ABSTRACT
This paper describes an empirical comparison for five-key text entry of three keyboard layouts (Alphabetical, Predictive, and Hybrid). The Alphabetical layout positioned characters in alphabetical order. The Predictive layout employed Bellman and MacKenzie's (1998) Fluctuating Optimal Character Layout (FOCL), relying on letter-pair probabilities to dynamically rearrange letters to minimize selector movement The Hybrid layout combined a fixed alphabetical and a reduced (seven-key) predictive character set Twenty-four participants entered four text types (Words, Sentences, Addresses, and intemet URLs) with each of the three keyboards. Empirical results showed no overall difference in novice performance (speed and error rates) among the three keyboard layouts across the four text types. There were, however, substantial differences among the methods in efficiency and user preference, with the Hybrid layout significantly more efficient than and preferred to the Alphabetical and Predictive layouts.
