4 ms·

The big difference between cars and smartphones is that you can focus all of your attention on your smartphone screen to use it. That's not true for the driver

The big difference between cars and smartphones is that you can focus all of your attention on your smartphone screen to use it. That's not true for the driver of an automobile. The best automobile interfaces use knobs of different sizes and with different textures so that you can find them without looking. Buttons should be arranged in rows, so they are easy to find with a short look. Extremely common buttons should be located on the steering wheel. Etc. Until screens can offer some kind of dynamic tactile experience, where you can feel what you're about to control, they can't be better than real tactile buttons and knobs. Don't get me wrong, I love a big screen in my car for advanced features (360 camera, vehicle settings, navigation, music, etc). But the basics need to be tactile.

sho_hn 5y ago
That's kind of what I mean: The discussion always starts and too often ends there. A couple of thoughts: - Already in the pre-touchscreen, button-dominated area, plenty of more complicated use cases had terrible UX, and in many cases touchscreens have made them much easier. For example, do you remember how difficult it often was to program radio presets in 80s head-units? There were a lot of real clunkers there. - There's plenty of newer use cases that would be hard or impossible to make easy with traditional hardware button UX, such as route and EV charge planning. - Good software offers many possibilities to avoid interactions altogether, or make them single-step instead of multi-step. A good screen-based system will now often suggest interactions pro-actively and situationally at the top-level of the system and reduce interaction to a single multi-modal (touch/gesture/speech) confirmation on what used to be multi-step processes. - Screens can offer tactile experiences - most recently with the introduction of screens that can do permanent haptic feedback via electrostatic friction coefficient modulation, emulating different surface feels as the finger moves. - You already mentioned features such as 360 degree cameras: If nothing else, screens are great for showing information. Interactions aside, this alone has a lot of legs left and there's absolutely quality differences between good and bad implementations. - There's absolutely nothing wrong with a good physical button (on the contrary - few things are more satisfying than a good, solid, weighty button) and having both for use in the appropriate situation. It doesn't mean there's no possible benefits from further investment into good car software. I feel a lot of this is rooted in first-hand experience with bad systems - which I agree are plentiful, and just as awful as a bad, clunky, unrefined phone onscreen keyboard. Dare to imagine a good one. Working on car systems means getting to own the form factor and the entire surroundings of a user, much more than working on software targeting rarely-changing phone and PC devices. It's an open field for software and hardware innovation and deserves a little more professional excitement and ambition to do it well.
- bonestamp2 5y ago
  I agree with some of that. The new HUMMER EV has some neat blended tactile/screen interactions where the button labels are on-screen above the buttons. That in itself is not new, but the way that some features work is really slick, such as sub-menu options replacing the functionality of adjacent buttons when appropriate. I wouldn't rule out that tactile HVAC and Audio controls are always better than a touchscreen, but I have trouble seeing the benefit for some of those basic controls. I mean, let's take the volume control for example. It's a very simple linear concept, there are no branches in the options for volume. There are no sub-menus -- you either want it higher or lower. You always want that control in the same spot so it's easy to find and you always want it available so you can use it at any time. For those reasons, I have trouble seeing what a touchscreen can add to that particular parameter. A knob seems better suited by nearly every metric. Touchscreens on the other hand, are great for controlling more complex parameters. Even a temperature control, while similar to volume, is actually more complex because the change is not immediate. You're setting a parameter that will be true in the future, and that is something where a dynamic display can be beneficial. That said, you still generally want that control to be in the same spot and always available so it's not an obvious win for touchscreens.
bahmboo 5y ago
I think that’s the point and also what sho_hn is saying. There’s nothing inherently wrong with screens and touch controls it’s that the car companies have done a truly terrible job in implementation. We have lots of tech including haptics and other feedback to make these interfaces work far better for humans that isn’t being taken advantage of. On the other hand a great counterpoint is the story of the Mac Touch Bar. (I’m a software person who also works on cars)

sho_hn 5y ago

That's kind of what I mean: The discussion always starts and too often ends there. A couple of thoughts: - Already in the pre-touchscreen, button-dominated area, plenty of more complicated use cases had terrible UX, and in many cases touchscreens have made them much easier. For example, do you remember how difficult it often was to program radio presets in 80s head-units? There were a lot of real clunkers there. - There's plenty of newer use cases that would be hard or impossible to make easy with traditional hardware button UX, such as route and EV charge planning. - Good software offers many possibilities to avoid interactions altogether, or make them single-step instead of multi-step. A good screen-based system will now often suggest interactions pro-actively and situationally at the top-level of the system and reduce interaction to a single multi-modal (touch/gesture/speech) confirmation on what used to be multi-step processes. - Screens can offer tactile experiences - most recently with the introduction of screens that can do permanent haptic feedback via electrostatic friction coefficient modulation, emulating different surface feels as the finger moves. - You already mentioned features such as 360 degree cameras: If nothing else, screens are great for showing information. Interactions aside, this alone has a lot of legs left and there's absolutely quality differences between good and bad implementations. - There's absolutely nothing wrong with a good physical button (on the contrary - few things are more satisfying than a good, solid, weighty button) and having both for use in the appropriate situation. It doesn't mean there's no possible benefits from further investment into good car software. I feel a lot of this is rooted in first-hand experience with bad systems - which I agree are plentiful, and just as awful as a bad, clunky, unrefined phone onscreen keyboard. Dare to imagine a good one. Working on car systems means getting to own the form factor and the entire surroundings of a user, much more than working on software targeting rarely-changing phone and PC devices. It's an open field for software and hardware innovation and deserves a little more professional excitement and ambition to do it well.

bonestamp2 5y ago
I agree with some of that. The new HUMMER EV has some neat blended tactile/screen interactions where the button labels are on-screen above the buttons. That in itself is not new, but the way that some features work is really slick, such as sub-menu options replacing the functionality of adjacent buttons when appropriate. I wouldn't rule out that tactile HVAC and Audio controls are always better than a touchscreen, but I have trouble seeing the benefit for some of those basic controls. I mean, let's take the volume control for example. It's a very simple linear concept, there are no branches in the options for volume. There are no sub-menus -- you either want it higher or lower. You always want that control in the same spot so it's easy to find and you always want it available so you can use it at any time. For those reasons, I have trouble seeing what a touchscreen can add to that particular parameter. A knob seems better suited by nearly every metric. Touchscreens on the other hand, are great for controlling more complex parameters. Even a temperature control, while similar to volume, is actually more complex because the change is not immediate. You're setting a parameter that will be true in the future, and that is something where a dynamic display can be beneficial. That said, you still generally want that control to be in the same spot and always available so it's not an obvious win for touchscreens.

bahmboo 5y ago

I think that’s the point and also what sho_hn is saying. There’s nothing inherently wrong with screens and touch controls it’s that the car companies have done a truly terrible job in implementation. We have lots of tech including haptics and other feedback to make these interfaces work far better for humans that isn’t being taken advantage of. On the other hand a great counterpoint is the story of the Mac Touch Bar. (I’m a software person who also works on cars)