Anything that takes attention away from driving increases danger.
Are they more dangerous than older interfaces? My feeling is overwhelmingly yes, but I would be willing to see a study or hear arguments that some touchscreens are an improvement. A touch interface is fine (not great) as long as it never changes. As soon as you have to search for a control or menu you are dividing your attention away from driving.
Specifically, text. Reading is "hard". Even things as simple as the title of the song on the radio. Especially when the text changes.
I have a modern LCD on my motorcycle, a BMW, that uses a WonderWheel (rotate to scroll up/down, and push or pull for right/left click) as an interface. It's very reminiscent of The Onions MacBook Wheel[0]. It is absolutely dangerous to use while riding. It's a cognitive black hole.
Obviously, the LCD is not alone in this case, the interaction pushes it all up to eleven. But the old school car interface was numbers and small words, and, eventually icons. Consider changing the temperature in a car, for me, I'd shove the hot/cold slider around until the air coming the from the vent was comfortable vs clicking up and down and deciding "do I want 72 or 73?".
And, yea, maybe it's just me. Perhaps I alone am a hazard when interacting with these things. So, maybe it's not fair for me to project my experiences to the population at large.
However, it is very person dependent. Personally, I am one of the fastest readers I know.
It's also day dependent. I've had days where my ability to focus switch is significantly impaired.
The big issue is that while there are people that touch screens are not going to impair their driving, you can't gear your system to them.
You have to aim it at the lowest common denominator.
Personally, I am a fan of my current vehicle which while being at 2015 because it's one of the police interceptors still has the basic ish radio. And has twist knobs for volume, tune, fan speed and temperature.
And while I probably wouldn't mind having the actual Ford sync stuff, I don't find myself missing it either.
- Have you benchmarked your speed on text vs non-text controls that are otherwise equivalent? (i.e. both are button presses, both are always in the exact same location, ...)? - Have you benchmarked how this changes as you loose the similarities? Does this benchmark measure "time to complete task" or "time spent looking at control" (turning a physical knob vs a screen slider) - have you benchmarked your speed for fixed-location controls vs controls which may be buried in a menu item on a touch-screen?
Do these benchmarks change if the control has delayed onset (pressing "play" takes 2 seconds to start the music, and you get no tactile response to tell you if you have successful pressed the button or not)
Have you benchmarked how these skill comparisons decay with impairment? Do they decay equally, or does the text-based skill decay faster?
Look, given this is HN I fully believe you are in the upper 99% on several aspects, making you with text controls faster than me with manual. But the question is would YOU be faster with text or manual? And how consistent is this?