The current work builds on the VelociTap touchscreen decoder
[4]. VelociTap takes a sequence of noisy touch locations
and searches for the most probable sentence given
those touches. The decoder has also been used to to research
eyes-free touchscreen input [5].
in 2015 by McGill et al. [2] provides evidence that use of
a mechanical keyboard with feedback does provide good
performance. However, this study focused primarily on providing
visual feedback in the form of a video of the user. In
the present work, we focus on using a decoder to correct
errors.