Modelling touch-interaction time on smartphones

Accurate performance models are important for interaction technique, application, and hardware design. The limited screen size of mobile devices and use of touch interaction require unique considerations, especially when interacting with large amounts of information. This paper considers the performance impact of target visibility on mobile smartphone applications that provide on- and offscreen content with the commonly used direct-touch interactions and four cursor-based interaction methods for precise selection. Three existing and 12 novel performance models are experimentally validated. Fitts' Law, which was not designed for modelling selection of offscreen targets, did not predict interaction times for mobile interaction methods as accurately as is commonly observed with desktop interaction with onscreen targets. Target visibility was found to greatly impact interaction times (particularly for direct-touch interaction). The presented models that incorporate variables related to target visibility greatly improve predicted interaction times. The use and merits of the top models are discussed, emphasizing the importance and implications of accepted user-interface design guidelines.