Hacker News Reader

Comment by cbhl | Hacker News Reader

The Boeing 787 is now about nine years old. I wonder how much of this is operator proficiency triggering latent bugs (aka, "pilots got too good at pushing buttons quickly").

The bugs in the Therac 25 were only discovered after the machine had been in production for a while (you had to be typing "too quickly" in order to get the software to trigger the bug with the safety interlock).

View on HN

It's not very typical for pilots to push buttons quickly. Especially not the thrustreversers and engine and flight controls.

But it is entirely possible that nobody pulled the reversers while still airborne in 9 years, because it's a very strange thing to do. A bit like opening your car door while still driving 60 mph... Nobody does that in normal operation.

A bit like opening your car door while still driving 60 mph... Nobody does that in normal operation.

You don't open and shut a door when you notice it didn't fully close because you can hear the wind?

A better comparison might be putting the car in reverse at 60 MPH.

You never missed a gear, thinking you were driving a 6 speed rather than a 5 speed?

No, but for as long as I've been driving manual cars they've had gates to prevent that very action. (1995)

Not mine (Honda Jazz). But you physically can't do it, as the reverse gear has no synchro and won't mesh unless stationary.

It can be done, but it's neither easy nor a good idea: https://youtube.com/watch?v=ARJENV0qGdY

You'd certainly struggle to do it by accident.

it should make a loud grinding noise though,

The sleeve of my coat caught on the gear shift while making a left turn and threw it in reverse while on the freeway.

It made a grinding noise and I yanked it back into drive.

Other than the noise nothing bad happened.

Non-suicide doors[1] in cars are very difficult to open at 60 mph. Between air pressure, and poor leverage, you are going to have a hell of a time doing it.

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

Just roll your window down first!

At 60 mph? No. I listen to the wind until I get a safe opportunity to properly close it.

(I might try to pull it more shut while moving. But not open it.)

Good luck even opening it far enough if you're going at 60 ...

In airplanes, a lot of procedures and diagnostics involve not re-trying a failed action, not touching anything before reading a check list, nor flipping a switch twice if it didn't work the first time.

> But it is entirely possible that nobody pulled the reversers while still airborne in 9 years, because it's a very strange thing to do.

That's overstating the case a bit. I'm 100% a layman here but I'd agree that the behavior seems like some sort of bug. But…… some planes (typically not commercial, although the Concord is one notable exception) call for reversers in flight under certain circumstances. Additionally there are definitely circumstances where a pilot might deliberately think reversers before touchdown are a good idea.

I've seen a number of videos on youtube of reversers been opened, if not fully spooled up, off the ground. It's not common but it does happen, especially in high cross-wind or otherwise iffy landings.

Here's the first video I found but there are certainly others https://www.youtube.com/watch?v=-RO66a_nvus

Make me think about could we use a robot to fuzz physical inputs.

Probably not cost effective as opposed to either making the software more resilient or just letting some monkeys/interns loose in the cockpit and pressing everything (With some focused searches obviously)

I don't know, I suspect software resiliency is one of those "gets exponentially more expensive as you approach perfect" situations and a robot could bash inputs at speeds a human couldn't in thousands of orders, it would certainly let you test those weird code paths that blow up when they are never exercised until they are.

It does get exponentially more expensive, which is why you have to pick a certain level of ~perfection~ and then stop.

Also, the robots would have to bash at the speed of a human: The point is to mimick human inputs not necessarily test every code path (Which you could do in software)

When testing resiliency software testing isn't the whole of the thing.

I spent a lot of time at new job (well 18mths now) on making the software side more resilient, we still had a site wide outage of internet access and comms and internal systems because one of the women in the sales office unplugged a socket in the ingress comms cabinet to plug her hair straighteners in.

Coincidentally my request to have physical access control on all the comms cabinets shot right up the list.

The boundary layer between software and hardware itself has to be tested, the real world is a messy place :).

..or not requiring physical hardware to test?

How? Planes won't not have cockpits for decades surely.

The software has some interface to the physical controls.

I wonder if selling tickets to watch would offset that somewhat?

I'd pay quite a lot of money to camp out in the desert for a week watching fuzzing robots try to crash airliners...

Considering the level of fly-by-wire that goes into airframes nowadays, I could certainly see such features added to older airframes through flight-software updates after they are already in service. I have no idea if that's how it's done in the field, though.

Software updates are done. However the manufacture has to recertify the airplane with the FAA first. This is expensive, requiring months of testing and a lot of paperwork. Each plane model getting the update needs to be re-certified, it isn't good enough to say that it works on this other model even if the planes are identical for purposes of that software.

The Ryanair booking website has several bugs exactly like what you describe. If you use it too quickly it won't sell you plane tickets. If you use it more slowly it works fine.

Probably to try and prevent bots

Reply on news.ycombinator.com