Planning for the User-Interface: Window Characteristics

While manipulating an interface, users often generate a proliferation of windows that overlap with each other and cause disorganization. Given a cluttered screen, many actions are possible to reorganize the configuration including moving and minimizing windows. This paper claims that three types of knowledge xist to aid the user and the system in intelligently reducing the clutter. They are the knowledge of window characteristics, window content, and user problem-solving context. We present a system called P4P. It implements a planner that currently uses the first and second knowledge types to achieve a state of reduced screen-clutter given arbitrary window configurations. Two examples illustrate the process. We conclude that, while necessary, knowledge of geometrical window characteristics alone is not sufficient to effectively manage the user interface.