Earlier quoted context omitted.
The "material-based renderer" sounds... fluff? Most any system should have it so that rendering a menu is a single line of code. Same for toolbars and other. Those are typically simple registrations of ("name" "description" handler). Anything more that that, and you are doing tailored menu/toolbar and are straying on purpose. That leads, then, to the spatial windows and APIs that go with them. And... quite frankly, I…
The gaze tracking is process isolated. Only when you tap is that location ever communicated to the application process — so they had to build a UI API that supports configurable hover behaviour where that hover behaviour is never readable by developers They've thought through the accessibility features, how VoiceOver works with the headset, dynamic type sizes, etc. Designing this sort of consistent API across a whole…
Then there is the juxtaposition of, "this let's you present as if there" with the, "of course you have to fly places to be with people, so when you do, zone out like a champ"
Of course, zoning out better not need you to interact much, as hearing on a plane is tough, and pinching your neighbor is not smiled upon.
I will mostly have to take your word for it that hover actions are not observable by the application. Will make games... More amusing than usual, with no memory of being looked at for characters.
And not showing any extra peripherals is a big part of "dreams work." Physical feedback is huge, and a big part of why controllers are needed. Even just the vibration of the standard PS5 controllers go a long way. For driving games, a haptic wheel is more immersive than the vr.