6. Related Work

7. Conclusions

This paper makes two main contributions. First, it presents a study of the impact of three user interface components on the time to enter information into a computer: handwriting recognition, adaptive menus, and predictive fillin. Handwriting recognition is slower than typing but is preferred by users. Advances in handwriting recognition may make it faster, but recognition would still be much slower than choosing a value from a menu or predictive fillin. All three components work well together and are preferred by users.

Second, this paper discusses principles for applying adaptive menus and predictive fillin to new application areas. Fields with a few, frequently repeated values are candidates for adaptive menus; functional dependencies indicate candidates for predictive fillin. Whether these characteristics can be learned at run-time is a topic for future research.

Acknowledgments

Kerry Hersh Raghavendra provided the names used in Section 4. Apple Computer developed and supports Newton and the Newton ToolKit programming environment. The Newton AI group at WSU provided many useful comments on an earlier draft of this paper. Geoff Allen, Karl Hakimian, Mike Kibler, and the EECS staff provided a consistent and reliable computing environment. Anonymous reviewers of an earlier draft of this paper provided many (many) valuable suggestions. This work was supported in part by NASA under grant number NCC 2-794.

References


Jeffrey C. Schlimmer, schlimme@eecs.wsu.edu, 5 December 1994