after reaching Vr. This is where the "sanity checking" appears to have failed, with the PF heeding the misconfigured flight director when it disagreed with other instruments and the situation visible out the cockpit windshield.
It's heading, altitude, and speed. For example, if you're currently accelerating to a target speed and climbing to a target altitude, it won't command you to pitch up so much that you stop accelerating.
The autopilot follows the flight director, and the flight director is dumb.
The autopilot wouldn't have been in use for the takeoff roll, though, would it? The question of automation smarts seems like a bugaboo in what I understand to be a highly manual flight regime.
Correct, autopilot isn't engaged until at least several hundred feet AGL; varying by airline. Some airlines don't engage autopilot until reaching transition altitude or even cruise altitude. Still, the flight director is giving you the same information, and most pilots will just follow that because it's the most accurate way of getting you to where you need to go... if you've set it up correctly.
They are, but there's a system that almost shouts "VEE ONE!" and "VEE TWO!" over a loudspeaker in the cockpit, at least if the flight director is configured properly.
I'm grossly simplifying but V1 is the point where you go "okay, are we SURE we're ready to go flying, y/n?" Around V1, if there are problems with the plane not conducive to flying (engine failure, fire, control surface malfunction especially), you reject the takeoff. On a loaded/fueled commercial jet you run a chance of tire/brake damage because the braking system will be at nearly its limit, and the heat might cause some of the thermal/pressure fuses in the tires to blow.
At V2, that's where you are going flying whether you damn well like it or not, with the only exception being (I believe?) a loss of control surface authority, because in that case, you're not flying, period. Plane on fire? You're going into the air, because if you don't, you're crashing because the plane won't stop before the end of the runway.
Then there's Vr, the point at which you go flying (rotate.)
Just about the only thing on the pilot/copilot's mind around V1/V2 is "does everything feel / sound / look OK?" and then getting as much altitude as quickly as possible once Vr is hit, because altitude is potential energy. Until you have enough altitude such that your glide distance is enough to circle back to the airport and land, you are completely boned if you lose both engines.
The runway is long enough to handle the plane rolling out from V1 to a stop. That means at Vr you aren't anywhere even remotely close to the end of the runway.
Going into the runoff area, but then also failing to climb at a normal rate even after you've finally taken off, is mind-blowingly incompetent. I would think it would almost be a career-ending mistake for both officers. If the first officer was flying the plane but the captain was PIC, the captain should have taken control a second or two after the first officer failed to rotate at V2.
My guess is that the captain was flying and the first officer didn't speak up; especially on the least desirable route people are looking to get promoted out of that hell, and arguing with the captain often isn't conducive to that. Lot of the captains know each other, and even if you get promoted, you stand a good chance of running into one of his buddies.
Edit: forgot to mention, the plane was likely screaming at them for overspeed on the gear and/or flaps, as well.
V1 is the takeoff abort decision speed; beyond V1, you are going too fast to stop safely in the runway length still available to you. If the plane is on fire and you've lost an engine and the computers and screens all die, but you've passed V1, you are still taking off anyway.
V2 is the safe single-engine-failure climb-out speed. This is the speed you pitch for after rotating, so that you're not flying too slowly that an engine failure would be dangerous.
Vr is almost always at V1 or a little bit higher; by definition, it cannot be before V1 (if you were rotating before V1, you wouldn't have to worry about whether you have enough runway left to not rotate) or after V2 (if you're at your single-engine-failure climb-out speed, you should've already rotated).
The V2 speed is entered into the autothrottle before takeoff (but the autothrottle won't maintain that speed; it will maintain a thrust setting instead). The flight director will command you to pitch up to maintain at least that speed at the current thrust, but only if you've correctly entered an altitude for it to take you up to (which did not happen in this case).
Essentially, they were relying on the flight director to command them to rotate, instead of rotating when Vr was called out. This is a catastrophic lapse in judgment.
There are more Pakistani citizens living in the UAE than there are UAE citizens. So if Pakistan has a problem with fake pilot licenses, then so does the UAE.
This logic makes no sense. First of all, the citizenship of Emirates pilots would not mirror the citizenship of the UAE itself. Most Pakistan citizens coming to the UAE are not coming to work for Emirates. Secondly, Emirates is able to do their own due diligance on pilot licenses, and since the problems with Pakistan licenses are well known throughout the industry, they would receive extra scrutiny.
The problems that occurred here are neither unique to Emirates nor likely to have anything to do with fake pilot licenses. Over-reliance on automation and complacency are global issues.
> It doesn't matter was in system. You run your checklists and configure for your own flight / your own departure etc.
I am not a pilot, but that's exactly what I thought when I read the memo too. They should be putting responsibility for the next flight on the next flight's crew, not the prior flight's crew. Train each crew to assume the plane they're prepping for their next flight is FUBAR in every way, checklist everything, fix it as necessary, and go from there. Don't rely on the prior crew for anything.
I’m not a pilot either but I am shocked that there’s not some standard “reset” procedure so that each flight starts with the plane in a consistent clean slate. It seems bizarre for any process to rely on the previous flight coincidentally leaving the plane in a certain state.
That's what pre-flight checklists are for. Even a "simple" one for a small plane like a Cessna 172 comprehensively covers setting or resetting literally every control in the plane.
It doesn't matter what the previous crew configured, because you reconfigure for the conditions and scenario you're about to fly. And this is all part of the preflight checklist.
Even if something was missed, it's very strange to not realize that the plane is still on the ground when you can see the end of the runway, despite what the computers are saying. Usually pilots have trouble trusting the instruments over their own senses, but in this case it's like the opposite.
Is there a reason for these checklists not to be automated? I would expect the pilot to set the plane in takeoff mode with destination and weight (if the plane cannot measure the weight itself) and the plane to run the checklist itself, calculate v1 and v2, check there is enough fuel to the destination, check all the control surface positions, etc.
The Boeing 777 (the plane in this incident) does have an integrated electronic checklist, which can automatically tick off items that are already completed, but the pilots still have to actually use it.
Nothing can help you when you refuse to operate the plane correctly.
Right, but what I mean is that it shouldn’t be possible to skip those steps and have things coincidentally work because of the state the plane was left in. I’m not a pilot or even much of an aviation geek, so maybe I’m misunderstanding the description of this incident, but it sounds to me like this crew did what they always do, and that normally works fine, but this time it didn’t.
There isn't much in these planes that works "coincidentally" because of a previous flight -- almost all important state has to be reconfigured each flight. The article is over-indexing on the flight director's altitude setting. These guys missed many basic things, like rotating at V1 speed and in general monitoring the plane's situation.
You rotate at Vr, not V1. Vr is almost always close to V1, though. The 777's flight computer can calculate all 3 (V1, Vr, V2) for you, put them on the speed tape (on the left of the Primary Flight Display), and audibly call them out.
There’s a lot that goes into deciding how and what tasks to do, how quickly, at what points in flight, and those numbers change as winds, weather, traffic, and fuel burn changes. Each flight, the pilots need to basically take all the factors going into the flight (# passengers, bags, cargo weight, routing, taxi time, and so on) to figure out how much fuel to use, what particular climb profile they’ll use, and a bunch more.
It makes sense that there’d be a “clean cockpit” magic button to push, but not only would the magic button need to be tested in every aspect of flight, in every configuration, in every combination of working and broken sensors… but airplanes are flippin’ complex, and flying modern airliners as part of an airline is even more complex.
Yes, assuming that the previous crew left you a 'perfect and ready to fly plane' is one hell of an assumption to make. With tiny little pieces to collect if it ends up being wrong.
Emirates has a pretty good rep but this doesn't help.
Their main problem is a very bad rep for working conditions. Crews being overly fatigued, arcane labour laws in UAE, that makes it harder for them to hire pilots and to retain experienced pilots. Combined with a local market that doesn't have a big aviation history so less people to build on.
> Yes, assuming that the previous crew left you a 'perfect and ready to fly plane' is one hell of an assumption to make. With tiny little pieces to collect if it ends up being wrong.
This is very bizarre. I once had a morning flight out of Denver (not too long ago, August 2008?) and talking to people there apparently the plane had an electronics malfunction of some sort. They said they would fly us out as soon as the plane was ready but about ten hours later, they gave up and flew us out on a plane with a different airline (which I assume was not free for them, I doubt American Airlines would fly a plane for free for United). Point is the initial problem, at least what they told us, was some lights failed to turn on but they refused to fly. I thought flight crew were very risk averse but clearly not everyone.
That's proper procedure. When in doubt: fail safe.
These guys even flew the plane back instead of being replaced by another crew, and that's with a plane that has just been used way in excess of its design parameters on more than one measure.
In the US this is explicitly part of training for major airlines. They will line up failures in The simulator and have the pilots solve/fly through the issue. You can’t fly until you pass these checks
Is it really unexpected failures or do they discuss scenario / do you expect one of a few typical scenarios?
Really throwing failures at pilots without them knowing what's going on seems to mean too many fail or rather there is not enough time to train pre-discussed scenarios and do scenarios that aren't announced.
It's truly unexpected failures,in the context of flying the plane. Basically every failure state in flying has a procedure, and as a pilot you're expected to memorize some, and know how to find and follow the checklist or red page for all emergency situations. Beyond that, you have to know how to assess and evaluate any emergency because that's your job.
It's not that unexpected, because you do recurrent training in the simulator every 6 months. And there is a set of emergencies that is done every time, for example engine failure / fire during takeoff must be done.
All pilots are trained to know the immediate actions required for a whole set of scenarios (we call these memory items) the point of the sim check is not to learn them but to check that you know all of them and to keep some recent experience. Because these things happen so rarely in real flights.
I've had for example only once had an engine failure after takeoff and it was a partial failure without fire. In the simulator I think I've done it at least 100 times and most times they put it on fire for good measure so you're in a hurry to get on the ground. The real world one was very uneventful, return to the airport, taxi back, do a lot of paperwork.
My father is a retired airline pilot and instructor. Everything that can fail in the simulator is fair game. Pilots have recurrent trainings all the time during their career. They get tested every X months. They have medical exams.
And that is how it should be. Pilots must be able to handle every possible failure. They don’t need to know every single procedure by heart, they’ve got the manuals in the cockpit for non-urgent procedures.
My father loves _flying_. He’s from a generation that trained on airplanes with little automation. He often talked about younger pilots relying too much on instruments. He would get annoyed with pilots that didn’t appreciate manually flying a plane.
I'm just asking if it's really a reality as i previously got the impression airline training wasn't quite at that level.
As in mostly you get an engine out because that needs training every time.
And if the unexpected scenario is an engine out most of the time ... it's not really unexpected.
I wonder how often stuff like static port blockages and even maybe the flight director being wrong during takeoff are trained in an unexpected fashion.
I'd expect quite rarely if at all.
That's ok i suppose as it's rare failures, but the descriptions here kinda suggested a level of random scenarios being thrown at pilots ... which isn't quite realistic i think due to limited simulator time alone.
I just fly small planes, so I don't know much about commercial airline sims except from what I've read -- but I think if you have just a limited time in the simulator then for sure you'd want to be hitting up flight failures more frequently to test yourself. These things are complex, so I'd imagine there would be discussions after of how it could have gone better etc, rather than just a pass /fail.
Some settings in airplanes are last left. And this is okay. If you had to reset every setting at all times it would cost time. Sometimes you find one that you go "man I wish that wasn't last left." Then software updates fix that.
> I am not a pilot, but that's exactly what I thought when I read the memo too. They should be putting responsibility for the next flight on the next flight's crew, not the prior flight's crew.
It's not a matter of responsibility, it's a risk mitigation measure. Adding the additional check is common in aviation. Yes, it is the next crew's responsibility, however if you can mitigate a risk by adding some procedure to what the previous crew does, at little or no cost until a software update is released fixing the issue, you do that every time.
I know nothing about aviation but it seems to me like the auto pilot software could use a lot of improvement.
Why on earth would you give the pilot an instruction that would almost certainly lead to a crash ? It seems like there aren’t enough sanity checks. Any such dodgy situation should trigger an alarm and disengage the auto pilot imo.
Auto pilots disengaging and sounding an alarm in unexpected situations have lead to multiple crashes, also the autopilot wasn’t engaged at the time, but the pilot was following a computer created flight path.
Commercial aviation is designed around highly trained people following defined procedures. Auto pilots are not supposed to second-guess the inputs because that adds to confusion and cognitive load on the pilots. Pilots tell the computer what to do, not the other way around.
What _should_ have happened here is the pilot monitoring should be watching the instruments and calling out flight stages. When the pilot flying didn’t initiate a proper climb they should have called this out, or taken over control until a safe altitude was established. This wasn’t an autopilot issue, this was a breakdown on crew resource management.
By my reading, auto pilot was not involved. Flight Directors are basically an overlay on the instruments showing where the pitch, roll, etc. of the plane needs to be to achieve the programmed navigational fixes.
Is there any situation when it would make sense to have the flight director set to an altitude of zero before takeoff ? If not, then maybe it shouldn't stay switched on and set to zero after landing when the crew leaves the cabin ? Couldn't it be reset automatically ?
That's what the memo the airline company put out said about this.
It's hard to say if it's a good thing or not, imagine an incorrect flight plan is entered but non 0, you could fly straight into another flight path or an obstacle, like a mountain.
The preflight checklist covers setting this (and all other parameters) to correct values. Not only did the flight crew not do that, the statement by Emirates heavily suggests that instead of making people actually follow the preflight checklists (which cover many other safety-critical settings), they're relying entirely on the plane being left in the right state by the previous crew.
> I’m a captain at a U.S. major airline with a lot of time in 777s. [...] Also, this is something that SHOULD have been caught by at least two separate checklists, which they obviously didn’t do.
I wonder what “these types of carriers” means here. I know there are low-cost carriers and I wouldn’t be surprised if they had less-experienced pilots, etc. But to me Emirates strikes me as a well-funded, reputable operation. Does he mean Middle Eastern carriers?
- Prioritizes "face" or "reputation" above literal accuracy
- Has lower-level officials that accept bribes
- Has had airlines and planes for fewer than 5 decades
Assume any complex high-risk engineering system run by that culture will occasionally fail in major ways.
I am expat in south east asia and his comment is correct
low level officials taking bribes means you can do stuff like never wear a seatbelt, just pay the cop $20 USD if you get pulled over. get pulled up for smoking indoors? $20 USD. run over an illegal immigrant and kill them? USD 1000 to have their records completely erased. Sim card? $5 USD cash. rape? what? the police here hung up on my friend on a 911 call because they spoke english and not the 'national language' (police do speak english here).on a different note, it is legal to rape your wife here
low level corruption causes immense problems and injustices in society. here, everyone drink drives. driving a car while extremely drunk here is default behaviour.
ye high level official bribery will cause probelms, but the low level stuff is what completely craters the system and causes the catastrophic failures in complex engineering systems.
You said it better than I could. I am a US citizen, and most of my middle/upper-middle class colleagues simply cannot imagine that their society is the result of many hundreds of thousands of people following totally arbitrary rules, and that if people don't follow those rules their society will collapse. They think that US cops (who for the most part can't be bribed by citizens) are normal, and non-US cops who accept bribes are somehow exotic or abnormal.
> the plane was flying funny because one of the aileron was replaced with electrical tape from dollar store by international airline technician
This is wild speculation, based on second-hand information from your friend, who, given that they referred to it as electrical tape when anyone in the industry would know it's called speed tape and has no similarities with electrical tape, probably doesn't actually know which part of the aircraft is the aileron.
Speed tape on the wing is quite common anywhere in the world on any carrier, and is safe if done properly.
Lion Air and Indonesian aviation has a lot of problems in general, but I doubt this issue was genuinely a safety issue. Your friend probably felt some turbulence on landing.
While I don’t know the specifics here, many people mistake tape on airplanes for normal tape, when it is usually aluminum speed tape[0] that is (in the right scenario) safe to use on airplanes.
It's true enough in western societies that we specifically needed Cockpit Resource Management as a tool to overcome deference to senior officers. A classic example being KLM captain van Zanten from the Tenerife disaster.
I hate that this thread descended into chauvinism and generalized cultural value judgments. The problem of excessive deference to authority plagues all cultures. By not being obsessed about saving face for our own society and acknowledging our own flaws, we make it easier for all of us to learn together.
I omitted that bit because I thought it was the least interesting part of Mark D's comment. The subthread which it has inspired here has reinforced my impression. :(
There was actually lots of additional interesting technical detail from Mark D which I also left out; the part that was germane to the comment I replied to was the bit about the checklists, since that backed up the notion that the Emirates notice was messed up and weird.
The conclusion though, "Don’t fly on these types of carriers", can easily be interpreted as a sociocultural swipe, even if I'm inclined to believe Mark D meant something else and just botched the wording.
Imagine having an intermediate system in your Tesla, one between fully manual and the autopilot. This system uses the navigation capabilities of the autopilot to draw in the big screen some squares that guide you to your destination without the need to look outside. Something you may find in a videogame. This system is helpfull and easy to use, and allows you to pay attention to more important matters.
Well, now imagine getting so used to this system that you forget how to drive without the squares telling you what to do every single moment. So used that if the squares guide you to drive straight to a wall at 100mph you do it without hesitation.
This is what supposedly happened in this case. The flight director is there to help you, but you are supposed to know how to fly without it.
When you rotate an airliner the initial pitch is around 15 degrees nose up, that way your rate of climb and the optimal climb speed is maintained.
They kept the nose almost horizontal, against any natural instinct for a pilot. They almost overrun the runway without rotating, and they barely rotated just enough to be able to keep the flight director centered in their screen.
It seems that the speed went beyond the structural limit of the tyres (around 200 kts) and if they didn’t retract the flaps, beyond their structural limit too. And then they proceeded to fly almost scratching the obstacles in their path.
Why does something like this happpens? Because in an airline like Emirates actual piloting skills are actively punished!. You are not allowed to fly the plane manually, nor disconect automatic systems if they are available.
They expect robot like precission applying procedures, modern airliners log dozens of instruments and have automatic reports when procedure limitation are exceeding (is a big brother like work environment). This may seem a good philosophy, but it actually creates an situation where pilots loose necessary skills, and when computers or procedures fail, as they often do in airplanes, pilots are not able to react properly anymore.
I’m sure Emirates will punish the pilots and will set new procedures over the current ones, to try to avoid this kind of situations in the future. They will solve nothing, they will only make it worse. Is a culture of fear and punishment. This also happens in other middle east airlines like Qatar. Airlines in Asia are making this mistake too.
In contrast airlines in the US and Europe (except Ryanair and some other British carriers) give pilots much more freedom regarding manual flight, which helps them to keep their skills honed.
Hope this gives a different perspective on the problem. The wrong altitude selection is the minor of the problems in this case IMHO.
I am an Airline captain with 22 years of experience (737, MD88, A320, A330 and A340) and more than 14k flight hours.
Edit, some typos and making a couple of sentences more clear.
This rings true. I live in Doha and a very good family friend is an instructor in the local flag carrier's training program.
He was recently bemoaning the following situation:
1. There was a go-around on some flight or other (for good reason!)
2. Management interpreted this as a failure, and announced that every go-around will now have a full post-mortem with involved pilots requiring more sim time.
3. Pilots now think go-arounds are now off-limits. Oops!
This is a very, very common pattern in the Middle East. When the waterfall goes as planned, service is excellent. When exceptions happen, management freaks out and adds more layers of regulation to make sure the exception can't happen again, failing to recognize that exceptions are sometimes good. For benign stuff like customer service, you get frustrated customers. For important stuff like air safety... well, you end up with close calls or crashed airplanes.
My dad flew for 20 years in the Air Force. He more than once defied procedure that would have put him in jeopardy on a flight. He told me he was not going to die because of a bureaucratic rule written by someone whose ass was safely behind a desk.
One was when formation flying, pilots were to keep both eyes on the lead, and mimic his flying. This led to several crashes where the whole team died because the lead flew into the ground. He said he'd be damned if he was going to do that, and kept one eye on the lead and the other eye on the ground.
He spent some years as a flight instructor, who sits in the back seat. There were incidents where the student would panic and crash the airplane. That wasn't going to happen to my dad, either, and he kept a length of iron pipe at hand to beat the student into letting go of the controls.
(That Airbus that crashed into the Atlantic a few years ago was an example of the junior pilot panicking and holding the stick back till it crashed.)
Yes about the airbus but the problem there is also that there is no feedback, so the second pilot can't feel what the other one is doing.
But in that case (I think you mean AF447) there was a lot more going on with confusing warning indications. They did however fly a perfectly functioning aircraft straight into the sea.
you’re correct that the pitot tubes had malfunctioned, and this was a significant factor.
but the gp point has validity.
if the controls were physically linked, like they are in boeing aircraft, the pilots who were putting in opposing inputs would have realized what was going on pretty quickly.
On that note: I have had a triple go-around and even though everybody was super happy about making their connections on time I personally thought that that was an extremely unsafe landing with the port side wing coming uncomfortably close to the tarmac. Severe windshear, ours was the last flight to land before the airport was closed. The pilot stood at the door to say goodbye to the passengers and was visibly upset.
Any kind of pressure on pilots to make connections or to reduce the number of go arounds or even alternates should be pushed back against hard. It can only lead to more accidents.
This rings very true. My dad is a retired commercial pilot and has a whole evening's worth of rant about this, albeit mostly in the 1970s-80s, before pilot hours were regulated to quite the same extent they are now (and with less automation). He said that his absolute worst was a "double washington" (i.e. London to Washington twice in "a day"), and that on the return leg either him or a co-pilot would really be operating at their physiological limits, at the stressful approach to Heathrow. Turnaround time and load factors are basically the thing that airlines compete on, with the net result that they were strongly encouraged to "man up" and "power on through".
After falling asleep for 10 minutes over the mid-atlantic at some point before I was born, apparently all the pilots in the airline agreed to sleep in shifts for safety reasons, and never told management about it...
He also has a story about a frozen Canada goose coming through the windscreen at FL300 above a dark cloud, but that's a story for another day...
The interesting thing about that flight is that I don't think the pilot knew how close to crashing we came whereas I had a 'front row seat' to what would have been a wingtip ground strike if we had been just a tiny little bit lower. I've flown 100's of times per year for a decade and I've never had a landing that bad. This was a short hop from Berlin to Amsterdam on Lufthansa.
I'd love to hear your dad's stories, there isn't a pilot that I know that doesn't have an evening's worth of material, so if you have the time to write that one up that would be great.
Interesting comparing this with the 737MAX crashes. There, I regularly get excoriated on HN for suggesting that the pilots not knowing to turn off the stab trim system when it goes haywire is unreasonable, despite it being a "memory item" and despite Emergency Airworthiness Directives reiterating this.
It’s because from a safety standpoint, procedural controls are not preferable to engineered controls. PPE controls are lower still.
It would be like alleviating a motorcycle manufacturer from responsibility of a bad design by claiming riders wouldn’t be hurt if they were wearing all their protective gear. It may be true while also being sub-standard design practice for a safety-critical system.
I made that same point repeatedly, to mixed reception. Boeing is absolutely at fault but the airlines that had crashes were also guilty of cutting corners. Boeing is (rightly) held to a high standard. I try to fly on airlines that have similar high standards for the operations.
Of course Boeing shares blame for this. But so do the parts suppliers, the installers of the bad part, and the pilots. All of that had to fail for the crash to happen.
I would like to know what the Ryanair pilots are ordered to do by their higher ups. Been on a few of them in Europe and they all hit the runway pretty hard coming back down.
An old running joke I've heard is if the pilot has a nice smooth landing then if the pilot had a military piloting background they must be from the Air Force, hard set downs meant Navy. Navy does hard set downs because when you are landing on an aircraft carrier you only have so much runway and you have to get the hook down.
Try KLM. Drop it down from 3' up and ensure firm contact with the runway, rather than those endless landings where nobody knows if you've touched down or not. Better to avoid the ambiguity. You have to break ground effect somehow and the decisive way is the best, then at least you know you're no longer flying but rolling.
I think GP is saying that Ryanair demands that pilots use the automation to the fullest extent, and disallows "hand flying" (with an exeption, one presumes, for when the automatic systems fail, and a hope that the pilots remember how to hand fly).
But a) not Ireland b) the amount of carrier pilots is low c) The intake of new pilots in EU no longer involves a military-to-airline kind of pipeline, especially with explosion in low-cost and charter airlines
FYI: firm landings are actually correctly done, especially with rain or drizzle. It should not feel like a crash though.
The worst landing I ever had was when I landed with no sensation of touching down on the runway (a zero feet rate of descent.) Both myself and my CFI went nuts trying to figure out if we were down, or still flying along the runway and about to bounce.
Gulf carriers in general share that reputation for bad working conditions for crews and bad safety practices.
Your safest choice is a US or EU legacy carrier. But for example in Asia you will not have that choice for local routes. Then there aren't that many things you can see on the outside. But one thing that can help is going with a large airline that also operates long haul flights into the US and Europe. There is a list for example of airlines banned from operating into Europe. You get on that list by having questionable safety or maintenance practices. So being allowed to operate flights to Europe is a small positive signal.
Yes Gulf carriers, but not all of them, Emirates and Qatar are the most important.
They really have high standard and try their best to have the best safety standars. Is not for lack of trying. Is that their system has gome far too much in one direction that is now backfiring.
Absolutely you may have an airline that allows too much manual flying and too low procedure standars (it is what happened at first in commercial aviation). There is a sweet spot between standart procedure and manual proficiency, and some airlines have been going way beyond that spot for some time.
> I really hope that's not the actual message Emirates is putting out there.
Why imagine there is hope of that? Is there some question as to the credibility of the reporting here? If this memo claim were fake I imagine Emirates would deny the report and initiate some lawsuits. The story linked here is based on the reporting[1] of one Gilbert Ott of godsavethepoints.com over five days ago which includes a image of the 'memo' Emirates issued, so someone has actually put their name on this story, including the report about the memo.
Blatant dysfunction should not be discounted through false hopes. This airline is clearly operated by incompetents, both in the cockpit and among its management. Someone will doubtless feel compelled to cite some safety record or another; this carries no weight if you cannot eliminate the contribution of aircraft designers and the automation they've provided. And you can't.
You'd hope a regulator would turn their wandering eye to this airline, both from a hiring, operations and management perspective. There are errors on all of those.
Emirates is a Flag Carrier for UAE, apparently owned by Dubai's government via 'Investment Corporation of Dubai.' There is, therefore, no meaningful distinction between it and its regulator, so don't hold your breath on that hope either. The FAA will demur; bet the rent money on that. We're talking UAE money here, and that means most of the US Senate is on speed dial. No FAA functionaries will be troubling UAE over this.
The only hope one may legitimately cling to is that the problem space is mostly solved by the automation and the autopilot operators occupying the cockpits of these aircraft continue to not interfere with it too much.
Airplane manufacturers and regulators being too close is not good, see the MAX MCAS issue, airlines should be completely firewalled from regulators for very obvious reasons.
That's pretty bad. Here in NL was have a similar situation where the state is shareholder of KLM/AF leading to all kinds of bizarre decisions in recent times. But at least I'd hope that some people at the regulator would resign rather than avoid doing their jobs when safety is on the line.
In NL you do have some somewhat-external oversight in the form of the European Aviation Safety Agency. This does _sort of_ apply to UAE as well, in the sense that EASA and FAA do some oversight of Emirates because they use their airspace and airports. But obviously that doesn't rise to the level of oversight that a proper arms-length local regulator would be able to provide.
>airlines should be completely firewalled from regulators for very obvious reasons.
Then we need to be prepared to fund the regulators appropriately. I think part of the problem is they are stretched so thin it becomes impossible to know enough about each system an individual regulator is expected to verify/certify.
Wow this looks like a classic corporate response, though in a place where it may actually affect lives.
Let’s not fix the actual issue (not following the proper procedures) but something further down the line which looks like the cause if you have no clue what’s happening.
Isn’t there some sort of reset settings button when you start a new flight? Like is there any setting set by the previous crew that should be left as is by the new crew?
Nope. The crew is responsible for verifying anything and everything important for the flight safety.
A button like that would be pretty dangerous. Just imagine somebody pressed this in an emergency...
In general flying a plane is already complex. You need to be able to understand what is going to happen to the plane when you press a button. Something that resets a lot of other settings would have extremely complex consequences.
Most of the switches are physical selectors and dials that aren't motorized so it would be impossible to reset those. And a partial reset would probably be even more confusing.
The flight computer isn't controlled by physical dials, and that's where you do pre-flight calculations (including the V1 and V2 speeds for given takeoff mass)
The incident description seems incomplete. With the high takeoff speed and slow climb rate, is it possible that the main error was an incorrect flaps setting?
777 has a takeoff configuration warning system, so unlikely. Far more likely is the pilots simply failed to command sufficient pitch up (higher pitch gives a lower airspeed and higher climb rate [at least in the airspeed and pitch regime we’re talking about here]).
Takeoff configuration warning can be silenced, but is an abort before the 80 knot callout.
Disclaimer: My experience in piloting is just a couple of hours flying planes in kerbal space program.
Wouldn't that much speed cause the plane to ascent naturally, even with zero pitch, just by the body lift? (Although, it is quite possible that real life planes have negative body lift, I have no idea)
I believe that while smaller planes will naturally pitch up as speed increases, jumbo jets are much more lumpant even with the correct flaps. That's the terrifying thing in my understanding: the plane basically had no manual pitch due to incorrect config values, and were barely gliding off the ground from minimal natural lift. Meanwhile the two potatoes at the helm were fiddling with the config values instead of pulling on the yoke.
The airplane will seek its trimmed speed (which is adjustable via a very powerful adjustable pitch horizontal stabilizer).
If there’s more power than needed for the trimmed speed, the airplane will climb. If there’s less, it will descend.
262 knots is well within the trim range, so it’s impossible to say without knowing the trim setting, but if the airplane was programmed to hold an altitude at 262 knots, it could (by driving the stab angle of incidence higher or lower, as needed)
They probably were given a command to stay at 4000 above ground (AGL) level after takeoff. The plane was to level off until further instructions from ATC. Since the plane was set to sea level on takeoff, the plane tried to level off at 4000 ft above sea level which…is not good. The plane was pitched down to descend.
Isn't the airport in Dubai barely above sea level anyways? I mean, if the difference between ground level and sea level is less than a hundred feet, then it seems like it wouldn't account for a plane thinking it was at 4,000 feet when it was really at about 200.
Oh no. So, the computer was actually trying to keep the plane pitched down to avoid ascending naturally and keep the altitude stable. This is really, really bad.
Well, the autopilot was disabled, and the flight director (which at a 1 sentence level is a visual indication of what the autopilot would do) was trying to keep the plane level at the configured altitude
The flight director did the right thing — displayed how it would manage level flight
The pilot did the wrong thing - copied the misconfigured flight director instead of flying the plane
The "black hole effect" is one mitigating factor - if you take off on a moonless night, it is surprisingly disorienting, and can become debilitating. These pilots missed a bunch of very basic things like rotating at V1. Maybe they got totally disoriented and it incapacitated them? Good training is supposed to prevent that. This wasn't a "missed an item on a checklist" story.
The black hole effect applies to visual flight, mostly the landing part. Not accelerating on a runway with runway lights and completely forgetting to look at the basic instruments.
If you look at the map, you'll notice that the plane was at 60m altitude in Downtown Dubai, at overspeed (600km/h) with highly lit skyscrapers of 200m left and right of you. Almost another 9/11. You cannot make that up even in desaster movies.
There is a pilot and a co-pilot. At any given time in a flight, one will be the "pilot flying" and the other will be the "pilot monitoring." That is exactly the point of the quote: there are two in the cockpit, and one should always be checking what the other is doing.
Pilot (piston engine, with only around 20 hours of light jet dual). As a pax on this flight, I might not have noticed much unless I was looking out the window.
This seems like a fault of the crew (obviously), but probably more on the PNF (pilot not flying) than the PF (pilot flying). Article says 4 pilots were on board, but fails to mention that 2 of them (relief pilots) were likely not in the cockpit on the departure.
It’s hard for me to imagine that a qualified crew would fly such a non-standard departure with “we were following the flight director” as an excuse. There’s nothing above you to hit on departure; get away from terrain! Fly the calculated airspeeds. Where was the Vr callout?! That rotation to a pitch angle is what takes you off the runway, not the FD command bars.
Once they made the mistake and corrected it, I don’t have much concern with the continuation. 262 knots indicated isn’t that big of a deal that I’d worry about shedding parts from departure flaps and leading edge devices or stressing the gear (they likely pulled the gear up well before breaking 250 KIAS anyway). I’d be more concerned with the safety of an overweight landing than continuing. (A 777 can dump fuel, but the crew that just moments ago couldn’t fly an entirely routine departure isn’t the crew that I want to be calling an audible and figuring out how to safely dump fuel at night and return for a max weight landing. How would they program the FD for that?! ;) )
My prediction is that almost every pax had no idea how close to dying they came until they read about it later.
I catch heat every time I mention that air crew training and certification standards are not uniform across the globe and that I am reluctant to fly on many foreign-flag carriers. (However, Emirates and most EU-flag carriers are ones that I would not hesitate to fly on.)
> My prediction is that almost every pax had no idea how close to dying they came until they read about it later.
This is why you should keep the window shade open during taxi, takeoff and landing (a handful of airlines require this, but IME it’s the minority). In the event of a crash or other incident you also want to know what’s outside the plane. Maybe there is burning debris on your side, so you want to exit from the other?
> Article says 4 pilots were on board, but fails to mention that 2 of them (relief pilots) were likely not in the cockpit on the departure.
The 777-300ER has two jump-seats in the cockpit. I believe standard practice is for the "off duty" crew to be in the cockpit, in the jump-seats, for takeoff and landing (basically, any time the "sterile cockpit" rule would apply).
> I catch heat every time I mention that air crew training and certification standards are not uniform across the globe and that I am reluctant to fly on many foreign-flag carriers. (However, Emirates and most EU-flag carriers are ones that I would not hesitate to fly on.)
Has anyone made some sort of a rough list of which carriers one can feel safe on?
My personal list is Tier 1 US, Canadian, British and German carriers as well as Qantas. Probably throw Air France and KLM in too. In Africa, Ethiopian would be the only airline I would fly if I had to fly an African airline.
I probably would have had a middle eastern airline on the list before I did a project with them last year. After that, I learned a lot that made me stick to the list above.
> I probably would have had a middle eastern airline on the list before I did a project with them last year. After that, I learned a lot that made me stick to the list above.
Are you able to elaborate? My email address is in my profile if you want to share in private.
Sure, in vague terms, there is a giant cultural problem in the Middle East where locals don't have the education or training or (more importantly) drive to actually do work, so most good employees are expats, but almost all leadership is locals because of hiring rules and nepotism.
Companies are actively trying to change it, but they aren't there yet and it's an uphill battle that has to start basically at the kindergarten level.
That culture also drives a major "image" focus, where how things look is more important that how things actually operate. I'm not from the Middle East, but this is what was communicated to us by our client / locals we interviewed.
I honestly don't know enough to have an opinion. I'm personally a fan of Star Alliance, but I don't know enough about the middle eastern airlines.
I DO know that even the airlines in NA/Europe etc. have their own share of... Challenges.. And a lot of those are saved by the regulatory / operating environment / culture, and those safeguards don't necessarily always exist elsewhere in the world.
I've heard Royal Air Maroc is pretty good when it comes to African carriers but I'm not sure. I know they have (or at least used to have) experienced crews and good airframes, but I don't know anything about their maintenance standards
Royal Air Maroc had a senior captain bounce an ATR-72 off the surface of the ocean through pure bad flying in 2018. Literally the belly of the plane scraped the ocean. Basically avoided killing 50 people by a few inches. Wasn’t gonna report it, but the damage gave him away.
I’m probably going to get downvoted for this, but focus on large, legacy airlines based in countries that have a long safety record (not just in terms of crashes, but also just general approach to safety).
> It’s hard for me to imagine that a qualified crew would fly such a non-standard departure with “we were following the flight director” as an excuse. There’s nothing above you to hit on departure; get away from terrain!
My understanding is that the 121 guys train to fly the FD and not "wing it." I think it's easier for us 91 guys to jump to the "just fly the plane" mentality.
That said, still a pretty baffling situation all around, especially the seemingly lackadaisical attitude towards post-exceedance inspection.
262 is still too high for either of those flap positions according to the above list. So while you may be correct with your comment, it is pretty much without point???
Edit: At least for the topic of this discussion. However, I'll now be able to win a bar bet when someone aks what flaps position for a heavy take off of a 777 would be
Sure. The question becomes "given an airplane that's had a limit airspeed exceeded by about 22% (so aerodynamic loads typically less than 1.5x the placarded limit), what's then the safest course of action for a crew who just minutes ago demonstrated that they were incapable to perform an uncomplicated, utterly standard takeoff procedure?"
Do you want them declaring an emergency, dumping fuel, flying around at night, and coming back for a max weight emergency landing? Or do you want them continuing the flight, sending someone back with a flashlight to look over the wing trailing edge, and continuing on as normally as possible to DC? If my butt or a family member's butt is on that airplane, I don't want these yoke holders trying to figure out and execute the emergency procedure after flunking the standard day one of flight school procedure, while flustered from almost milking themselves, and knowing that they’re looking for a job in the new year. Take a few hours; paint a magenta line all the way to DC and fly it.
"The thing that stands out here the most is the complete lack of control that the crew had over the plane. The crew was managing the computers rather than flying the aircraft."
I am not a pilot and I have no interest in flying nor any particular interest in aviation, generally ... and I find this talk to be fascinating.
EDIT: you can jump to 3:35 where he gives a specific example of a situation where pilots need to choose between automation dependency and dropping down a level of automation. He's a really good speaker and you can tell the audience is very engaged.
I find it weird that we learn so little about aviation when the industry has so much to teach us. The checklist manifesto was a great attempt to try to implement checklists in the hospital and other places, but from CRM to workload management there is so much to benefit from what the airline industry has pioneered.
As someone who works in the airline industry I find it infuriating when other industries don’t match my expectations. I saw behind the curtain at a couple of medium sized online retailers. The lack of process and procedure was just astounding to me. I know some airline pilots have had success in consulting in hospitals but maybe there exists a broader market for that service.
Unfortunately, it is difficult to get senior management to engage. If it's just a botched order they let the cowboys be cowboys and consider it a cost of business. Rigorous process requires too much painful self-reflection, so it only happens maybe after they get sued. Even in hospitals, death from improper dosing is surprisingly high.
Regulators play an important role in life safety situations. Beware of regulatory capture (737MAX not getting new type certification).
If you do find this video fascinating, I'd encourage you to take a few flight lessons, especially if you find an instructor that can be on board with your goal of not really learning to fly per se but exploring some of these more nuanced issues around task saturation and automation -- on several single seater prop planes today you have just as many avionics bells and whistles like you would in an airliner. As I've learned to fly, the mix of psychology, HCI, systems design, etc., has been awesome to contemplate. Nothing I've ever done has made me so keenly aware of my own way of processing information, how to deal with stressors, etc. It's an absolute blast, and if you have any interest in any of these areas and are so aware of it as to really like the magenta video, I promise you you'll like it too!
As I went through my pilot training (even for just single engine piston planes) there was so much focus on decision-making and interpreting automation and info from the plane that it absolutely has made me even more skeptical about the current state of self-driving cars. In the flight world, you get trained so much about how to properly monitor systems and how to be a step ahead of it at all times, and how to react to failures, etc.... It seems like we're completely not in the vicinity of that kind of thinking with self-driving cars... its as though for cars we are to just trust the automation and the law of large numbers will take care of things assuming it is safe enough!
I am reminded for some reason about an Air Asia (?) flight a few years ago that departed SYD/MEL for KUL (?) where by some fat finger accident one of the pilots initialized the plane's location (while on the ground), skipping or misplacing one decimal point in the lat/long somewhere.
He had to do it manually because some FD/GPS system was not correctly functioning or was inop, and then after doing this there was no warning that the plane's current position was suddenly registered as several hundred miles away from its last position. (correction, there were some warnings that were noticed but disregarded as they were very short and disappeared momentarily)
On takeoff, the autopilot was engaged and dutifully turned its heading to the course believed to be from the current (erroneous) position to the first desired waypoint -- throwing the plane into an unexpected large turn in a low altitude risky phase of takeoff. The pilots correctly saw some problem, but incorrectly then carried out a series of mistaken recovery steps leading to full disengagement of the plane's normal coordinated flight controls ("normal law", etc) to try and get it to behave. Apparently, everything except just shutting A/P off.
Well, one thing cascaded on another, the plane couldn't be reset back into normal flight control mode, there was fog, etc. and they couldn't land back in MEL and had to divert several hundred miles to SYD or wherever and a whole incident. It could have been even worse.
Seems that complicated systems... can get complicated easily if you overlook or start off unsure of what has gone wrong.
Bad FMC data entry during preflight caused pilots to fly west, rather than east, until they were too low on fuel and crashed into the Amazon rainforest. Sun was setting, so cardinal direction was obvious, leading some passengers to ask flight attendants why they were flying the wrong direction, but they were brushed off.
Well, that's just unforgiveable, reading about the details. The passengers even tried to alert the crew but were rebuffed. To be so situationally unaware for that long, and with that many people's lives in your hands. I mean, it was 1989 but still.
So instead of telling the next crew to configure the system properly they are instructing the previous crew not to set it wrong? I really can't follow that. Of course it should be checked, every time.
But the next crew has been told to configure the system properly many many many times (training, checklists, recurrent training). Telling them once more won’t achieve anything. On the other hand, telling previous crews not to set it wrong might achieve something, removing one failure mode and adding another level of redundancy.
The ghastly Helios accident [1] was also triggered by a misconfiguration (by maintenance staff). Sure, you’ll tell the pilots again to actually run their checklists, but you’ll also tell the maintenance crew not to screw up the standard settings.
Counterpoint: if you have a checklist item, and it is always in the desired/expected configuration already, you might start to skip it (with the concomitant consequences if one day it isn’t).
Maybe flight crews/maintenance staff should configure the machine randomly after a flight/maintenance, so as to force the next crew to actually run the checklist carefully?
Is there no functionality to reset the computer between flights? It seems insane that the first item in the checklist isn't "press the reset button". And the computer shouldn't let the aircraft take off if some required setting hasn't been set. Imagine if you were on an e-commerce website and you had to delete the previous customer's billing details before checking out!
Manually reconfiguring is what happens today. What I'm suggesting is to reset to a known good state.
Some settings should clearly start as blank and be required. Other settings could start as as a default value (defined by the airline) and the pilots would merely have to confirm that it is correct, or deviate from it.
The point is, we can do a lot better UI-wise than just reusing the settings from the previous flight, and with a little thought put into it the configuration process could be both safer and faster.
Maybe flight crews/maintenance staff should configure the machine randomly after a flight/maintenance, so as to force the next crew to actually run the checklist carefully?
This reminds me a little of Adam Langley's advice for keeping software protocol extensibility points well-oiled:
You might want to define and implement dummy extensions once a year or such, and retire old ones on a similar schedule. When extensions contain lists of values, define a range of values that clients insert at random. (https://www.imperialviolet.org/2016/05/16/agility.html)
Reminds me of a pistol class I once took. The instructor said he didn't believe in using the physical safety, because people tended to use it as a crutch to ignore other safety measures.
Setting autopilot immediately on takeoff is very tempting because it massively reduces workload vs hand flying during a critical flight stage (high workload). So there’s a strong temptation to do this. But if you miss a few checklist items and don’t fully understand the systems, this is the result. The numbers seem to indicate they didn’t immediately recover the situation either I.e immediately switch to hand flying and adjust throttle. At 1000 ft they’re still 80 knots too fast.
I also wonder about the decision to continue on after overspeeding the flaps. Takeoff flaps with full load on the 777-300ER is 15 degrees. Max airspeed with that setting is 215. They hit 260 at 175ft.
I have an autopilot in my small single-engine GA aircraft and don’t even think about touching it until long after takeoff when I’m comfortably at pattern altitude. The reason being I’ve read way too many NTSB reports when the pilot engages the thing 50ft after takeoff, it has a mind of its own and the plane goes careening into the ground. There are “failsafes” built in, for example you can easily overpower it with control forces or disengage it, but that’s one more problem you don’t want to deal with during a critical phase of flight.
Yeah I got into that nasty habit in sims using pilotedge because it’s easier to switch to departure/approach on the climb. Had to break that for real life. (I’m also a GA PPL)
All hail engineering safety margins I guess. Hope they check out that plane, especially anything related to flap mounting and actuating as well as the flaps themselves. That's a lot of force.
edit: omegant points out that the tires have been over-speed:
>This wasn’t an autopilot problem. The problem occurred while still in takeoff roll.
Just an amateur here, I could be wrong.
It is true that according to the article, autopilot was not engaged and the pilot was following the flight director during the take off roll. However, had the pilots obtained positive climb, and satisfied autopilot activation requirements, and turned it on, wouldn't it be catastrophic if the altitude had not been set properly? [0]
> While it would be standard to return to the airport to assess damage when something like this happens, the pilots made the decision to continue to Washington
That reminds me of Air India Express Flight 611, where the plane hit the antenna at the end of the runway and the brick perimeter wall(!) Even though ATC notified the pilots that they hit something, the pilots decided to keep flying for 4 hours. On landing, the plane was found to have a bunch of damage including a hole in the underside.
(The underlying problem was the captain's seat backrest collapsed during takeoff and he accidentally yanked back on the throttle, reducing their takeoff speed.)
The point is that pilots can hit something during takeoff, do severe damage to the plane, and totally not realize it.
No, the point of this one is that pilots can hit something during takeoff, be told about it, and decide to continue and climb only to be forced to land at the request of their own airline company! What is wrong with people?:-)
You missed the best part, their own airline asked them to return to Mumbai, probably to avoid bad press from International media, thats the general approach from anything Indian Govt.
> Shortly thereafter, the company contacted the flight crew and instructed them to divert to Mumbai, despite the stricken aircraft being much closer to Dubai. AI made it more dangerous and made them fly back another 3 hours.
On another note: I think that having Boeing 777 in the title here is a bit misleading, this was an Emirates aircraft and Boeing had nothing to do with this incident.
Suggest replacing 'Boeing 777' in the title with 'Emirates aircraft'
I'll play devil's advocate and suggest that it is a systems error that the 777 accepted an altitude of 0 without sounding an alarm or at least requiring prior confirmation. It shouldn't be that easy to almost total the airplane.
Besides: you can't fix stupid, if someone switches on the autopilot seconds after wheels up without checking what altitude it is set for it could be anything, including below ground level. Autopilots aren't magic, and pilots have a very large amount of discretion in how they fly the airplane, but these pilots weren't really prepared to fly the plane for one second more than they had to.
There is a reason they got sacked. Oh, and as to this being a Boeing issue: this setting can be entered on any autopilot. I'm all for giving Boeing shit when they are the ones that messed up but here they didn't.
There are all kinds of airports at altitude 0. You’re arguing for vastly more complexity in the flight director for it to know what altitudes are valid for particular locations.
> It shouldn't be that easy to almost total the airplane.
It’s not, the pilots are complete fucking morons who ignored multiple checklists and basic knowledge about flying (210 knots on the ground is insane).
This wasn’t autopilot or anything like that. It’s just hints for an already manual operation.
> You’re arguing for vastly more complexity in the flight director
The airplane should already have a terrain database in the ground proximity warning system, so triggering a warning in this case doesn't seem entirely unreasonable, should have very low false positives, and could improve flight safety.
Should it be necessary? Theoretically not, but pilots are human, and we have to choose whether we want to be smug about their stupidity while piecing together 300+ charred corpses, or account for humans doing human things.
While true, my understanding is that the altimeters automatically calibrate themselves to air pressure as well as use auxiliary readings from the radio altimeter to provide an altitude system where 0ft is sea level, always. However given that the pilots seemingly didn't follow the pre-flight checklist it's possible they did not calibrate the altimeter either.
Altimeters are only pilot calibrated to ensure planes travelling in various directions can trust their vertical separation is constant even amid rising or dropping temperatures and pressures. In that context an altimeter that calibrates itself is quite undesirable
There is the radio-altimeter which uses ground based radar but that is for spoken-aloud altitude readouts and terrain warnings and as far as I’m aware doesn’t factor in Boeing or airbus autopilot.
At least in the U.K., the lingo has us refer to this always as height (ie, above the ground) and not altitude (within the air above a hypothetical sea level)
I don't understand why the automation feature in question is even available during takeoff. What would be a situation in which use of the FD at this point in the flight would be a preferable option?
IFR doesn't mean "Let go of the controls and let HAL sort it out." This equipment seems like it's meant to assist with pilot duties, not dictate to them.
Read what is the current top comment in the article linked above, from a U.S. captain also flying 777's, for the list of checks the crew apparently missed.
Biggest flag is flying pilot apparently selecting autopilot on immediately after takeoff. Emirates note to their crews,is even more bizarre. Like an hospital reminding Anesthesiologists to check if patients are fully knocked out before operations.
Incident still under investigation but it looks bad.
This! While the maintainer of the site is a "strong character", I would trust him for timeliness of updates and correctness of the information provided.
How is he a strong character? The only strong characters I've ever seen were other people in the comments, and I'm surprised he hasn't disabled them years ago.
The style of the web site has not changed for 20 years, insisting on his conventions of writing airline names (France instead of Air France), not allowing links in comments. The list goes on, but I haven't followed the site for a while so I don't remeber the details.
Not that I would have a problem with that. It's his site, I have not paid, so he can do things his style. It must be an unbelievable effort to gather all the information.
For the first 24 seconds, the plane didn't leave the ground. Then, while a "normal" departure goes from 2000ft to 2875ft in 30 seconds, this flight went from 1050ft to 3000ft in the same amount of time. Not only the take-off was scary for being to slow to gain altitude, when it started gaining altitude, it did so very quickly.
For passenger who are knowledgeable, this was certainly frightening. Ignorance is bliss for most of us.
So does someone monitor each and every flight for anomalies like this?
Or is there some software that takes in the data from each flight and spots differences like this, there could be several more instances in the global data if this is a manual process.
Could be an interesting little project, although I guess the data isn't publicly available.
There aren't any traffic police of the sky per se but at least in the US controllers can and do report unusual and potentially violating events like this, I would assume the UAE has similar provisions. For the most part though commercial pilots are assumed to be adequately supervised by their companies, which have various means of doing so.
From a research perspective, the US has a strong model of voluntary reporting of safety events to NASA ASRS which is used by researchers. This is all human-generated rather than automated, but that's probably a good thing since am explicit goal of ASRS is to not be viewed as a punitive or disciplinary system. It has long been found that if people think safety incidents will be used for discipline or evaluation, then they won't report safety incidents - which is the origin of NASA acting as a neutral non-regulatory party to collect this data.
I'm surprised commercial planes don't do full software reset before each flight, and especially before a new crew takes over. Checklists only get you so much.
The checklists bring the aircraft back to a known state.
One 777 pilot commented to say that this issue should have been caught by at least two checklists, implying that neither checklists had been followed.
Tangentially: if my 2006 Mercedes is reset (battery disconnected) it needs to idle for 20 minutes to re-learn the idle pattern. The vehicle then needs to be driven around for a bit to re-learn driving patterns. I wonder if there are similar settings on a 777 that would need to be preserved between 'resets'.
> The checklists bring the aircraft back to a known state.
No, it doesn't, clearly. It provides a procedure that can be followed to put it to a known state, but its existence does not put it into a known state, with certainty. A reset puts it into a known state, with certainty. That known reset state probable wouldn't be useful, but can be set to be least-damaging, where a checklist can then be used to then put it into a "good" state.
My assumption is that many planes don't allow a reset to be trivial, since levers and whatnot are involved.
I also assume the FAA/military has done research into this and, probably, found that requiring a manual checklist resulted in stricter adherence to the checklist.
I'm surprised they don't do a full reset then set every checklist variable to an invalid state, in which the aircraft will not permit starting until all have been set by the flight crew.
I mean if you are able to do that, why to even have a checklist. I mean the computer must be able to complete the checklist, unless specified, no? Or is the checklist important because the pilots must know what they’re dealing with?
I can't say I have much of an idea of what goes on in a cockpit but presumably not everything is related to something a computer knows, or can change. Clean windscreens, functioning levers, coffee cup filled ;-)
Also having to poke the computer also helps check that the computer isn't obviously broken.
It takes a while (think like 10min+) to reboot these planes. Checkout "cold start 777" on YouTube [1].
There are a ton of computers (FADECs for engines, fuel management computers, FMS (flight management), PFC (flight computers), weight and balance computers... and much more). All of these have state. Which way is up for the gyros. Proper direction for the compasses. Proper location on Earth for the navigation system... etc. Basically, a 777 has the mother of all POST sequences, so frames in line operation are rarely restarted, for operational efficiency.
Reminds me of the navigation systems for some rockets (Atlas, IIRC), that require 24 hrs sitting level on the pad to fully initialize their starting ground state (super precise gyros). This ain't like restarting your internet connected juicer.
It takes too long and would achieve little. Flight turnaround times are very tight. The process for updating software on aircraft is, understandably, tightly controlled and only takes place via physical access. Pilots wouldn't know how.
Without knowing much about this Flight Director system, it’s surprising to me it would even attempt to operate in such a configuration—taking off to an altitude of zero feet. I would think it would just display some sort of message or code that’s the equivalent of “you people are on your own with that shit”.
As somebody who doesn't regularly interact with aviation I don't have an intuitive understanding of knots and feet so to me it is helpful to have these translated to units that I "understand".
I’m flying an ultralight in Europe and my IAS is in km/h. And to be frank, probably all UL and PPL aircraft I’ve seen nearby use km/h, not knots. Not sure why it is so, but it’s definitely easier for my brain. :)
I know who the pilot was on this flight and I’m damn happy she won’t be getting near the controls of a plane again anytime soon. Coworkers were astonished she didn’t rip the flaps off.
Planes taking off from Boston Logan can fly as low as 300 feet over the neighboring town of Winthrop--and average about 700. I've kayaked in the area and you basically have to stop talking every few minutes when a plane takes off. Arriving flights are even a bit lower (although not as loud).
There was an episode of Cheers about this, where Carla buys a house extremely cheap and then believes it is haunted. After sleeping a night there, she finds out it’s because it is close to where planes land at Logan Airport, which calms her nerves.
I was kayaking with a long ago girlfriend and we were in the parking lot for the Winthrop boat launch. And I do remember we got into some conversation with a local and HE TALKED VERY LOUDLY. Which was something we joked about for a time.
You could actually land your boat at the end of Logan airport from there. But I assume if you were to do so, serious people with serious weaponry would pay you a visit very quickly.
Following up from another thread, excluding reflections and what not, a plane at 75 feet should be 16x louder than the same plane with the same throttle settings at 300 feet. Sound follows the inverse square rule, where a change in distance of X reduces the perceived noise to 1/x^2. Since 300/75 = 4, you get a factor of 16 in perceived difference. Yikes.
I live on the coast, 15km from major airport and sometimes in the summer planes fly directly over my apartment, including really huge ones, just a few storeys above (I'd say 100-150 ft, maybe more). With open windows, it's madly loud. With closed windows it's a loud but somehow tolerable.
Perhaps an order of magnitude more? You'd expect an airliner at about 4000 ft at 15 km from the airport, assuming it's trying to land. It's also illegal to fly lower than 1000 ft above inhabited areas in most of the world.
Typically sound follows the inverse square rule, so the difference between the same plane at 1000 feet should be a 1/169th (1/13^2) of the noise at 75 feet. So two orders of magnitude.
I am not a pilot, but here some very fundamental rules of flying a plane have been broken and from my point of view I can't classify this as a simple error.
The crucial job of the crew planning/taking off is to
a) calculate the decision points (speeds, distances) at which they have to decide to take off or not
b) monitor speed and progress along the runway
c) abort if they can't take off safely before the decision point
What they did suggests to me a complete breakdown of discipline and process in the cockpit.
The autopilot settings have absolutely nothing to do with it.
A similar incident happened to Qatar Airways Flight 778 (another 777-300ER), from Miami in 2015 [1].
Incorrect runway lineup led to the plane only getting airborne after the end of the runway, after smashing through the ILS antennae. They then continued onto Qatar (13.5 hours away), with bits of antenna embedded in the tail.
You put your life in many other people's hands every time you step into an airport. If you use Chickens, Goats, and Passengers Airlines, expect not to make it to your destination in one piece.
737 NG (-6xx (736), -7xx, -8xx (738), -9xx (739)), MAX, and 787 are reasons not to fly on Boeing equipment. They have both obscure (Ducommun) and well-known history of cutting corners in design, testing, manufacturing, and who knows where else. Furthermore, Boeing pilots have huge cognitive dissonance issues around these matters because their salaries depend on it (remember Upton Sinclair's quote). They will overconfidently swear on their mother's graves that their equipment is "safe" on blind faith while unable to face the overwhelming evidence to the contrary. Boeing isn't as bad as kit planes made in someone's backyard or Tupolev's, but the cutthroat profit-obsession culture is different than at Airbus in terms of priorities.
Human-piloted aircraft for routine transportation makes no sense. In terms of safety and ecology for long-distance travel, trains will dominate in the end because flying adds inherent risks and consumes much more energy.
You, clearly, have never flown EK. I have six flights in EK first class under my belt. I've taken showers on their airplanes. I've enjoyed $1,500 bottles of 1964 port wine and cognac poured generously. I've eaten caviar and socialized at the onboard lounge.
Believe me, EK is nothing like a "chicken and goat airline". It's world class all the way.
What makes you think emirates doesn’t have good pilots? Do you have any source for that?
Emirates is not even a budget airline, like RYAN AIR. I mean unless you are the one who trained all emirates pilots, I don’t think you should be calling them a goat airline because an incident happened. If you look deep, every airline and even Airbus had issues.
>The maiden flight of A320 Air France 296: Official reports concluded that the pilots flew too low, too slow, failed to see the forest and accidentally flew into it. The captain, Michel Asseline, disputed the report and claimed an error in the fly-by-wire computer prevented him from applying thrust and pulling up. In the aftermath of the crash, there were allegations that investigators had tampered with evidence, specifically the aircraft's flight recorders ("black boxes").
Sure, but the reference to chickens and goats in the parent comment indicates a severe lack of knowledge about Emirates, the UAE, and (I would venture a guess) perhaps most of the world outside the USA...
Even the fastest passenger trains are quite slow compared to average airliners, and there's a lot of places connected by air that don't have train tracks between them. Air travel should stay dominant until both of those things change.
Only beyond a certain distance — there are a few routes in China and Europe where the train will get you from A to B faster than an aircraft due to the aircraft's speed advantage mostly being in cruise flight. If you take into account the time spent in the airport at either side, there are actually a lot of such routes.
> Human-piloted aircraft for routine transportation makes no sense. In terms of safety and ecology for long-distance travel, trains will dominate in the end because flying adds inherent risks and consumes much more energy.
OK cool I'll just take a train.... uhhh absolutely nowhere as I live on an island with extremely limited inter-city rail.
Scarry as it already appears, it also looks as if the flight were considered overloaded. Thus longer speedup, shallow climb. Was their landing profile also out of ordinary?
Wrong details keyed in, or some undisclosed and supposedly heavy cargo? Though such a Bond-style mystery may not make any reason to it, as the destination was IAD. Well, the pilots should feel lucky that everyone is safe and sound.
I'm surprised fatigue is not mentioned more in the discussion of this incident. The takeoff was at around 3AM local time, and Emirates is well known for pushing their pilots hard (many pilots take a job there specifically to build hours quickly, despite the unpleasant working conditions, with the intention of moving to greener pastures once they've done that).
True, I was more referring to the time relative to the pilots' sleep cycles. They likely spend their days off in UAE, and probably sleep at night on those days off, so 3AM would be a difficult time to be alert and performing at your best.
I don’t think you get an overspeed warning on ground. Also by the time pilot realised he was already over-speeding, there is no way he is going to bring back the throttle, once he is up and climbing, until he crosses the transition.
Can someone ELI5 why the plane was not gaining altitude despite higher-than-usual speed? (The physics behind it)
The autopilot set to the lower-than-current altitude effectively set the flaps into descent mode so it needed higher speed than usual to start climbing up?
Also how come it finally started getting altitude even with destination altitude being zero?
Roughly speaking, if you push forward hard enough the plane goes down. There are limits to that of course, but in this case the FD wanted to maintain an altitude of 0 so it was "suggesting" a downwards pitch to the pilot, and the pilot was doing as told. The FD is only an advisory system, it basically lets the pilot choose a goal and then it shows on the attitude display where to point the nose to achieve that goal. It helps pilots to be more efficient and precise and, in theory, avoid errors. In this case though it was misleading and apparently the pilot did eventually realize that and stop following it, but very very late, after other indications like high airspeed and the view out the window would have strongly suggested that the climb was not going as intended.
The plane took off because the pilot pulled back on the stick. There was no automation involved. The flight director is a suggestion, just like GPS navigation in a car. If you can’t turn right because the road is flooded, you go straight, and ignore the GPS which is telling you to go right.
USA is one of Emirates key markets, I wonder why they didn’t mandate higher quality pilots for the routes. I’ve flown quite a few times and remember hearing American-sounding pilot names on several occasions so I know they have them on staff.
Why would Americans be any better pilots than people of any other nationality?
Presumably what matters is the standards of pilot experience and training that an airline demands, not what country the pilots are from. Perhaps some of the Americans working for Emirates do so because they couldn’t get jobs at a US legacy carrier. (I don’t know whether this is indeed the case; I’m just claiming it’s plausible).
You are so close to imply US pilots are superior in flying? I have many incidents where US pilots failed and pilots from other countries heroically landing a damaged plane on US soil. But if yo didn’t mean that. You must understand Emirates doesn’t train pilots from their childhood. Who know/ the pilots might’ve been perfectly capable and even extraordinary. He might be under fatigue, may be he went into shock. Until FAA comes back we must avoid this talk. Recently people started making fun of pilots on Boeing 737 max, until they found out the plane had issues.
>Emirates is telling pilots not to set the altitude to zero on approach, for fear of the next crew not changing that.
that sounds like a typical result of Six Sigma/5 Whys/<whatever spawn of the PM/MBA methodology hell they are practicing there>. If one thinks a bit (note: not a part of any such methodology) there is a reason why previous - ie. landing - crew sets the target altitude to that of that target airport... I wonder whether there is a narrow specialization - ie. the managers doing safety analysis of take offs are different from that of landings :)
- So that when kph/mph is heard a pilot knows they're referencing ground speed, whereas knots is air speed.
- Pilots use latitude/longitude, and a knot is "exactly equal to 1/60th of a degree of latitude. A 1/60th of a degree of latitude is known as a minute of latitude. Therefore, a knot is equal to one minute of distance. There are 90 degrees from the equator to the pole."
But reality is that changing international standards is very hard, and nobody wants to have the battle when knots works "well enough." It likely started because existing nautical maps were used for over-ocean flight planning and stuck.
seems slightly odd to use an angular unit for airspeed; the linear equivalent would technically depend on altitude. I guess airplanes don't actually fly high enough for that to matter.
It was only ever approximately accurate anyway, a NM is no longer an arc minute because that is inconsistent even along lines of latitude, so it had to be pinned to 1.852km.
Aviation, broadly speaking, inherited all of its navigation techniques from the maritime world. In particular, prior to radio methods and to some extent today aviation navigation is based primarily on dead reckoning periodically corrected by observations, which is a technique that was pioneered in marine navigation. Since early pilots learned to navigate from seafarers, their units and conventions became strongly embedded in the field. It is, after all, still not unusual to refer to an aircraft as a ship, especially informally.
Ironically the metric system suffers the same theoretical "error" as the not-an-arc-minute-anymore NM because the meter was originally defined by taking the distance from pole to equator (to the best of their knowledge at the time) and then shifting the decimal point to get something human scale.
Initially, yes. The story itself is quite fascinating, and the Wikipedia page on the metre is a good starting point.
This was a hot topic during the 19th century as different improvements in the techniques and devices used to measure the Earth showed that the first measurements were imperfect, however good they had been at the time, and that the whole idea is not workable because the Earth is not a sphere. This lead to the formal definition of the metre as the length of a prototype in 1889.
I imagine that the mental adjustment from sphere to almost but not quite a sphere must have been harder than a switch from flat earth to spherical. You just expect reality, if it matches a certain mathematical principle, to match perfectly.
This has not been the case since 1889, when the metre was formally defined as the length of a prototype metre bar. So the metric system has not suffered the same error for more than a century.
Of course, the metric system has not been in use since 1960, but this is also true for the SI.
Oh, I think I didn't even read that correctly since it seems like it's just a mistake. A knot is, for historic reasons, a term for NM per hour. An NM is defined by latitude.
I believe that the term knots originates from a method of measuring speed at sea by use of a twisted rope and a sea anchor, but perhaps that's apocryphal. Presumably though the definition in terms of NM was a later improvement.
While I don't have any real evidence for this, I would speculate that part of the reason for the enduring use of knots in aviation is that, much like the old twisted rope, it links the unit to the instrument. The speeds usually quoted for aircraft are in KIAS, Knots Indicated Airspeed, a measurement which is basically defined by what the instrument shows. It is intentionally uncorrected for various errors in the instrument design. This happens to be convenient because many of the same factors that affect the instrument calibration also affect the aircraft handling, so handling at 100 KIAS tends to feel similar even though the actual speed varies by altitude etc. KCAS and KTAS, Calibrated and True Airspeed, are both KIAs after correction for various errors and are useful for certain calculations but not nearly as often as KIAS. My point is that it might be helpful to use Kts rather than MPH or KPH because it helps keep these factors in mind and helps to reduce ambiguity about whether a number is indicated or true, although people very frequently just write knots when meaning KIAS.
Of course for ground speed none of this matters but ground speed isn't really a "flight" concern per se, more just used for navigation calculations.
Really depends on the context. Usually things that were designed for pilots to use will display ground speed in knots since it's likely the pilot will be comparing against calculations done in knots. Things that were designed for the general public usually show mph since most people don't have a good sense of how fast a knot is.
Taxi speeds are always written in knots, so I suppose you could say it is "official" (in terms of aircraft handbooks, airfield directories, etc) to use knots for ground speed. Of course the situation of an aircraft's ground speed while on the ground is sort of an edge case.
On the topic of odd units, the operating handbooks for jet aircraft often use mach number for limits, which has some of the same properties as KIAS as far as being density dependent but is a little more precisely defined.
> So that when kph/mph is heard a pilot knows they're referencing ground speed, whereas knots is air speed.
Who is using MPH/KPH in the course of flying a plane? Why would this ambiguity resolve to “that must be ground speed”? Why would there be ambiguity in a speed readout at all? Surely context and procedure should disambiguate this rather than units?
I'm not actually sure where that quote came from, but yes, there's not likely to be much confusion. Ground speed is usually displayed by some kind of computer, could be a glass panel or just an aviation GPS unit, and is usually marked "GS" or "Kts Ground." And it's sort of considered, philosophically, to be more a navigation thing than an actual flight instrument, since you only really pay attention to it in flight if you're doing navigation calculations.
Ground speed is speed over the ground. This includes when the aircraft is in the air. Airspeed is the speed of the plane through the air. These are not the same thing. Both are reported in knots.
Other commenters say it's historical are, well, mostly right, but there is a distinct advantage to knots: a knot is 1 NM per hour, and 1 NM was historically defined as one arc minute latitude. So if you are looking at a chart and doing calculations by hand or in your head, NM and kts are very convenient. It lets you mostly do chart -> distance -> time and fuel measurements/calculations in your head.
Of course this is not perfectly accurate in several ways but given chart projections it's usually Close Enough.
The speeds given in the article are a little ambiguous in the details, the number given in the text is probably KIAS (but maybe not?) and the numbers from FlightAware will be ground speed.
Maritime navigation techniques were adapted to aviation.
For a similar reason airplanes have captains. A few decades ago they also had navigators. With a dedicated position in the cockpit.
It’s also customary to measure altitude in feet. Except in China (except Hong Kong), Kyrgyzstan, Kazakhstan, Tajikistan, Uzbekistan and Mongolia. Russia is actually moving to feet at some altitudes. Turkmenistan similarly uses feet or meters depending on altitude.
The units themselves are irrelevant as long as everyone uses the same ones. Switching costs are significant both in terms of training and equipment.
History is surely one thing, but another key point is that these speeds are always KIAS (indicated air speed). This is a measurement that doesn’t correspond to either true ground speed or true air speed (but is far more useful for flying than either of these).
Certainly I would find kph/mph slightly confusing as you might expect it to correspond to some actual velocity, rather than something to do with air pressure.
Maybe we should measure airspeed in mbar instead?
(ofc navigation is a different matter, and there just history prevails)
Using knots made navigation calculations easier with paper maps marked in degrees and minutes. A nautical mile is the same as a minute of latitude, and a knot is one nautical mile per hour. Now with electronic navigation systems kph would be just as easy, but aviators are all used to thinking in knots so changing would be difficult.
It's partly true. However, a lot of planes outside of the US would have used kilometers per hour historically. For example German WW II planes did that. Likewise a lot of Russian cold war era planes did the same. And lots of modern planes (e.g. French ones) still do. And some planes in the US actually use miles per hour instead of knots; especially in general aviation.
It's just that post war, the US aviation industry was very dominant and the metric system never was popular there. For the same reason feet are used instead of meters for measuring altitude. And unlike with nautical miles, there's not much of an advantage for that. However, the units for altimeter settings are metric (i.e. millibar instead of inches of mercury) in much of the world except in the US.
For flight controllers and procedures they had to pick a unit for flight levels that everybody understands in the same way and since the US aviation industry after WW II pretty much dominated, a lot of planes would have used feet. Internationally, flight levels are used everywhere above certain altitudes. Those are basically feet divided by 1000 (kilo feet?, LOL). So, FL 29 would be 29000 feet at 29.92 inches of mercury in the US and 1013.2 millibars in the rest of the world.
The Russians actually used meters until very recently apparently and only switched to feet a few years ago after of course first supporting flight levels. And apparently, the Chinese still use meters: https://ivao.aero/flightops/divprocedures.php?id=CN and planes must actually be capable to switch their altimeters to using meters to fly in China. Likewise, North Korea uses meters apparently.
At the very beginning, aviation had to solve the very same problems that sailors used to encounter at sea (navigation, precise position), so they naturally used the same algorithms and equipment.
That is how naval terminology entered the lexicon.
By now, it is so deeply ingrained that switching the entire world (well, there are exceptions) to the metric standard would likely cause some preventable loss of life.
At least the knot/mile system is fairly internally regular. We tolerate a much less regular one, with a lot of weird exceptions (the Gregorian calendar) out of inertia.
Both aviation and marine navigation generally use nautical miles for as it‘s a little easier to use for long distance travel (plus tradition and inertia). Measuring speed in knots make things like dead heading calculations a lot easier.
Historical reasons mostly. For the same reason that various units are used in all sorts of different industries. And guess what? If you changed you'd probably end up with some accidents in the transition as a result.
When you translate maps with vast distances down to units, a nautical mile equals a minute of latitude. With long distances, you refer to lat/long over following a road.
A nautical mile is equal to (about) one minute of latitude, which probably made calculations easier back in the day, with the tradition carrying forward.
> suggesting [...]the previous flight crew had left the altitude setting at [...] 00000 feet
Does the system really display as "00000"? That seems like an odd design choice, I'd have thought just a single "0" would be far more legible in this scenario, especially if it's on a 7-segment display.
I don't think passengers would have noticed this as anything other than a regular takeoff, because ATC does request slow climbs (ie. slower rate of ascent) every now and then.
Speaking of bad UX choices, the FAA is constantly dropping one or two or more digits from numbers like altitudes, runway lengths, heading, timestamps, etc. For example, "the runway from 180° is 5200 feet long" becomes "RWY18 52" which I suppose you get used to, but it's just begging for mixups.
As a student pilot, this drives me nuts. The volume of abbreviations, especially in NOTAMS, METARs, etc, is a bit obscene.
What’s wonderful too is that the abbreviations are often based on non English words as well, so they aren’t intuitive. As an example, instead of typing “mist” in a METAR it is “BR” based on the French word brume.
It still matters since most communication is over the radio, and in many areas it’s highly congested.
When there are 50 pilots on the radio and things are moving fast there isn’t time to slow down and spell things out, and doing so can actually be more dangerous when extraneous stuff ends up on the radio and is misinterpreted.
AIUI it is a holdover from the days when such reports were sent by telegraph/teletype, and saving characters mattered. The abbreviations became standard convention and changing them now would be more disruptive than teaching new pilots how to decipher them.
> I don't think passengers would have noticed this as anything other than a regular takeoff
It looks like the aircraft didn't leave the runway until the runway safety area at the end. That is something any passenger looking out the window could notice. A normal takeoff has the aircraft leaving the runway well before the end.
The runway lights and taxiway lights would be on. Sure, you wouldn't be able to see the ground itself outside the range of the lights, but you would be able to see enough to tell that the plane had not taken off at the end of the runway.
You can see the lighted markers along the side of the runway indicating the number of thousand feet of runway remaining. I’d definitely be concerned as we pass by 2…1…and still not rotated.
Just compared the 13th December to the 20th December. WAY off, where the proper departure was at 200kts 700ft, the bad one was still at 0-75ft, at 216-234kts.
But engineering/mechanical problems can be identified, analyzed, and eliminated (or mitigated). Human error still finds a way around our many attempts (training, checklists, standardized procedures, etc) to prevent it.
Is this a cultural problem where some areas of the world tend to defer to authority even in the face of obvious error on the part of the authority? In this case the flight computer?
I thought this was solved by training none-western pilots, in particular co-pilots, to be empowered to challenge decisions by another pilot instead of just doing nothing.
The article made me wonder if weather could have been a factor (although this seems sort of far fetched anyway). NCEI's archive is, well, 503ing, but Iowa State has a METAR archive and it shows good VFR conditions for the whole day of the flight with only moderate winds at most.
Arab societies are intensely hierarchical = orders from down from the top, there is no counterflow upwards.
This is exacerbated by the political appointees that populate the top ranks.
With an incompetent's political appointee as captain, his authority can not be challenged - even as planes crash. This manifests in quasi military groups, the the civilian air forces, which have a preponderance of ex-military officers at the top.(it has been a long time since the last war, so the older air force people move into the civilian fleet and retain their culture)
These political appointees pay competent pilots to pass exams, even flying ones.
This forum discusses this culture as it manifested in Arab armies repeated losses to an enemy with fewer men and less equipment, but with a culture that demanded a win with no hierarchical encumbrances.
https://www.meforum.org/441/why-arabs-lose-wars
Bear in mind that the cultural difference that exists in the military, which the air force is part of, has demonstrably spread to the command hierarchy where junior flight crew DARE NOT raise any issue that is at all critical of the Captain or other senior crew members. These hierarchical differences are well known in Arabic as well as Asian flight crews.
Hard to see why the demonstrated dismal failure of the Arab armies and Air forces against the Israelis is called clap trap. It is the dismissal of a serious introspective analysis of how the army and air force failed even when they outnumbered their enemy almost 10 to 1 as clap trap that speaks to the ineffective state of those forces.
Turn it off and on again. Shouldn’t these things have a standard baseline that they reset to? I would have thought you’d reset everything after a flight / first thing you do when you get in the pilot seat . The plane shouldn’t have memory or the previous flight.
I'm not a pilot and have never touched the controls of an aircraft, but I always mentally call out in my head: V1, rotate, V2 when I'm in a window seat.
I may be wrong but we just don’t forget how to drive or how to code after a vacation? It just takes some time to get used to work again, but that lasts for few days. And we opened back months ago. That was a weird statement to read for me.
I wonder if a contributing factor is that they have local noobs flying. I know quite a few very highly experienced pilots (15+yrs at Emirates) were sent home due to the pandemic.
The only pilots in the middle east with absolute job security are the locals, and they get away with shit that would see any expat fired on the spot. You can hear stories about this from all the airlines in the region... Emirates, Qatar, Oman, etc.
UAE doesn't have a high local population to begin with. Majority of people living there are expats. And among that low population only a few choose to become pilots.
You know, from the rest of the world American pilots flying in America also look like "local noobs".
Let's wait until the relevant authorities release their final report.
Maybe those particular pilots are questionable, but I don't think it's fair to blame the entire carrier.
Emirates is very responsible and it's definitely in their interests to run a safe airline.
----
(I'm probably going to get flagged but I don't care)
I've seen it again and again in HN. Whenever something goes wrong in Asia or Africa everyone's like "the noobs over the don't know what they're doing".
> "the noobs over the don't know what they're doing".
I've heard from 3 different pilots from three different ME airlines, that the locals are untouchable, and get away with all sorts of bullshit that expats would get fired in an instant for.
There are two or three pilots in every cockpit. If two or three pilots miss something that checklists should have picked up, that suggests it's not the individuals, but the airline and it's processes that are at fault.
> I've heard from 3 different pilots from three different ME airlines
Oh really????? I find that hard to believe.
Anyway, I'll assume you aren't lying.
If you have conclusive evidence, why don't you do us all a favour and give it to the FAA and try to ensure those middle eastern airlines stay grounded or atleast out of American skies?
(If you aren't from America, do the above with the aviation authority in your country, and maybe FAA as well anyway)
If you're gonna say FAA is in the arabs' payroll and the whole system is rigged, then you're screwed regardless of whether the ME "local noobs" are bad.
> that suggests it's not the individuals, but the airline and it's processes that are at fault
No it does not necessarily.
If this kind of incident regularly happens, then yes the entire airline might be flawed. But at this point there is no evidence to believe so.
> the locals are untouchable, and get away with all sorts of bullshit
First up, why don't you ask your ME pilot friends the ratio of local and expat pilots in those airlines?
What makes you think the local pilots there are into "bullshit"?
Those middle eastern airlines aren't any worse than any other major airline from anywhere else. If you can prove otherwise, contact FAA.
TL;DR - Don't screw with the defaults (MCP) after landing, especially when it's unnecessary and when the next takeoff crew likely will not remember to set it to proper value before takeoff.
I wonder if this is not a detectable condition for the flight software - maybe in some cases you want the MCP to be set to low? Otherwise autopilot could either auto correct it or issue a warning?
Sure, but the TL;DR was the approach Emirates took - they literally communicated that to the crew and I can't say I disagree with it - why add to process/work if you can leave it at a sane default? (Unless I misunderstand - which is why my question.)
> the next takeoff crew likely will not remember to set it to proper value before takeoff.
If the takeoff depends on it, the next takeoff crew had better remember to set it. What's more, there better be at least one checklist that requires them to set it. And they had better follow the checklists.
Standard takeoff procedures would have them rotating at a calculated airspeed and pitching for a climb target before transitioning to follow the FD. This crew’s issues went way past a pre-select mistake.
Why is it even possible to enter a desired altitude of 0 feet into an automatic pilot? As a layman, it’s hard for me to imagine why anyone flying a 777 would want to maintain an altitude of zero at full speed for more than a few seconds (if that).
The Germanwings Flight 9525 tragedy [1] was effectuated by the suicidal pilot inputting an auto-pilot target altitude of 100ft, while flying over high mountains. The plane dutifully flew into the mountains, murdering 150 people.
While there are airports with elevations lower than 0ft, this is a good design question to ask. In general, it's assumed the pilots will do the right thing to save the day, if needed.
I'm guessing it's done on approach so that the autopilot won't keep the plane from landing -- which would beg the question of why the autopilot would even be on during takeoff or landing.
As one would guess, it shouldn't have been on. Insight comes from a comment in TFA:
"I’m a captain at a U.S. major airline with a lot of time in 777s. Something this doesn’t mention is the that even if ground level was set into the mode control panel, the aircraft wouldn’t have descended if it was being hand flown. That means that the flying pilot here selected the autopilot on immediately after takeoff. This is unusual at Western carriers as we like to hand fly. To me this indicates a lack of experience or confidence or both."
This seems a very important question, and I'm not finding answers to it satisfactory.
To elaborate, why isn't an autopilot "aware" of whether the aircraft is taking-off/landing or in-flight?
It seems completely mad that an automated flying mode could be engaged in such an obviously suicidal way on a modern bulk carrier aircraft, without so much as a blaring warning and red flashing lights.
> It seems completely mad that an automated flying mode could be engaged in such an obviously suicidal way on a modern bulk carrier aircraft, without so much as a blaring warning and red flashing lights.
You are assuming that there weren't blaring warnings and red flashing lights. Probably there were many.
> To elaborate, why isn't an autopilot "aware" of whether the aircraft is taking-off/landing or in-flight?
Autopilots are much simpler systems. I don't know how much you know about them, but people in the general public believe they are some AI which flies the airplane from A to B. In reality they have much more common with a thermostat than an AI. You set a target heading and the plane will bank to match it. You set a target height and the plane will pitch to match it. Yes yes, they have many modes, and some modes can do much more complicated things than these, but it is not there to replace the pilot. It is a tool to alleviate dead simple piloting tasks so the pilots fatigue less on long flights.
Just as an example how simple stupid they are: You can command the autopilot to climb, and if you don't adjust the throttle appropriately (either manually, or by enabling the autopilot to adjust it) then the plane will use up all the airspeed during the climb and the autopilot will happily fly it into a stall.
Also just a general rule about aviation incidents/accidents: It is a good idea to wait for a proper investigation report before running far with all kind of assumptions about what happened, and therefore what you think should be changed.
You're talking about a GA aircraft autopilot, modern airliners go way beyond that. I don't know about the 777 but I believe all newer Boeings are FBW. Airbus has been FBW since forever. I'm mostly familiar with the A320 but that things is pretty much impossible to stall under normal operation no matter how much you misconfigure the autopilot.
No I am not only talking about GA aircraft. What do you think is going to happen if you turn off the autothrottle, manually set the throttles to idle and set the AFDS to vertical speed mode with a positive climb rate?
Here[1] is an NTSB report detailing how the AFDS and the A/T work together in a Boeing 777. They enumerate several different combinations in which the autothrottle does not provide stall protection.
It also calls out as contributing factor the crew’s mistaken belief that the airplane is pretty much impossible to stall.
As I said, I was talking about the A320. You get alpha floor protection which triggers TOGA thrust even with the autothrottle disabled. I would have hoped newer Boeings offered similar protections, the 737 always struck me as being super "analog" compared to the A320.
It doesn't matter was in system. You run your checklists and configure for your own flight / your own departure etc. Memo should be to training / recruitment - where did you find these pilots to fly a 777?
So they obviously didn't pre-flight the flight. And they don't seem to have used their takeoff checklist either? That should get them through flap retract and 2,000 feet or so.
Takeoff checklist might have rotate at 160 or so. Then you get to positive climb and V2. You then clean up the plane (flaps / gear etc), maybe hit 2,000 feet / 210kts?
This is all checklist stuff. I've no idea what's Emirates has as checklists, but this is like 101 entry level flying stuff. What was pilot monitoring doing during all this?
They still on ground at 215 kts? 262 kts at 175 feet?!