Towards Better Command Selection Mechanisms for GUIs

点击翻译按钮获取中文摘要

Continuing improvements in input devices and the development of new interaction techniques make it difficult for researchers and practitioners alike to identify the key traits of graphical command selection mechanisms. In this project, the PI will address this problem by developing a new taxonomy of command selection mechanisms used in graphical interfaces. The goal is to identify the fundamental traits of each technique, understand the pros and cons of each of these traits, and study their possible interactions. To this end, the PI will develop a notation that provides a powerful tool for comparing existing technique, and that will also serve as a road map for fruitful future development in the domains of sensing and interaction design. Following the example of previous models (e.g. the GOMS family), the PI's notation will be grounded in systematic study and empirical evaluations of the key traits relevant to command selection. This will allow for accurate predictions of the performance of new techniques and provide a clear, quantitative understanding of the relative importance of different options in creating command selection mechanisms for new families of devices. As part of this phase of the work, the PI will investigate how users transition from a mode of operation requiring visual feedback to a mode of operation relying mostly on muscle memory. While the former is relatively well understood and captured by Fitt's and Accot's laws, the latter is still experimentally unexplored. This line of study will use eye-tracking technology as its main investigative tool. Finally, the PI will illustrate the potential of his notation by re-evaluating existing techniques, and by developing an efficient new interface for tablet computers based on goal-crossing instead of pointing-and-clicking. Recent empirical results have shown that goal crossing is at least as efficient as aiming in many common tasks. This makes the goal-crossing paradigm a strong contender as a fluid and efficient command selection mechanism for pen computing.Broader Impact: By providing a better understanding of the interplay among hardware, interaction structure, and user skill level, this research will create a solid foundation for a systematic exploration of the design space. The PI's findings will reduce the intrinsic complexity of basic graphical command selection. In particular, the studies on the transition from procedural to autonomous interaction will help intermediate users reach expert performance more rapidly, and improve ease of use for difficult interactions such as navigating a cascading menu. Finally, exploration of a new interaction paradigm based on goal-crossing will allow users of pen-based systems to interact more fluidly with these devices. One of the main outcomes of this part of the project will be a new toolkit based on goal-crossing interactions, which the PI will make freely available to other researchers.