This is the issue with imposing semantics of the programming model on the behaviour.
User behaviour is the only _real_ thing, it happens. Everything else is in your head. If people in the real world use PiP, then it should happen. The programming model has to bend and change to support it. It simply does not matter if the window manager does something or the window does something.
Sure, there is always the security argument wayland folks fall back to. But what ever is the problem with making a one-time permission popup? "Google Chrome wants to open in PiP: allow | allow once.". Just expose the existing PiP code in the window manager as an API guarded with an `if` that apps can call. It's not even that much real work, just pure bikeshedding and architecture astronauting.
Right right, and I'm not saying users shouldn't be able to have a floating window with video (or whatever) in it. I'm saying it shouldn't be Chrome making that window floating and always visible.
Permission prompts still only allow things that have already been thought of so we will see less innovation in the future. I don't think this kind of security model is needed at all for an open source desktop where we can enforce directly that programs respect the user instead.