[kube-system]

I don't think it is insanity. It's part of the complexity inherent to managing state with auto high beams. My Toyota worked differently but had other complexities. I just looked up a 2021 Legacy manual and it sounds like Subaru High Beam Assist works about the same way as my Toyota did.

The annoying thing about that design is, if I want to turn my high beams on manually, and the car decides that a street light in the distance is a vehicle, it will ignore my request to turn on my high beams when I push the stick forward. I then also have to set my primary headlights to manual mode in order to override the car. This requires turning the knob at the end of the stalk which is not backlit and therefore not legible while driving at night. This is one area where I like the Honda design better: I can always turn my high beams with one operation.

I think the page count is relevant because it is indicative of complexity and is likely inversely related to the number of people who read it.

While I am sure you probably do understand the controls on your vehicle -- I think it is important to note that the people driving around with lights in an improper state at night are also under the impression that they understand their vehicle's controls. They're just mistaken.

I personally wish licensing and enforcement was more stringent. But this is not politically possible in the US.

[nawgz]

> The annoying thing about that design is, if I want to turn my high beams on manually, and the car decides that a street light in the distance is a vehicle, it will ignore my request to turn on my high beams when I push the stick forward. I then also have to set my primary headlights to manual mode in order to override the car. This requires turning the knob at the end of the stalk which is not backlit and therefore not legible while driving at night. This is one area where I like the Honda design better: I can always turn my high beams with one operation.

It's an Outback for what it's worth, but... High beams work directly by pushing the stick forwards. I haven't experienced it denying such a request, only correctly toggling off when traffic is both oncoming and within a reasonable distance, but if it did behave incorrectly in this way, I would simply pull the stick back and push it forwards to re-activate high beams. I don't think there is any complexity in managing state with this implementation, it seems inherently obvious to me it would be this way. I am very surprised with everything else you've written here because it seems quite extraneous.

[kube-system]

> if it did behave incorrectly in this way, I would simply pull the stick back and push it forwards to re-activate high beams

In the Toyota that I had, if it thought there was an oncoming car, asking it to do the same thing again wouldn't result in a different outcome. Maybe Subaru's system is tuned better, or maybe my driving conditions are an edge case.

Either way, this was just an anecdotal tangent to support my above point that vehicle controls are more complex than they used to be. What was once one single boolean value is now a logic tree containing four boolean values ([auto headlight system engaged][vehicle detected ahead][vehicle speed within operational parameters][highbeams commanded by driver]).

[nawgz]

> In the Toyota that I had, if it thought there was an oncoming car, asking it to do the same thing again wouldn't result in a different outcome

Why wouldn't the result of the action the second time be based on exclusively the stick's position and the current state the vehicle is in? Then it is a pure function of time. If they truly mixed up previous states in there, and somehow it's aware that in the history of positions it was at this position and at that time it rejected your request so it will also reject it now, that's braindead and you should phone Toyota and scream at them

My car's headlight's current state is a pure function of time T and position P. That's my mental model and I haven't observed deviations. I will look at these manuals later perhaps to understand if they really argue there are more stateful factors than this