Optimizing Human Performance in Mobile Text Entry

Although text entry on mobile phones is abundant, research strives to achieve desktop typing performance "on the go". But how can researchers evaluate new and existing mobile text entry techniques? How can they ensure that evaluations are conducted in a consistent manner that facilitates comparison? What forms of input are possible on a mobile device? Do the audio and haptic feedback options with most touchscreen keyboards affect performance? What influences users' preference for one feedback or another? Can rearranging the characters and keys of a keyboard improve performance? This dissertation answers these questions and more.

The developed TEMA software allows researchers to evaluate mobile text entry methods in an easy, detailed, and consistent manner. Many in academia and industry have adopted it. TEMA was used to evaluate a typical QWERTY keyboard with multiple options for audio and haptic feedback. Though feedback did not have a significant effect on performance, a survey revealed that users' choice of feedback is influenced by social and technical factors.

Another study using TEMA showed that novice users entered text faster using a tapping technique than with a gesture or handwriting technique. This motivated rearranging the keys and characters to create a new keyboard, MIME, that would provide better performance for expert users. Data on character frequency and key selection times were gathered and used to design MIME. A longitudinal user study using TEMA revealed an entry speed of 17 wpm and a total error rate of 1.7% for MIME, compared to 23 wpm and 5.2% for QWERTY. Although MIME's entry speed did not surpass QWERTY's during the study, it is projected to do so after twelve hours of practice. MIME's error rate was consistently low and significantly lower than QWERTY's. In addition, participants found MIME more comfortable to use, with some reporting hand soreness after using QWERTY for extended periods.