Bring it DOM!
(Somebody please help us with better pun's) ;]
Delivering HTML-like experiences in WebXR (without HTML) is quite a jump back in time.
An hybrid approach, would be interesting to combine the best of both worlds.
Let's see what's available:
As you can see, as soon as things get smartphone/standalone stereoscopic, the CSS/DOM-possibilities dissappear.
For stereoscopic experiences, SearXR emulates a (stripped-down version of) DOM in WebXR.
Just take a look at SEARXR, and see how this hybrid mix can work on various devices:
The HTML search-textfield 'lives' inside 3D world (as an overlay), but is not rendered using the Webgl <canvas> element.
This hybrid HTML/WebGL combination allows for the best of both worlds.
In VR mode, the textfield is automatically projected as a canvas-texture, without losing its DOM-reference.
Spatial dragging of searchresults is another area which SearXR is researching.
NOTE: Big respect to all VRML research, which has directly and indirectly paved the way for many DOM-like tools & technology (also AFRAME, react-three-fiber ).