Computer-aided window identification in Trident

Multi-windowing capabilities of modern graphical workstations and personal computers provide fundamental features which are necessary to achieve a usable user interface (UI). This usability depends heavily on judicious application of these capabilities. Modifiable variables include reliance on task analysis, respect of guidelines, respect of ergonomic criteria. The problem of correct window identification is therefore crucial in this context. After reviewing the state of the art, this paper examines how windows can be identified in a computer-aided fashion in TRIDENT. The use of an Activity Chaining Graph (ACG), resulting from task analysis, is one of the basic assumptions of this project. Theoretical elements for formalizing UI presentation are defined with respect to this ACG. Then, a complete methodological approach using these elements is provided, consisting of: · an identification of presentation units for each interactive task leading to five types of window identification; · an identification of windows for each presentation unit: according to interaction styles and priorities, a particular type of window identification is retained and applied algorithmically to obtain windows of first rank; these windows become higher rank windows when they are aggregated by techniques to be applied if specific criteria are satisfied.