We develop an integrated writing method named GWP that seamlessly switches between typing, word prediction, and predictive gesturing. Predictive gesturing enables a person to enter text by moving the mouse cursor toward a few letters in a word without requiring precise targeting of the letters or gestures towards all the letters in the desired word. Our preliminary work on predictive gesturing demonstrated that people with disabilities created text using 47% fewer precise letter selections than when using word prediction. We propose to improve our predictive gesture analysis algorithm so that it is more error tolerant. We incorporate our integrated writing method into an on-screen keyboard that operates within web-browsers and accesses cloud-based word processors. Developing a browser- based, on-screen keyboard provides access to a wide variety of devices, including smartphones, tablets, and computers. We will evaluate this writing method during real- life use by people with disabilities over a three month period. It is our hypothesis that assistive technology experts will rate GWP as the overall better method for entering text compared to word prediction for most people with disabilities. Experts will base their judgment on user data collected during the three month evaluation period.
The proposed research offers people with severe disability an efficient method for entering text into a broad range of computing devices that alleviates the need for repetitive, precise letter selections.