You may not be familiar with Fitts’s Law, but designers have been using this scientific law since the birth of the personal computer. Fitts’s Law predicts that
the time required to rapidly move to a target area is a function of the ratio between the distance to the target and the width of the target.
The equation for Fitts’s Law was first introduced in 1954, but it has since been modified include adjustments for accuracy. The most common form of this law is: 
Although Fitts’s Law was originally designed for only one dimensional tasks, it has been extended to two and three dimensional tasks, such as stylus and touch interactions.
Examining Inputs, Fitts’s Style
As a predictive model, Fitts’s Law is heavily reliant on predefined constraints. However, calculating user tasks in terms of bits is a unifying method to compare various input techniques. If we account for input with the dominant and non-dominant thumb in a similar way to Gibson’s study on desktop affordances, then we can also look for biases in bimanual control when examining mobile keyboard affordances.
Using Fitts’s Law would grant insight on the throughput of user interaction on mobile devices between typing and swiping methods. UX research has shown we use different ways of interacting with the touch keyboard on a mobile device. Using this information, I can think of several variations for trials:
- Two-Handed Portrait with thumbs
- Two-Handed Landscape with thumbs
- Holding in non-dominant hand, typing with non-dominant thumb
- Holding in dominant hand, typing with dominant thumb
- Holding in non-dominant hand, typing with dominant thumb
- Holding in dominant hand, typing with non-dominant thumb
- Holding in non-dominant hand, typing with dominant finger
- Holding in dominant hand, typing with non-dominant finger
We could accurately calculate the index of difficult for each of these variation using the Fitts’s Law model. I believe we are moving away from the standard QWERTY keyboards, especially on mobile devices. How would we design mobile UIs differently if we analytically knew the most efficient method of input?