[jerf]

It seems to me it might be easier to put localizer gloves on the user instead. They wouldn't even have to be gloves in the traditional hand-covering sense, just enough to stay on the hand and provide orientation, just little skeleton things.

It seems like the real insight is the creation of a gesture system with spatial persistence by using one hand as a reference point shared by both the computer and the human, rather than the lapel-camera. The way in which the innovative gesturing system gets into the computer seems less important than the gesturing insight in the first place; this is the first gesture system that I've seen that seems like it might actually be useful. And now that someone's had the base idea, a couple of iterations of refinements and we might actually be somewhere.

(It seems like the gestures in Minority Report represented what people thought this would look like, but those always seemed almost impossibly floaty; one metric I use to determine how well a computer could possibly perform is "could a motivated human even figure out exactly what you meant?" and I'm not convinced the Minority Report interface meets that; pointing is a pretty imprecise thing, more than you might realize because the average person's 3D model of space is shockingly bad in some ways. Anyone who has learned to draw late enough in life to be able to observe the process of learning to draw, as I am fiddling with now, quickly learns that.)