I haven’t had time to fully digest this yet, but I have a feeling it may have implications for us.
There’s an interesting counter-point to the idea of the shadow dom here:
I don’t really see any implications. Users will be able to render a DOM scene into any element, including a shadow DOM element, and our engine won’t care. Our engine will merely apply transform:matrix3d to the elements in a scene, regardess of where that scene is.