The Hidden Cost of Touchscreens

I love that the landing gear levers have little tires on the top. It's perfect tactile feedback.

The flap controls have little airfoils on the top.

The cockpit is full of things like that, all learned the hard way. People are very bad at inventing intuitive interfaces.

For example, aviation has standard phrases for things. Like "takeoff power" is maximum power, as that's what you use for takeoff. One day, a pilot was landing and suddenly had to abort. He yelled "takeoff power" to the copilot, who heard "take off power", and chopped power to the engines at the worst possible moment, and they crashed.

The phrase was changed to "full power".

How could a generation of pilots and aviation experts never notice the ambiguity inherent in "takeoff power"? But they didn't, until there was an accident.

iPhones also let you use the volume buttons to take a picture

> High contrast, color-coded buttons take less time to mentally digest.

I agree with this strongly and recommend using high-contrast features and color wherever it can assist in rapid recognition of important UI elements. I remember being surprised by how much font-lock-mode reduced my bug/typo rate; I started fixing many errors immediately after typing them because the color was obviously wrong.

However... it's important to remember that a lot more color blind people will use your UI than you might suspect (if you even remembered to account for color blindness at all, which is unfortunately common). Color can still be used, but remember to also include other no-color based differences and/or use a tool (or consultant) to verify everything is still high contrast and easily distinguishable.

edit:

> “and when this one becomes tolerable, they’ll change the software on me.”

I had an echocardiogram a few months ago, and the technician had to restart after making a minor mistake setting up the first image. Apparently the test equipment was modular, and the plug-in features would show up on the touchscreen UI in the order the hardware was plugged in. The medical center had two of these devices, and the technician was used to the other machine. The one we were using had several buttons in opposite locations. Obviously this was only a minor problem at the time, but medical equipment is another situation where it's probably a very bad idea to make the UI layout unstable.

Having physical buttons has the advantage of limiting how much the UI can be changed in a software update.

>we never tested it on the road

That is an even bigger issue than the decision to use a touchscreen. Don't these companies want to make good products? I can't imagine releasing something without even attempting to put it through some realistic usage. Given the sorry state of in-car interfaces, I suppose I should not be surprised.