I noticed that users who tried out the leap motion controller do not have an intuitive sense of how large the interaction box is.
This produces some frustrating user experience when performing scrolling, panning, rotation actions. In these kinds of actions, users must signal the end of each action in order to repeat the same action. For example, scrolling action typically requires the user to repeat their motion in specific direction. It requires the arms to return to the starting position without scrolling being activated - which is achieved by moving outside the interaction box so that such motion is not registered as trying to scroll back.
Without knowing where the boundary of the interaction box is, the users have to guess the boundary by trial and error.
Another frustrating scenario is that users have been found to be instinctively hovering their hands right above the sensor (< 10cm distance), thinking that the closer their hand is to the sensor, the more likely they will get some kind of reaction.
One idea I have is to create some kind of 3D holographic display delineating the boundaries so that it is clear and obvious to the user when their hands are "in" and when their hands are "out".
Have anyone thought of any ways of letting the users know the physical boundary of the interaction box?