Abstract
Research investigating factors in the design of tilt-based interfaces is presented. An experiment with 16 participants used a tablet and a 2D pointing task to compare position-control and velocity-control using device tilt to manipulate an on-screen cursor. Four selection modes were also evaluated, ranging from instantaneous selection upon hitting a target to a 500-ms time delay prior to selection. Results indicate that position-control was approximately 2× faster than velocity-control, regardless of selection delay. Position-control had higher pointing throughput (3.3 bps vs. 1.2 bps for velocity-control), more precise cursor motion, and was universally preferred by participants.
