Analyzing the kinematics of bivariate pointing.

Despite the importance of pointing-device movement to efficiency in interfaces, little is known on how target shape impacts speed, acceleration, and other kinematic properties of motion. In this paper, we examine which kinematic characteristics of motion are impacted by amplitude and directional target constraints in Fitts-style pointing tasks. Our results show that instantaneous speed, acceleration, and jerk are most affected by target constraint. Results also show that the effects of target constraint are concentrated in the first 70% of movement distance. We demonstrate that we can discriminate between the two classes of target constraint using Machine Learning with accuracy greater than chance. Finally, we highlight future work in designing techniques that make use of target constraint to improve pointing efficiency in computer interfaces.