| The effects of latency, latency jitter, spatial jitter, and signal dropouts are evaluated for tasks involving acquiring stationary targets (pointing) and following moving targets (pursuit tracking). The properties of the computing system are thoroughly measured and the experiments were carefully designed to separate all the measured effects. The results of the experiments are incorporated into predictive models allowing one to calculate the effects of the aforementioned factors on pointing throughput and tracking errors in advance, i.e., without performing time-consuming empirical evaluations.;For tracking tasks, the throughput measure is computed. Although this measure was introduced as early as in 1960, it has not been widely adopted. Using the same metric of throughput for two input tasks of different nature provides an insight at how the fundamental efficiencies of different input modes compare and can be used to estimate the speed at which information could be manually entered into a computer. Parallels between this measure, applied to a tracking task, and the Steering Law, applied to a task in which one needs to navigate a pointer through a tunnel, are highlighted.;The findings can be used in designing human-computer interaction scenarios, in particular, when developing electronic games controlled by pointing devices. In addition, device throughput as a measure can be used to predict the theoretical limit of information entry rate with a pointing device both in pointing and in tracking mode. |
