Touchscreen is such a stupid low value move for driver. Physical knobs are safer in many ways, last much much more, I don't need to lose contact with whats happening in front of the car to manage these, just muscle memory. BMW has nice big physical knob (turn&press) to control all menus, much better than hunting on flat glossy screen with fingers.

Some stuff can be managed from steering wheel, but instead of putting there things like AC there are voice controls, while we all know very well this increases danger on the roads and calls shouldn't be made while driving.

That being said, I choose (used) cars based on many metrics, and this is just one of them and not the most important one. So its sort of self-inflicted move to worse.

> Physical knobs are safer in many ways, last much much more, I don't need to lose contact with whats happening in front of the car to manage these, just muscle memory.

None of which some up on a spreadsheet in the manufacturer's accounting department.

Similarly: open concept offices with hot desking lowers the amount of rent (per employee) that shows up on an expense report, but the productivity drop of open offices does not.

I used to agree, but the point I made in GP is that the touchscreen front and center could be a requirement, with what to do with one being open to local market interpretation.

IF that's the case, and that's one big if, debating touchscreen vs buttons might not be meaningful, because touchscreen MIGHT be a requirement that cannot be removed. With un-eliminable touchscreen, buttons becomes just added cost, rather than being a resilt of cost calculations between all-touch vs all-buttons.

Button-y interface can be handled on basically any IO-heavier 8-bit microcontroller with some muxes as all it needs is to handle interrupts and emit event messages over whatever-bus. The solution can even be hard real-time if there is such requirement.

Touchscreen will probably run atop whatever OS powers infotainment system. In theory touch events CAN be handled on the same 8-bit uC, but converting click coordinates to button-press messages requires knowing what is visible on the screen, which either hinders expressiveness of touch interface (fixed islands for controls) or complicates implementation (constant updating of element-coordinate maps). The moment safety-critical control runs along with non-safety software huge validation challenges arise - the performance characteristics of the whole system, including UI, must become at the very least bounded.

Hard button solution can be internally commoditized, whereas soft button solution requires constant babysitting throughout development lifecycle. One is much more expensive than the other, however, once the move is made incremental cost drops dramatically.