I would also like to see a more radical approach. E.g. one making use of graphic features. Why is my auto-completion completely text-based instead of being full

I would also like to see a more radical approach. E.g. one making use of graphic features. Why is my auto-completion completely text-based instead of being fully graphical like in a real browser? It's a waste of screen space having fixed row size. Maybe evolution will also go the other way around: Nautilus and other file browsers have rudimentary shell-like use & feel, e.g. when searching for something. Extending that could create some interesting features.