Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

USS McCain collision ultimately caused by UI confusion

arstechnica.co.uk

Post reply on HN
Print view View original

81–90 of 369 posts

Re: USS McCain collision ultimately caused by UI confusion

#81

Post by manicdee »

Animats wrote: post #73

There was a ferry accident involving a similar transfer of control problem in New York harbor.[1] Ferries have a lot of maneuvering modes and directional thrusters, because they do so many dockings, and so they have a more complex control problem. The pilot put the system into a backup dumb mode while in cruise, then changed stations to where he could best see the dock while still in the wrong mode. Rammed the dock a…

Split command also provides plausible deniability.

“No, we did not deliberately ram a ship suspected of smuggling contraband to North Korea, that would be illegal.”

Re: USS McCain collision ultimately caused by UI confusion

#82

Post by azernik »

mschuster91 wrote: post #58
tlb wrote: post #3

Amazing that someone would design in multiple steering wheels, only one of which is active, and no big indicator to make it super-obvious which one it is. Also amazing that there's a mode where moving a control sometimes controls all engines, sometimes just one.

> Also amazing that there's a mode where moving a control sometimes controls all engines, sometimes just one. This is a fallback in case the primary rudder fails. You can still shift left or right, but you will take more time and it will be far less precise than using the rudder. This is also valid for aircraft, where not only the rudder can be compensated but also the elevator (aka up/down) by simultaneously modifyi…

The UI issue being that ganged control should have been the default, but separate control was instead.

Re: USS McCain collision ultimately caused by UI confusion

#83

Post by cocoablazing »

A key occurrence during this mishap:

0127 The Officer of the Deck ordered course to the right to course 240T, but rescinded the order within a minute. Instead, the Officer of the Deck ordered an increase to full speed and a rapid turn to the left (port). These orders were not carried out.

0129 The Bosun Mate of the Watch, a more senior supervisor on the bridge, took over the helm and executed the orders.

I've witnessed multiple similar situations on a warship (imminent loss of extremely critical systems), and had to forcibly push the watchstanders out of the way to do their job for them. Many people are going to freeze when the SHTF, and there may be no prior indication.

Major attaboy to the BMOW who saved many of his shipmates. Given the incident angle of collision, many more would have died from a perpendicular collision.

Re: USS McCain collision ultimately caused by UI confusion

#84

Post by kylec »

monocasa wrote: post #38
drsopp wrote: post #32

Earlier quoted context omitted.

I’m sorry for bringing this up here, but why do I need to scroll sideways in Safari on iphone on Hacker News every time there is a quote written in an unproportional font? Just, why?

Because he used block quote syntax. It's designed for code on HN's implementation, so it preserves newlines, instead making you scroll. It's a poor choice for quoting non computer languages.

People wouldn't use it if a better option was available. Lots of people are familiar with Markdown, which lets you start a line with > to have it quoted, but HN does nothing with this syntax so it looks ugly and as a consequence people don't do it.

Re: USS McCain collision ultimately caused by UI confusion

#85

Post by Johnny555 »

A billion dollar ship seems like the kind of place it's worth having real physical controls and indicator lights for functions like this.

I get why auto makers love touchscreens, because buttons, knobs and lights are comparatively more expensive than adding another menu entry to the touchscreen, but that's not the case in an expensive ship.

Re: USS McCain collision ultimately caused by UI confusion

#86

Post by swixmix »

Animats wrote: post #73

There was a ferry accident involving a similar transfer of control problem in New York harbor.[1] Ferries have a lot of maneuvering modes and directional thrusters, because they do so many dockings, and so they have a more complex control problem. The pilot put the system into a backup dumb mode while in cruise, then changed stations to where he could best see the dock while still in the wrong mode. Rammed the dock a…

Do you have references for the Submarine story? The submarines I've seen have a rudder, stern plane, and another on the sail. With mechanical controls, I imagine it's difficult to have one person control all three of those movements. Two people controlling three movements seems right.

Re: USS McCain collision ultimately caused by UI confusion

#87

Post by mikerathbun »

I would be interested to see how much liberty the contractors who develop these systems are given. Are they given a rigid spec by the Navy and expected to implement it exactly or are they given a general proposal which allows them to iterate on UI from lessons learned from previous accidents caused by confusion.

Re: USS McCain collision ultimately caused by UI confusion

#88

Post by noobiemcfoob »

Zak wrote: post #21

Aside from the lack of clarity about which station has control over what, which should be made really obvious, with additional warnings when someone sets a control input on an inactive station, this seems bizarre to me: While the Ship's Control Console has a wheel for manual steering, both steering and throttle can be controlled with trackballs, with the adjustments showing up on the screens for each station. I've ne…

You use trackballs to input relative movement commands. The position of the ball doesn't need to relate to the position of the cursor. It's like a mouse if you picked it up and replaced it after every stroke. They have a couple nice properties in that they can be fixed in place and offer more freedom of movement to a joystick.

Re: USS McCain collision ultimately caused by UI confusion

#89

Post by caconym_ »

foldr wrote: post #65
mikeash wrote: post #55

Earlier quoted context omitted.

I believe this is the official transcript: https://www.bea.aero/uploads/tx_elyextendttnews/annexe.01_04... I see that the first link I posted doesn't contain everything, but I don't see any major omissions either. At 2:11:37.5, Robert takes control of the plane. According to the report, he pushed the stick to the left twice. Bonin then took back control. Figure 69 on page 96 of the report you linked shows Bonin's pit…

Figure 69 shows Bonin's stick inputs before Robert took control (and a few seconds after). As explained on p. 182 of the report, Robert pressed the override switch at this point and took control, so Bonin's stick inputs would have been ignored. Bonin then took back control by himself pressing the override switch. In effect, then, at this late stage, the pilots were fighting over the stick. If it had been a physical f…

With physically linked controls, direct physical feedback will tell you if the person in the other seat is fighting your inputs. Without them, that doesn't happen, so any explanation for the aircraft not responding to your inputs is potentially valid.

Re: USS McCain collision ultimately caused by UI confusion

#90

Post by ineedasername »

Nope; No, just not at all:

UI Confusion is never "ultimately" the cause for accidentally killing people through poor operation of a $2 billion piece of military equipment. A UI may suck, many do, in or out of the military. So, training. and then more training.

I don't have a high stakes job, but if I allowed code-completion to drop tables, commit the change, delete the backups, and then burn down the building... well, that's not on the UI, that's on me.

So unless we're talking MS Word Clippy-- and I think the article would have mentioned that-- UI can't take the fall on this.

Besides: It began with the Helmsman sending all control instead of just throttle control to the Lee Helm station. That mistake isn't explained as a UI issue, it's glossed over, but everything came from that initial screw-up. Sure, afterwards there was confusion, but even then, "OMG WE LOST ALL STEERING INCREMENT THE IDIOTIC PANIC OPTION VARIABLES!" isn't really an acceptable response. How about, calmly, "Helmsman double check that control assignment if you please "

Post reply on HN