[gervase]

I think motion sensing is probably the way forward; voice control is intrinsically limiting in most (crowded) environments where computers are used today.

Considering humans can ride bikes, drive cars, play instruments, etc. (including typing on keyboards!), I think that indicates that non-verbal, physical interaction is not nearly saturated as a transmission channel.

Conversely, it's hard to imagine someone verbalizing "navigate to HN" in a loud open-space office, or "Excel, create a pivot table" or whatever. I think it's fine in private spaces like your home, but in public spaces, you're implicitly broadcasting your activity to everyone around you, which I consider to be a strong negative.

[BatFastard]

The Deamon book has some cool motion interaction in it. They call it the Shamanic interface! I really want to make one of those, too bad tech is not readily accessible yet.

Is subvocalization a possibility? Mic or EEG setups might need to be slightly different.