A mechanism of window switching prediction based on user operation history to facilitate multitasking

Abstract

Multitasking on a PC has long been a common practice for users. However, recent rapid technological advancements often complicate multitasking. At this moment, users can utilize many applications on one or more wide displays simultaneously. Thus, the number of tasks and their windows are dramatically increased. Under these circumstances, window switching costs cannot be underestimated. In this paper, we propose a window switching prediction mechanism to facilitate finding the desired window. The mechanism predicts and suggests which window will be needed for the next task based on user operation history. Prediction precision and efficiency of multitasking were investigated by collecting and analysing logs of actual multitasking on subjects’ PCs. The experimental results show that the mechanism was able to reduce the time costs of window switching. We also found that the subjects tended to use both traditional and the proposed window switching methods.