The Silent Anniversary: Fifteen Years Since Our Last Major Crash | Hacker News Reader

The Silent Anniversary: Fifteen Years Since Our Last Major Crash

askthepilot.com

272 pointsmaxcan163 comments

The effect of culture should not be underestimated here.

There really is a cooperative spirit among pilots, air traffic controllers, weather briefers, mechanics, the FAA, etc. Controllers are willing to help pilots with special requests if able, pilots file weather reports for other pilots if what they encounter aloft is different from what they expected, mechanics take pride in their work and are highly regarded by pilots. I generally take a dim view of regulators, who generally seem to have less expertise than the people they're regulating. That's mostly not the case here. I'm libertarian and a pilot - and I like the FAA. Nobody is out to nail you for a minor technical infraction, but they will yell at you for something that's unsafe but technically legal. This encourages people to report near misses, even if casts them in a bad light. The regulations themselves are generally reasonable, and frankly feel like they were written by pilots, controllers, and airport operators.

The point is this - regulations matter and budgets matter, but if the culture is toxic, people will find a way around it anyway. When I see a problem elsewhere in society - investment banking in the 2000s for example - people propose heavy regulatory solutions. That may work, but wouldn't be so much of a battle if the culture were better.

I haven't thought deeply about what "better" means in this context, nor about how to intentionally change a culture. I'd rather just fly.

> I'm libertarian and a pilot - and I like the FAA

Me too. For every frustration about what they've done to the costs of things in GA, and those are valid, their accomplishments in general are incredible.

I've often heard it said that FAA regulations are written in blood. Meaning, if a rule exists, it exists for a damn good reason.

For anyone interested, NASA's ASRS [1] (Aviation Safety Reporting System) is what the parent commenter is referring to regarding reporting near misses. The idea is, anyone can submit an incident report with NASA, a neutral separate party from the FAA, which is placed in an anonymized public database and cannot be used to justify punishment against anyone involved in the incident. This promotes transparency and learning from the mistakes of yourself and others, instead of covering them up. Frequently, the person who caused an incident in the first place will promptly report it to ASRS, as the FAA looks favorably upon those who come forward to ASRS on their own and dimly upon those who do not.

I think more industries could use this kind of transparency in reporting sketchy situations.

[1]: https://asrs.arc.nasa.gov/overview/confidentiality.html

I often remarked in the same vein to my daughters as they were growing up and learning to drive. Everywhere you encounter a stop sign, a light, etc., is a place where people were badly injured or killed on multiple occasions. That stuff is there for a reason.

New road construction starts with stop signs and traffic lights. They don't just leave it a free-for-all and add them later on.

Now, when you see quadruple stop signs (i.e. 16 total) and a jury-rigged flashing red light attached to them, as there was in my home town, chances are that one came about after some accidents.

There's this place in Atlanta where one makes a 180 turn from one interstate to another. They added flashing lights, paint all over the place, signs. Still, there are so many vehicles going off the road that no grass ever grows there. Here's the Google Street view of the dirt patch, complete with vehicle debris which accumulates almost every day.

https://www.google.com/maps/@33.7939354,-84.3917037,3a,75y,1...

At what point do you change the turn? Signs and paint is great, but maybe the fundamentals are just bad.

That road is knackered too - is that a commmon state for roads around there?

The roads in the US are in general horribly maintained. I've driven across it north/south 8-10 times and east/west 4 times along different routes/not just freeways and it's rare to find great roads. Michigan was the worst but luckily starting/ending in Ontario we weren't often on their roads for long.

Sorta depends where you're at. In cities that have hot summers and cold winters (Spokane, etc.) you get terrible roads. Poorer parts of cities have worse roads than richer parts (compare north Tacoma to basically the rest of Tacoma).

In suburban, high-traffic†, low-traffic†, or moderate climate areas the roads are fine. In fact some of the freeways around where I live have been putting in the "quiet asphalt" stuff— driving is quiet, even in my 2000 Corolla!

† High traffic areas get attention because they get beat up easily. Low traffic highways do not get as beat up. This still kinda depends on who maintains the road, how rich the surrounding area is, climate, etc.

Weather is really important. For all the traffic it has, the freeways in LA are generally not that bad. 300 days of sunshine and no freezing temps helps a lot.

That's why I specified how widely travelled I was. I definitely understand weather related issues having grown up in Ontario with a 50-60c swing in temps over the course of a year + salt. But I saw lots of awful roads in the north east, the south, midwest, south west, west, and north west.

It's pretty safe to say that in general US roads are poorly maintained. Like saying in general UK rail is poorly maintained.

From street view that doesn't seem that difficult to negotiate - is it just distracted/low skill drivers? Speeds too high?

On a side note - zooming out from that, the sprawl of Atlanta looks unlivable. On several nearby highways I counted 18+ lanes!

Having negotiated this particular turn many times, I'd guess it's a combination of factors. Atlanta is spread out so the highways are typically long and straight, so apart from rush hour the fast lane typically goes 70-80 mph comfortably. Even on highways, you're usually ok to reduce speed to 60 mph to make interstate turns, e.g. spaghetti junction (85/285). Interstate turns this sharp are very rare so even natives need to be alert. I-75 is a long multi-state road so many out-of-towners may not be expecting the sharpness of this turn. Not sure if this plays into it, but it does also exit from the fast lane.

There's a "Right Lane Ends" sign right there, and I'd imagine people unfamiliar with the turn see that sign and their brain just can't catch up with trying to process both slowing down for the 25mph turn and a left merge.

Might be better to have the left merge occur before the turn, then do the turn. Otherwise you're doing several things at once: Merging, slowing down, and maintaining direction around a turn with a changing radius. Two of those would probably be challenging at once.

Ha, I grew up in Atlanta but left 12 years ago and haven't been back. I immediately knew exactly the spot you were talking about.

Sometimes it's quite explicit: in Korea there are signs saying 사망사고 발생지점, literally, "Place of a fatal accident".

The problem is that sometimes that part of the road isn't even particularly dangerous, so I wonder what was the rationale for forever reminding people that somebody died there...

I remember little white crosses on the side of the road in Montana. If a turn had a number of crosses that was a pretty good indication that you really need to be careful.

In Costa Rica, for a number of years a few years ago, they would paint a hollow yellow heart, with a halo, inside a box, on top of the part of a street or road where a fatal accident occurred. However, after some time, they stopped repainting them after each time the road was resurfaced.

Obviously that's not even remotely true, so why are you telling your daughters that?

Don't lie to your kids. They're smarter than you think.

Napkin math suggests you are wrong, the grand parent is right.

I can't find the exact stats but in the US there are 2+ million injuries per year caused by road traffic crashes. We know there are around 300k signalized intersections [1]. If we estimate non-signalized intersections (ie. those with STOP signs) to 20x that number, that's 6 million non-signalized intersections. If 20% of injuries happen at intersections and are randomly distributed across intersections, then on average there will be 1 injury at a given intersection every 15 years.

So yeah it is likely that "multiple injuries" have occured at mosts intersections in the country.

[1] http://mutcd.fhwa.dot.gov/knowledge/faqs/faq_part4.htm#tcsgq...

Where I live, there are at least a couple hundred stop signs and two (2) traffic signals. That's probably closer to a typical ratio for suburbs and smaller towns. I doubt there's ever been a serious accident at most of those intersections, and certainly not as the catalyst for putting up all the signs. The major accidents happen on the highways outside of town, during blizzards.

Perhaps you should ask your local roads department how they decide where to put traffic control structures before talking about lying :).

If you think the answer to that question is going to be, "Whenever somebody gets hurt," then I'd say you're the one who needs to place that particular call.

We know that highway speed limits aren't generally optimized for safety, for instance: https://www.motorists.org/issues/speed-limits/faq/

So in your mind how come brand new roads have road signs and lights? Nobody has been killed or injured on a road that hasn't existed before.

And don't you think they could apply some historical knowledge that a new road similar to an old one could use similar signs? Do you think the actively wait until someone gets killed before they add any signs to obviously dangerous new intersections?

That's certainly true for new construction, so I should probably qualify that I live in the northeast, where totally new roads are very rare, and almost all of our roads evolved from much earlier thoroughfares. The reality in other parts of the country is probably much different (and informed by our experience here, where the first roads in the U.S. were constructed). In this part of the country it is indeed true to say that almost all traffic control structures were placed because a particular spot caused accidents.

This is some piece of reading about the odd history of government and aviation... https://en.wikipedia.org/wiki/Air_Mail_scandal

"elsewhere in society" also education, health care

It helps that everybody's economical interests are aligned into a hight safety level.

This environment would never exist in an area where companies could be created overnight, sell something for a while, and safely fail before their social (PR) costs become obvious.

If bank executives had a device implanted in their chest that would explode if their bank ended up insolvent they'd probably run things very, very differently.

There is also a fair bit of "systems thinking" in aviation, which focuses not on who to blame when things go wrong, but how were systems set up to permit failure. Barring gross negligence or malice, most things go wrong for reasons other than the immediately involved actors.

> Barring gross negligence or malice, most things go wrong for reasons other than the immediately involved actors.

Anytime I hear of systems failing, I can't help but think about Dr. Cook's paper on How Complex Systems Fail - http://web.mit.edu/2.75/resources/random/How%20Complex%20Sys...

Culture yes, but also consequences.

The consequences for a failure in aviation, is often death, and this is probably why everyone is on their best behavior. Consequences for a failure in banking policy are less obvious.

I get the feeling that people don't regard wiping out someone's savings and destroying their retirement, as grave as death, and so profiteers will either consciously or subconsciously, use that as justification for their actions.

> The consequences for a failure in aviation, is often death, and this is probably why everyone is on their best behavior.

Well, except for the pilot who took off with inadequate fuel and killed a football team.

> In a sad coincidence, that this article was posted on the same day that a plane crash in Colombia killed 71 people. Not every region of the world has been as fortunate as ours, and we are wise not to gloat.

The consequence of death does absolutely not discourage people. Several people die everyday in the US because they were texting while driving. The problem with the consequence of death is that no one believes it will happen to them.

Or rather, we have a terrible intuition about relative probabilities. Explains not wearing seatbelts, playing the lottery, and lots more.

I think fear of death is a universally strong motivator. The problem with drivers is that they're allowed to drive with a pittance of training, and not fully educated about the various ways they can put themselves at risk.

Also there's a rationalization that occurs even in drivers who should know better: even if you put yourself at risk and get into an accident, many types of auto accidents, even catastrophic ones, are survivable. A pilot would never make such a rationalization because aviation accidents have much worse survivability rates.

> I generally take a dim view of regulators, who generally seem to have less expertise than the people they're regulating.

The thing that always worries me is — if we try to prevent the "revolving door" whereby regulators hire from the industry they're regulating, then how will we ever have regulators who know what they're doing?

Most of the problem isn't people retiring to a government job, it's people leaving the government to a private sector job.

Exactly. The promise of a well paid private sector sinecure after your time in public office is what creates the perverse incentive for regulators.

> When I see a problem elsewhere in society - investment banking in the 2000s for example - people propose heavy regulatory solutions. That may work, but wouldn't be so much of a battle if the culture were better.

The toxic culture of investment banking is a feature, not a bug. If you cleaned it up, it wouldn't be high finance anymore.

No, for a long time starting with the new deal, we had high finance with a much less toxic culture. The change happened because a bunch of conservative politicians persuaded the government that high finance is natural self-regulating and so it would be just fine to gut many of the key regulations. The crash of 2008 was one consequence. And Trump is very much on the financial deregulation side.

" That may work, but wouldn't be so much of a battle if the culture were better."

The culture of guys moving people around in the air or the culture of guys making bets

As a former Boeing engineer designing flight critical systems for the 757, the safety of airliners is deeply embedded in the culture at Boeing. Nobody wants to be responsible for a fault that results in a crash. Furthermore, Boeing is acutely aware that if an airliner design acquires a reputation for shoddiness, unsafety, etc., it can very easily put them out of that business (that has happened to other airframe companies, look what happened to Lockheed with the Elektra).

They got the largest military aircraft contract ever?

https://en.wikipedia.org/wiki/Lockheed_Martin_F-35_Lightning...

In all seriousness, I've done a floor tour of the boeing factory, its really an incredible place. The work you and all the engineers do there is amazing. Thank you.

The Elektra's reputation caused Lockheed's exit from the airliner business.

(Getting a tour of the Boeing factory floor is really quite an experience!)

Fair point.

The tour was amazing. AOPA set it up as part of the Bremmerton (sp?) fly in. Coming from the software world, seeing the efficiency and precision of the floor was unreal.

> a problem elsewhere in society - investment banking in the 2000s for example - people propose heavy regulatory solutions

Air flight is not "adversarial", has the ability to directly and spectacularly kill people, and doesn't have the same upside potential for making extraordinary amounts of money. You can cut costs, but only asymptotically towards 100%.

Whereas in finance, the potential upside to misbehavior is unlimited, the negative side effects are indirect, and there is already an adversarial culture of "beating" other people in your deals. So the regulatory culture ends up having to be adversarial as well, because you can't get people to give up multi-millions for the sake of playing nice.

What wasn't mentioned directly was that since 2009, US passenger airlines have had a perfect record of zero fatalities, which is mind blowing.

Especially mind blowing considering that not all that long ago, it was common to see hundreds of airline fatalities per year, despite much reduced traffic. If the 1960 level of safety carried through to today, we'd see over three thousand airline deaths each year.

Furthermore, fatalities are still in the hundreds world-wide, often with the same equipment flown by US airlines.

This number tells us that the FAA+NTSB, despite some other criticism, _usually_ has its act together where it counts the most. They iterate on identifying safety problems and force the airlines to adopt solutions.

US airlines often sell their older planes to foreign airlines, which may account for some of the increased issues those airlines have. Your point still stands though - culture and processes rule.

You could argue that those planes should be forced to retire instead.

A different perspective might be to require that every 10000 hours of service see a new mark painted on the hull. This way passengers could be informed of how much 'age' a plane has had and the regulations could also apply to any plane which /might/ fly in to the US (or Europe).

The hull's age only tells part of the story though; how well it's maintained, and the ability of the pilots flying it would need to be factored in as well.

Those are probably far more important. I'd be far more comfortable boarding a 30-year-old airliner with great maintenance and great pilots than boarding a brand-new 787 with a clueless maintenance crew and pilots who aren't quite sure which way to move the stick (edit: yoke) to make the houses get smaller.

If you look at the causes of crash, it's usually pilot error or improper maintenance, occasionally design flaws, but not equipment age.

It brings to mind a flight just last week that crashed due to likely negligence (not carrying enough fuel).

http://www.cnn.com/2016/11/29/americas/colombia-plane-crash-...

So it turns out not only was that guy negligent, but there was also another plane in the pattern who got priority because, at the time, it looked like that other one was low fuel also (maybe a false alarm).

Here's a reputable source.

https://fearoflanding.com/accidents/not-enough-fuel-the-disg...

Plane crashes have been going down not only in the U.S., but worldwide, which is commendable.

A bit unfortunate for this to be posted a few days after the Chapecoense crash shook the world (noted in the postscript), but that was a regional charter flight, not a major U.S. airline.

Interesting infographic:

http://www.ibtimes.com/how-many-planes-crash-every-year-how-...

Edit: the Chapecoense crash shows exactly why the FAA's work is so commendable - disregard of proper procedure and safety measures led to that accident.

About the Chapecoense crash. From [Portuguese] http://www1.folha.uol.com.br/esporte/2016/12/1838707-lamia-j... Autotranslation: https://translate.google.com/translate?sl=auto&tl=en&js=y&pr...

> On at least two occasions, the Bolivian company had flown beyond its declared autonomy. [4 hours and 22 minutes]

> One between Cochabamba in Bolivia and Medellin in Colombia (4 hours and 27 minutes), and the second between Medellín and Santa Cruz de la Sierra in Bolivia (in 4 hours and 32 minutes).

Wouldn't the difference in 5-10 minutes not be the reason for the lack of fuel, though? I would guess a 4 hour flight would have enough fuel for at LEAST 5.5 hours if they couldn't land at an airport and had to be diverted to another one (right before landing).

Flying is so amazingly safe now that you're far, far more likely to be killed on a cab ride to the airport than on the plane itself.

"... the first officer of flight 587, Sten Molin, who was at the controls, overreacted, rapidly and repeatedly moving the widebody jet’s rudder from side to side,... . The vigor of his inputs caused the entire tail to fracture and fall off."

I was interested in learning more about this, and it turns out there's an alternative interpretation that argues he was flying according to best practices at the time rather than "overreacting": http://articles.latimes.com/2005/jan/11/opinion/oe-garrison1...

IIRC from watching the episode of "Air Crash Investigation" (aka "Mayday") on this flight, it was a pretty tragic confluence of factors: it was a best practice on other planes, but not on the A300, but American Airlines training manual had not been updated to reflect this change, nor did the simulator models account for this eventuality. Kinda surprised the author of this piece wrote it up as he did.

If I'm remembering the details of the case correctly, it turns out he did overreact and that did cause the tail to fail: Certainly the tail did fail, and the pilot's input caused it.

The important part and the part that has been noted above is the NTSB's (and the industry's) focus on not blaming the pilot but figuring out why the situation turned out the way it did, what factors allowed it to happen, and how things could be changed to insure the same chain of events doesn't keep occurring.

(Again if I recall correctly), It was the confluence of a few different events compounding each other and resulting in the accident:

One factor was the airline company sim training on recovery from upsets. Basically the pilot going through the training could come away learning entirely the wrong lesson. The simulation was hardwired so that the pilots' inputs were disregarded until the plane was in an upset and then the pilot took over. Pilots unfortunately seemed to learn that control inputs before the upset (which were ignored) affected the outcome, so they might randomly use large control inputs which were unrealistically ignored in the sim. Pilots also came out of training with an exaggerated expectation of wake turbulence because they were unable to control the plane during the early part of the upset.

An additional problem is that simulators are based from data gathered during aircraft testing and represents the normal flight envelope of the craft. Using the simulator outside this regime can result in aircraft responses which do not simulate the actual aircraft responses.

Another issue was the way the limits of the rudder pedals were handled between different revs of the A300. Like the steering wheel of a car the rudder pedals require less and less input to be effective as speed increases. To prevent overstress on the airframe, the mfr will somtimes shorten the distance you can physically move the rudder pedals. Another way to achieve this effect is to change the ratio of rudder pedal movement to actual rudder movement (at cruise speed, a given movement of the rudder pedal will result in a smaller rudder movement than at low speeds). The mfr changed from one scheme to the other between A300 models and additionally lightened the amount of force needed to move the pedals.

All this added up to a pilot making the same rudder inputs and getting a vastly different rudder response depending on the revision of the A300 they were flying.

Finally, I believe there was concern with pilots' understanding of the V_A speed (one of many V_* speeds that the mfr provides in the aircraft's operation manual). V_A is the maximum aircraft speed at which you can jam one of the flight controls to its stops and keep the plane in one piece. Most pilots the NTSB surveyed didn't realize that this only applied to a single maneuver/single control input. Putting the rudder pedal to the stops back and forth repeatedly, even under V_A would allow the energy to build and eventually cause the airframe to break up.

Full NTSB report: http://www.ntsb.gov/investigations/AccidentReports/Reports/A...

I think the NTSB reports are fantastic reading for anyone interested in the failure modes of people and machines as well as how to really approach and investigate difficult problems.

There's likely no media coverage of the anniversary because only including major airlines is a very arbitrary cutting off point.

The NTSB lists five major passenger air accidents since 2001:

http://www.ntsb.gov/investigations/data/Pages/paxfatal.aspx

No agency breaks out "major" airlines from "regional" for the purpose of regulation or statistics, nor does any other nation. It is purely self-serving.

What makes this post weirder is while cutting out regional airlines from the statistics, the causes he attributes with improving safety for the majors are the same things that regional airlines have also been doing.

Of those 5, 4 occurred in propeller craft. I think that it is relevant to include that kind of difference. Regional airlines and major airlines do work differently.

He does mention that in the post, and, yes, I think the author would also agree that it's self-serving. However, I would also argue that this is a symbolic milestone of the increased safety of US commercial flights, and an interesting thing to reflect on. I wouldn't take it too literally as any sort of real statistical measure, but it's certainly worthy of a blog post.

"2009: A Colgan Air (Continental Connection) Dash-8 Q400 crashed outside Buffalo. Fifty people were killed, including the occupant of a house struck by the plane."

"2006: Forty-nine people perished when a Comair (Delta Connection) regional jet crashed after attempting takeoff from a too-short runway."

I'm not sure, maybe "major" has an official cut off number, but 49 or 50 doesn't seem "minor". It's a long way from 260, but it's far from 0.

They're counting Major US airlines, not including regional carriers

FTA:

>To be clear, there have been a handful of tragedies involving regional carriers and freighters [...] It’s the majors that set the standard, and it’s only fair that we measure from there.

Colgan and Comair aren't counted in the metric because they're regional, not major/legacy, airlines. They codeshare with the majors, but they're different organizations.

Tangential: Frontline had a great episode on regional airlines' safety issues. http://www.pbs.org/wgbh/pages/frontline/flyingcheap/

Tangent on the tagent: Phillip Greenspun touches on regional airline safety in this post comparing US to foreign airline safety - http://philip.greenspun.com/flying/foreign-airline-safety

If having a newbie in the right seat represents a challenge to safety, how do U.S. regional airlines manage? The U.S. regional airline is hiring pilots who are more experienced than those hired by foreign airlines (the U.S. regional F/O typically has been a flight instructor) but less experienced than those hired by a U.S. major airline. The answer is that U.S. regional airlines have an accident rate as much as 5-10 times that of U.S. major airlines. The culture is the same, but apparently there is no substitute for experience.

From the comments: "The total number of departures is split just about evenly between the majors and the regionals. Yet the regionals, at least over the past several years, are carrying a much higher share of the accidents."

This, I think, is the issue here. Some of the commentators have suggested (plausibly, I think) that the number of legs per day flown by regional aircrews may be an issue. We should probably also consider experience.

> the number of legs per day flown by regional aircrews may be an issue

More take-offs and landings, more human and equipment fatigue, more opportunity for mistakes and breakages?

Most importantly the human fatigue. There have been some very alarming reports about regional airlines requiring their flight crew to work very long hours and scheduling them such that it's nearly impossible for them to get a proper night of sleep between working days. They're remaining compliant with the regulation, but in general getting as close as they can to violating. The major airlines, on the other hand, tend to have a more senior and thus more demanding flight crew (in terms of work life balance) so most flight crew have significantly more generous schedules.

The regional/major distinction doesn't really exist from the perspective of most passengers, though. The article isn't wrong, but its certainly a little misleading

I don't know about that. I can usually tell when I'm on a regional. The equipment is different, and the planes smaller. There may be one or two regional airlines that fly major-grade equipment like B737s, but my experience has been that they fly different equipment.

> They codeshare with the majors, but they're different organizations.

Indeed they are organisationally different, but to the passenger holding a United through-ticket from Smallville to LA there's no difference except one part of the journey is on a 'tiddly airyplane'.

If they're good enough to carry a major's flight code and livery and do the grunt work for feeding hubs for them, they're good enough to count in accident statistics.

Would we accept lower security standards for, say, AWS if that was outsourced and was only 'branded' by Amazon?

Oh, I don't disagree, and I get the jitters every time I get on one of those regional jets. I was just trying to clarify the parent's confusion.

If you include them, the picture is still remarkably safe. How many years since a regional airline fatality?

At least in the Buffalo case (crash was down the road from house I grew up in) the pilots didn't know what they were doing and it was a very cold night with ice on the plane.

If I remember correctly they pulled back when flying too low and the plane dropped - was a freshman in college at the time though and don't really remember specifics. The recording recovered of the pilots didn't inspire confidence.

This would play into the distinction the article makes about 'major' airlines and the culture differences probably.

Children of this post are discussing/arguing about the regional vs. major distinction ... ok, fine ...

But, in relation to:

"Forty-nine people perished when a Comair (Delta Connection) regional jet crashed"

I continue to read:

"after attempting takeoff from a too-short runway."

What the fuck ? How do you manage that ? That sounds like some pretty broken procedures there ...

I know perfectly well that a fear of flying is often an irrational, though very real, fear for many people. But it's exactly the irrational nature of it that boggles my mind: they should be terrified of driving, but they aren't.

I'm way more comfortable flying than driving.

It's weird how the brain works.

I know very well that airliners are safe, and I'm aware just how much work goes every day to ensure they stay safe. And yet, every time I'm flying, I have this uncontrollable thought, "maybe this is the day we all crash and die". I always get a little anxious during takeoff and landing, I listen to all the noises and worry about them. "This doesn't make sense, the plane didn't usually make that sound". "Oh shit, the lights blinked, wiring must be damaged." Stuff like that. It usually takes me few minutes after liftoff to get those thoughts out of my head.

At least for me there's a feeling of autonomy. If a plane crashes, there's very little that I as a non-superman can do to stop that from happening (with me inside it.) While driving, I can trust my own reflexes, awareness, and decision making to put myself in better or worse situations. Now this may still absolutely be a false perception, but this is where the feeling is rooted for me (from lengthy introspection on similar situations when I used to do trick-skiing and never really questioned my safety, but would be very hesitant to e.g. skydive).

To some degree I think it's potentially also "failure magnitudes." Having been in a handful of very minimal car crashes, that calibrates my "something goes bad" prior to a not very impactful bar; whereas if something goes wrong on a plane, it can go VERY VERY wrong. Again, this is likely a false perception due to observations of outlier events in both cases and sparsity of signal, I'm just trying to question the origin of my feelings.

There's this thing that will happen to me when I doze off as a passenger in a car: When I wake up, there's often a split second of panic between the visceral understanding that I'm in a moving vehicle and the slightly delayed recollection that I'm not the one responsible for driving it.

Somehow, this turned into a problem for me when flying (even though I'm not a pilot). I wake up, I instantly know I'm moving at 500 miles per hour through the air, then I have a singular panic of being in a vehicle not under my control. This has only happened once – I haven't slept on a plane since.

Speaking of irrational, when flying I think, "What if this is the day I discover my superpower to walk through objects and fall to the earth." I think that's about as irrational as I can get for a fear.

Curious phobia! Do you imagine yourself disintegrating on the surface of the earth or do you continue to fall through it?

Hah yeah. I have no idea why I think of that in planes. I assume it would even be worse on ground level, fall into the bedrock and never come back up.

It must be something from my childhood and stories or TV shows, and just stuck with me. I know I had recurring dreams when I was young, a futuristic belt that let me jump through walls, but detected and kept me on the floor without falling through.

I tried to find the stats on this and haven't found the right ones yet. There are odds of dying in a lifetime, which is flawed because people drive more than they fly. There's also deaths per mile, which isn't right because planes go faster than cars. It seems that what people care about is what is the probability of dying in a time period of engaging in the activity?

Are there stats for driver-deaths-per-time-spent-driving? It'd also be interesting if it could be broken down to driver-deaths that 100% clearly weren't the driver's fault. I wonder how that compares to deaths-per-time-flying (via a major carrier). Or probability of death.

Best I could find is the micromort wikipedia page [1] - apparently to spend one micromort, you can drive 230 miles, fly 1000 miles, or train 6000 miles. Looked at that way, you're burning micromorts at a faster hourly rate on the plane, so according to that it's riskier if you squint your eyes just so.

[1] https://en.wikipedia.org/wiki/Micromort

Hmm.

If I'm reading that page right myself, then if you travel 1000 miles in a car there's ~4.3 times as much risk of death than if you travelled in a plane instead.

So if a plane travels at least 4.3 times as FAST as a car on average, then your risk over time in a plane could be higher.

I think your post makes a good point in general. Way too often, people compare statistics that don't consider major factors like how often people drive vs. fly.

Couldn't you just divide the deaths per mile by the typical speed of an airplane or car to normalize the data?

That's probably a good first approximation, genuinely wonder about how accurate it gets. How many people die in LA gridlock?

If your driving experience is "hours moving at 5mph", and your flying is "lots of shaking", easy to get scared of one.

As someone who has developed a fear of flying (I still fly because I'm not taking a boat but...), one small thing that scares me is looking at plane wings in flight.

A lot of bending, and a lot of parts that look old. On one flight I could see a piece of rubber stuck on the rudder , that seemed to slowly get more worn out as the rudder moved up and down.

I'm sure it was by design, but at the time, I was legitimately scared.

EDIT: it was a flap, not a rudder

Modern aircraft have engineering tolerances that are absolutely absurd: http://www.popularmechanics.com/flight/g2428/7-airplane-wing...

The shake and wobble in wings is by design. If they didn't flex they'd shatter.

People have a place they want to go when they travel, so deaths per passenger mile is much more relevant than deaths per hour. Otherwise you could reduce your risk of being in a car accident by driving really fast and spending less time driving!

When you go to visit someone or to an event, the place is fixed. When you plan a vacation, travel time is a factor (certainly not the only one)

Nervous flyer here.

I tell myself this on a loop while flying and my lizard brain just does not care. Weirdly, I've flown hundreds of times and this didn't start until I was near 30.

I think one aspect is not the fear of death, but the fear of being in a situation of knowing death is certain for minutes as you fall out of the sky. (Which is a situation I also know is a small fraction of the already tiny chance of dying in an airplane).

Completely agree with you. I fly tons for work, and was just fine until I got into my mid-late twenties when I started to have horrible anxiety.

And like you, I don't necessarily fear death, but I definitely don't want to see death coming. More than anything, I fear helplessness.

When I was a teen (long time ago) I remember telling my dad that I thought it would be great to take a flight with many stops because that way we got to takeoff and land multiple times! What fun! A bit later of course it became exactly the opposite and not just for time travel reasons, price or schedule.

Ditto. Was traveling for a living, and didn't care at all about it for years, then I got worse and worse, and had to change careers. I think part of it was the increase in stress overall after 9/11 in travelling, but maybe another part of it is doing it enough times, having enough unpleasant experiences that it's added up to me just hating everything that can happen on a typical flight.

Also, dropping 10-20 feet and seeing/feeling the frame of the aircraft flexing about is just plain unsettling.

> dropping 10-20 feet and seeing/feeling the frame of the aircraft flexing about is just plain unsettling.

If it helps, Boeing tests the wings to failure by mounting them to a diabolical torture device that bends them. They have to be bent near vertical before failure.

A little turbulence flapping the wings is about as dangerous as your car's suspension flexing over rough roads.

Have you tried taking a flight lesson to see if it helps?

I used to be a white knuckle flyer until working at Boeing. Knowing how the systems are designed and how they work and how they are designed to be resistant to failure helped a great deal.

I tried to explain all this to a friend who was also a nervous nellie flyer, and he told me he didn't want to know - he felt safer with it being a black box.

I'm sorry, but if I'm in a tin can 30,000 feet in the air, defying gravity, I'm going to freak out at the smallest bit of turbulence, whether it's rational or not.

Moving at 100 mph in a car does not seem exactly natural either.

My lizard brain disagrees with you.

A "lizard brain" (besides being a hypothesis largely dismissed by neuroscientists) seems unlikely to have any reference point or other handling for "I'm sitting in a tube 30,000 feet up at 700mph". Similar to how we're basically incapable of thinking about numbers like a billion dollars effectively.

Interestingly there is a great deal of work that weather services, air traffic control, and pilots do to avoid turbulence for this exact reason.

I think people fear the absence of control, and the absence of understanding how flight works; even though there is a lot that can happen to you driving, you or a person you trust is usually the pilot operating a vehicle who's physics you are familiar with.

As another commenter said, for me it's about my personal lack of control over the outcome -- that you're completely at the whim of others (the pilots, the maintenance crew, air traffic control, etc.), which compares unfavorably with the illusion of control one has with driving.

Traffic fatalities do occur to people who were not 'at-fault', and despite there being lots of good statistics on driving accidents [1][2], I can't any good ones on the ratio of serious injury or death for non-"at-fault" drivers vs. at-fault ones.

Meanwhile, it would be illuminating to interview people who never drive themselves -- rather, they are always passengers -- whether they feel safer being driven, vs. flying, and the like.

[1] https://crashstats.nhtsa.dot.gov/#/ [2] http://www.iii.org/issue-update/auto-crashes

While I get that, it's a little odd to have that "illusion of control" in a situation where a legally blind 90 year old with mild dementia might well be barreling down the wrong way of the highway towards you.

Others downthread have mentioned it, but you still have an element of control in that situation. I've seen someone driving down the wrong side of the highway, and I pulled my car over and didn't get hit.

Even if the % of control you personally have is small, as long as it's > 0%, you feel like you're in control of the situation. In a plane, it's 0%.

Sometimes. There are situations where you're going to have no real options. A drunk driver could sideswipe you into a pole with zero warning.

The chances of your plane running into another plane piloted by a drunk is pretty slim in comparison.

True. But in a car, it typically takes 2 people doing something wrong for a wreck to occur. I prefer to give myself a chance and be one of those with a choice, and to be able to train myself, and so on. On a plane, it's outside of my control. Furthermore, the failure is much more likely to be mechanical, and often at that point there's nothing anyone can do, even with the best-trained pilot in the world. So many people (maintenance personnel for instance) have to do their job well to keep the plane in the sky.

I also have almost no fear of being in a car wreck because I still find it highly unlikely a wreck would kill or seriously injure me. I have a safe car and have lost (and regained!) control of cars plenty of times in my life without incident. Hell, I even choose to race cars in circle for fun! I only fly because I HAVE to.

Motorcycles obviously feel more dangerous to me and I have greater reservations whenever I ride one.

Ironically, I am posting this from a UA flight.

I was once on a flight that lost an engine. That was fun. We took a "normal single-engined" tour of LA, and a "normal single-engined landing" 45 minutes later. The lack of concern was both reassuring and terrifying. I expected the pilots to spin the plane on a dime and drop it back down the runway the same way we came, emergency landing, rather than keep the thing flying on one engine.

If you're just after any % so you can have a semblance of illusion of control, you could always spend some time researching safest seating positions in planes, how to position ones self to change direction in a free fall, and ideal landing terrain.

Your ability to influence outcome in a plane accident are minuscule but they do technically exist.

Yes, as an anxious flier/safety obsessive I do this. I would say they are actually more significant than you would estimate. Even catastrophic crashes can be survivable, and most crashes are not the plane falling completely out of the sky but rather mishaps around takeoff or landing at slower speeds and involving fire-related fatalities on the ground.

Here's a list of things you can do to increase your personal safety while flying:

* Sit as close as possible to an emergency exit.

* RTF(safety)M

* Know how to assume a brace position, including ideally the particular brace position which is optimal for your body shape/size in your given seat.

* If you're particularly paranoid like me, you can preemptively assume the brace position on takeoff and final approach/landing. Full disclosure: I don't have any evidence that this is beneficial and people may think you're crazy, so there's that, but it makes me feel better.

* Wear sneakers and keep them on or at least in reach at all times.

* Wear layers, preferably made from less flammable materials.

* Don't be intoxicated. I break this rule by having some drinks to ease my anxiety, but the point is to not get to anywhere near the point of incapacitation.

Interesting - I never drive myself and I feel far more nervous about being in a car than my driver friends. I think there'd be confounding factors about who drives though. Perhaps you want to measure whether people who sometimes drive feel a significantly different level of safety when driving or being driven?

It's about feeling that you are in control. You do not have control when you are flying because you are not a pilot. You are giving control over your safety and life to other people. When you drive a car you have a feeling that everything is under your control (which is not true most of the time).

People know they're statistically safe when they go to a scary movie or a haunted house too. Fear is a primal emotion, not weighed to meaningful statistics.

Airline passengers are trapped in a cramped, shaking metal tube 30,000 feet in the air. I'm surprised fear of flying isn't more common.

Yes. It's completely rational for people to be scared out of their wits in an airplane, or higher than say twice the distance of a tree. No primate is in either situation, it's just not natural. What's different is the ability for many people to distract themselves, they have a more effective coping mechanism. There really isn't much distraction on an airplane for people who either do not read magazines/books, or watch tons of TV and movies. That's pretty much the only distraction an airplane has to offer.

> they should be terrified of driving, but they aren't

Driving you can control for everything pretty much but the 'other guy'.

When flying you are depending on multiple 'other guys or gals' all doing their job correctly, not making mistakes in multiple areas where there is no way for you to verify. So the unknown factor is considerably higher, regardless of statistics.

With driving if the weather is bad you can typically drive later/earlier a different day. With flying that is not the case. You are locked in financially to perhaps taking a bigger chance than you might if you could just work around the weather (in two cities).

> Driving you can control for everything pretty much but the 'other guy'.

Yes, and all the evidence says you're terrible at it.

I don't know about that. Nearly every adult American drives nearly every day of his life, and only about 30,000 die each year. That's still too high, of course, and I look forward to autonomous cars, but it does mean that while many of us know of folks who've been in fatal accidents the vast, vast majority of us won't.

Yeah, I think it's because people feel in control when they drive, whereas on a plane you're just sitting there. Of course you actually don't have any control over what other drivers do, but the feeling of control is there. Also, a car accident may not kill you, but if your plane crashes, you're dead pretty much.

I can't explain this, but on the day I'm going to fly somewhere, I'm scared about it from the moment I wake up. Maybe 'extreme anxiety' would be a better word, since I'm not actually afraid that I'll die at that exact moment; just anticipating it later on.

I remain in that state during the whole trip to the airport, during check-in, while going through security screenings and sitting in the departure lounge, and while walking through that arm thing that connects to the plane, right up until the moment that I actually sit down in my seat and buckle in.

And then I'm fine.

Happens every single time. And I've no clue why I could be terrified of flying when I'm on the ground, but totally fine when I actually am flying.

I get this too. The second I close the zipper on my suitcase I start to feel anxious until I'm in the skyway. I wonder if missing a flight would cure it

I am surprised there wasn't more than a token mention of terrorism. People like to bemoan airline security and I think there is a valid argument regarding the effectiveness of the TSA, but this is also the safest stretch the US airline industry has had against terrorists and hijackers since the early days of the industry in the 40s and 50s.

I think the current level of airline security is overkill, but I'm always a little surprised to see people advocating for having none. In the late 60s and early 70s, US airliners were getting hijacked at a rate of once every week or two. Once basic security screening was introduced, it became far more rare.

Making people take off their shoes and empty their water bottles is a bit silly, but basic checks for guns and bombs are a good idea.

    In the late 60s and early 70s, US airliners were
    getting hijacked at a rate of once every week or two.

https://en.wikipedia.org/wiki/List_of_aircraft_hijackings is pretty dramatic

Hijacking was pretty different back then, which is why passengers were used to the idea of staying uninvolved and not escalating things. That's part of why 9/11 was just a (culturally) unimaginable shift, it wasn't really in the gestalt that a hijacker's main goal was other than to kill/die.

Yeah, 9/11 could work exactly once (and even then, it only in three out of four concurrent attempts attempts)

Keep in mind too, there are also things like air marshals and bullet proof cockpit doors that can't be opened from the outside and other measures like that.

Not all the security improvements are at the screening check level.

Very true, I did not mean to imply that all of the new stuff is bad. Air marshals and solid cockpit doors are both excellent ideas.

We could lose an entire airliner once a month to terrorism, with everyone on board dying, and air travel would still be 100 times safer than driving a car.

No way. That would be something like 3,000 deaths per year, versus 30,000-40,000 per year for cars. Airliners do about 1/7th the passenger miles that cars do, so the fatality rate would be pretty close. It would still be a bit safer than driving, but not much. Certainly not 100 times safer.

You have to look at total passenger-miles. Currently planes are over 1000 times safer. I think it actually might approach 10,000.

I did: http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pub...

For 2014, the latest year listed, airlines did 607,772 million passenger miles, highway vehicles did 4,371,706 million.

The current level of safety can't be accurately measured, because airline fatalities don't happen often enough. Whatever the level is, it's far better than cars. But the hypothetical one-full-plane-crash-per-month scenario does not compare particularly favorably to cars.

Maybe the figure was 10 times safer then. A plane carries an average of 200-300 people. One flight per month is somewhere between 2400-3600. Current highway fatalities number about 35,000. So the ten times figure works.

That said, I feel like either passenger-miles figure is either too low for airlines or too high for vehicles. I remember the study I had read had them much different.

Ten times fewer fatalities and seven times fewer passenger miles means safety is nearly the same. Not ten times better.

Check out these charts:

Injured persons by transportation mode: http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pub...

Total Accidents by mode: http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pub...

Rates for Air fatalities: http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pub...

Rate for Vehicle fatalities. http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/pub...

The two key numbers here I think are: Fatality Rate per 100 million vehicle-miles: 1.10 (2013) and Fatality Rate per 100 million aircraft-miles: 0.117 (2013)

The 10 year averages for both of those respectively are 1.254 and 0.189.

What's your point? These numbers agree with me.

If you're still arguing that airlines would still be 100x (or 10x) safer in the face of monthly fatal crashes, these numbers do not support that at all.

If you're merely arguing that airlines as they actually are today are 10x safer than driving, that is clearly true and a vast understatement, but I never said otherwise.

There are essentially 0 non-fatal injuries in major commercial air travel. Add in the (hundreds of thousands?) of auto-related injuries and even in the scenario described, air travel is easily 100x safer.

They said that if a full airliner were lost each month it would still be 100x safer. That is completely untrue.

In no way directly related to this post, but I'm always looking for an excuse to link to my two favorite airline stories.

1) The Turn - http://www.theatlantic.com/past/unbound/langew/turn.htm

2) The Crash of United Flight 232 talk by Capt. Al Haynes - http://clear-prop.org/aviation/haynes.html

Exploiting this aviation-related post to see if any aerospace folks will comment on my two-month old Ask - Ask HN: Why is the National Airspace System so dated? https://news.ycombinator.com/item?id=12651244

In regard to the safety of regionals vs. the major airlines that the article's author is writing about: I can see why the statistics are separate. Even though, according to the FAA, we're all under the "same standard", in practice, this isn't the case. As long as the paperwork looks the same, though, we're "safe". Also, within the industry, some pilots tend to think of themselves as shouldering the responsibility of safety, however, I sometimes think we're safe despite the pilots. Engineers of aircraft, airspace, arrival/departure/approach procedures, airports, ATC systems, etc. are silently keeping you safe. That the FAA is able to manage all of these elements is quite impressive. Still, its massive bureaucracy can be quite Vogon-like.

I think a lot of software engineers would be blown away at how strict software standards are for avionic systems. Like 100% modified condition/decision coverage for all software tests, strict. No pointer math and no dynamic memory allocation, strict.

https://en.wikipedia.org/wiki/Modified_condition/decision_co...

Being a party to lackluster software development on a regular basis, I somewhat pine for projects this strict.

As part of my daily routine, feigning surprise at the fact that, given a tree of decisions, I'm only permitted to code for the ones that make money, simply to find that "Quality Testing" reveals the 500 other permutations will inevitably result in thrown exceptions with hideous, glaring error messages.

"WHOOPS."

This sort of thing is exactly what some of the software development processes are all about.

When I was in a CMMI training class 10 or more years ago, the instructor we had said basically that much. By instantiating a process, it takes the mental weight of knowing (or checking) if everything is correct.

No one builds cars that way - why should we build anything that way. The processes should catch the inevitable human errors.

Does anyone know what this is referring to?

> We’ve engineered away what used to be the most common causes of accident

I'm not enough of an experienced pilot to authoritatively answer, but here's a guess at some items.

* Sterile cockpit: a rule that pilots avoid any distracting activities during critical parts of flight (mostly takeoff/landing, flying around busy areas).

* Crew resource management: a shocking number of major commercial accidents have involved either confusion in the cockpit or one pilot who knew something was wrong but didn't feel he/she had the authority to say something. Tenerife is a very sad example. CRM is a set of procedures that specify how concerns are brought to the PIC (pilot in command) in a way that's unlikely to make things worse. Basically, it's an escalation procedure that is recognized on both speaking/listening sides as something that needs to be paid attention to.

* Checklists: Not sure if this is new, but it is completely routine in flight to consult standardized checklists for everything, whether ordinary takeoff/flight/landing or emergency situations. It's thus a lot less likely you'll make a stupid mistake in flight like running out of fuel because you forgot to switch tanks. (Actually people still run out often, but the out-of-fuel checklist includes ensuring you've remembered to switch away from the empty tank, meaning people don't crash planes with half their fuel anymore.) There is no cultural shame in consulting checklists; in fact my (limited) experience is that most pilots would question the judgment of someone who didn't use checklists. This is a bigger deal than you might think because most aircraft incidents are fundamentally attributable to human error.

* Better weather technology, both for prediction and real-time assessment.

Flight 1549 (the one that landed in the water with no deaths and just a couple serious injuries) is a good example of the first three. You can listen to the cockpit recording to see them all in action.

When the birds hit, the pilots were already 100% focused on the critical takeoff portion of the flight -- zero distractions that could have delayed reaction time.

The copilot immediately got out the POH (plane's owner's manual containing tons of checklists) and started all the restart procedures while the captain reflexively put the plane in the maximum glide position to extend range.

Then right as they're lined up for the river, Sullenberger asks the copilot "Got any ideas?" While this might sound like comical desperation, it's actually good CRM in action -- after they'd tried the checklists and were in a somewhat stable descent, the captain solicited feedback from the second-in-command, just to make sure they hadn't overlooked anything else.

Anyway, I'm sure my list is far from complete. Someone more knowledgeable can continue it.

Checklists in aviation started in 1939, I think. Airplanes had gotten complicated by then and pilots would routinely forget to set things up properly for takeoff, etc.

Checklists have been incredibly effective in improving safety. It's a crying shame that they are not used elsewhere, like for surgical procedures. Doctors need to give up on the idea that checklists insult their competence.

Also, the health care industry could learn a great deal from the aviation industry about eliminating human error. The health care industry just wings it as far as I know, and hospital deaths due to stupid mistakes are legion.

There's a movement towards checklists in healthcare, and the UK NHS is picking them up, as I understand.

I'm currently reading The Checklist Manifesto by Atul Gawande, a surgeon. Devops shouldn't be life-or-death, but I'd still like to learn and improve.

https://www.tailstrike.com/150109.htm has the 1549 transcript incase anybody's looking.

https://www.youtube.com/watch?v=5S5hRRio-E8 has the radio audio with the CVR transcript overlaid as text.

> Some will argue that it’s unfair to gerrymander statistics in this fashion, separating airlines into categories of convenience. However, it’s necessary. Regional carriers, for instance, can have substantially different cultural and operational environments than their legacy carrier affiliates. It’s the majors that set the standard, and it’s only fair that we measure from there.

Yes. It is gerrymandering, and no you can't wash your hands of the regional carriers' crashes and deaths after you've stamped your 'trusted' name and brand on their flights.

I am appalled at how they callously dismissed those plane crashes since 2001 like this.

> To be clear, there have been a handful of tragedies involving regional carriers and freighters... in which dozens died.

Really? 'dozens died' is ok, but 100s died is not? Are we cattle?

> Really? 'dozens died' is ok, but 100s died is not? Are we cattle?

Yes. Some people are choosing to pay less for a little more risk. The market should allow for that in the fringes. The vast majority of flying happens on the major carriers, so having stricter standards for the majority of your flying passengers is the right way to approach safety.

Also: Air Crash Investigation / Mayday / Air Disasters is an interesting series if one is interested in real-world, root-cause engineering analysis.

RIP to the freight dogs though:

"2013: A UPS flight crashed on approach to Birmingham, Alabama, killing the two pilots."

Why the downvote? UPS is a major airline - even if they only transport cargo, humans still had to give their live.

Also, "freight dogs" is the gentle term for the pilots in cargo planes.

This would be why I always tell people to fly major U.S. legacy carriers if possible, or their western equivalents (e.g. British Airways, Qantas). As a general rule, major carriers from countries that have a pool of predominantly native English speakers to hire from.

Actually, only two of the world's safest airlines hail from the US: http://www.telegraph.co.uk/travel/news/The-worlds-safest-air...

And 6 of the top 10 are not from English speaking countries (ANA, Cathay, Emirates, Etihad, EVA, Finnair).

Actually, only two of the world's safest airlines hail from the US

The list you just cited includes Alaska and Hawaiian on that list, for a total of four US carriers. My original statement was in context of major legacy carriers, so I could understand you excluding those—however, each of the airlines you cited has a comparable—if not smaller—fleet size to that of Alaska Airlines.

Also, keep in mind that major US carriers have underwent major consolidation in recent years:

US Airways and American Airlines

United and Continental

Northwest and Delta

>And 6 of the top 10 are not from English speaking countries (ANA, Cathay, Emirates, Etihad, EVA, Finnair).

If you total the fleet sizes of each of these, you end up with 907 aircraft. American Airlines alone operates 952 aircraft.

Moreover, many of the airlines you listed tend to hire Western pilots formerly of major legacy carriers—captains in particular—in order to meet high PIC experience requirements.

It's also worth mentioning that AirlineRatings.com[0] (the source of the list you cited) currently assigns Malaysia Airlines a safety rating of 5/7, despite the fact they managed to lose two 777 hulls (MH370, MH17) only months apart—one of which was perfectly preventable (don't fly over active war zones?).

The same site affords Asiana Airlines a similarly lofty rating of 6/7, despite a 777 hull loss (Asiana 214) due to its pilots being unable to perform a simple visual approach in ideal conditions at SFO. A truly stunning display of incompetence that belied a terrible training and safety culture at the airline.

Of course, going by safety records alone, I'm sure both airlines looked great on paper about five years ago.

[0] http://www.airlineratings.com/safety_rating_per_airline.php

One word. DO-178B.

Reply on news.ycombinator.com