There is a baked in assumption that there is a single "user" and that this user is the one that has their hands tracked. What about the use cases of third person hand tracking?
"Is there a way of knowing which hand(s) belong to the user, vs. hands detected by near-standing neighbors, or hands detected at a distance attached to other people in the viewport? (I assume there are heuristics that try to limit the hand-tracking only to the user's hands.)"
A device is free to report whatever hands it sees as additional input sources, though I'd expect input source 0 and 1 to be the users' hands and further hands to slot in after that.
There is a baked in assumption that there is a single "user" and that this user is the one that has their hands tracked. What about the use cases of third person hand tracking?
"Is there a way of knowing which hand(s) belong to the user, vs. hands detected by near-standing neighbors, or hands detected at a distance attached to other people in the viewport? (I assume there are heuristics that try to limit the hand-tracking only to the user's hands.)"