I've been mulling this thought over in my head as well and the best idea I've come up with is to allow separation of keyboard keys and logic "channels".
My main trouble is remembering what key I've bound to the output of which sensor, which leads to crosswiring and chaos.
If I could assign an output to a named variable like "proximity front left" and select that from a dropdown im my logic gate, that would free up a lot of mental RAM. This would not actually change functionality except allowing more channels than keys on a keyboard.
I'm not quite sure how the interaction of classic "keyboard keys" logic and my named "logic channels" should be implemented. Either allow parts to use keys and channels simultaneously, wich would be a "free" OR-gate or allow switching the inputs and outputs from "keyboard mode" to "channel mode".
Accessibility should be an an important as well. Ideally for a non-hardcore-logicker nothing would change, everything can be directly bound to keyboard keys, and simple keyboard logic should still work.
Edit: Apparently the devs already had that idea and i didn't discover it yet.
Like Lurkily described, I'd like to see some kind of logic visualisation, too but probably only in the editor. Maybe a toggleable function that shows inputs and outputs of a hovered part as coloured lines with each colour corresponding to an in- or output.
