> maintaining a separate, hidden outline of DOM elements mirroring what’s on screen. Is there any analysis on the efficacy of this approach? This is always my

> maintaining a separate, hidden outline of DOM elements mirroring what’s on screen. Is there any analysis on the efficacy of this approach? This is always my first thought in these discussions. Performance of WASM with native text IO. Being fully "hidden" one could just use semantic elements, skipping most layout div's, maybe in turn reaping untold accessibility benefits? Being such a simple idea I cannot believe that it wasn't tried and abandoned before.