Quill: A gesture design tool for pen-based user interfaces

This dissertation describes the motivation, design, and development of a tool for designing gestures for pen-based user interfaces. Pens and other styli have been ubiquitous for recording information for centuries. Recently, pen-based computers have recently become common, especially small devices such as the Palm Pilot. One benefit pens provide in computer interfaces is the ability to draw gestures—marks that invoke commands.; Gestures can be intuitive and faster than other methods of invoking commands. However, our research shows that gestures are sometimes misrecognized and hard to remember. We believe these problems are due in part to the difficulty of designing “good” gestures—that is, gestures that are easy to remember and are recognized well—and the lack of tools for helping designers create good gestures. We believe that an improved gesture design tool can help interface designers create good gestures for their applications.; Since people confuse similar objects and misremember them, we performed experiments to measure why people perceived gestures as similar. We derived computational metrics for predicting human perception of gesture similarity.; Based on the results of our experiments, we developed a gesture design tool, quill. The tool warns designers about gestures that may be hard to remember or recognize, and provides advice about how to improve the gestures. It also provides a convenient way to test recognition of gestures.; To evaluate quill, a user study was performed with 10 professional user interface designers and one professional web designer. All designers were able to create gesture sets using quill, but not all designers benefited from quill's suggestions. More work is needed to make suggestions useful for most designers.; The primary contributions of this work are: (1) Improved understanding of the gesture design process, including the types of problems people encounter when designing gestures. (2) Computational models for predicting human-perceived gesture similarity. (3) Confirmation of the importance of good naming for gesture memorability. (4) An intelligent gesture design tool, quill, which automatically warns designers about potential problems with their gestures and advises them about how to fix these problems.; This work also suggests several areas for future work in the areas of gesture design tools and gesture similarity and memorability.