Mostly segmentation of your code. Sure, you can have a single view with a child RealityView, but that also means that this will always be your "master immersive space". It makes the usage of the immersive spaces be very limited, and forces you to handle any code surrounding your RealityView (including attachments) and other outside actions (like onAppear) in a very clunky fashion.

This is also what I have for plan B, but it surely would be nice to be able to define transitions (animation, transition styles, etc) between spaces, that could play while loading.

This will still produce a gap between immersive spaces, unfortunately.