Performance evaluation of input devices in trajectory-based tasks: An application of the steering law

Abstract

Choosing input devices for interactive systems that best suit user's needs remains a challenge, especially considering the increasing number of devices available. The choice often has to be made through empirical evaluations. The most frequently used evaluation task hitherto is target acquisition, a task that can be accurately modeled by Fitts' law. However, today's use of computer input devices has gone beyond target acquisition alone. In particular, we often need to perform trajectory-based tasks, such as drawing, writing, and navigation. This paper illustrates how 'a recently discovered model, the steering law, can be applied as an evaluation paradigm complementary to Fitts' law. We tested five commonly used computer input devices in two steering tasks, one linear and one circular. Results showed that subjects' performance with the five devices could be generally classified into three groups in the following order: 1. the tablet and the mouse, 2. the trackpoint, 3. the touchpad and the trackball. The steering law proved to hold for all five devices with greater than 0.98 correlation. The ability to generalize the experimental results and the limitations of the steering law are also discussed. Copyright © 2012 ACM, Inc.