My thesis consisted of studying bend gestures and how they can be used as a form of Around Device Interaction to address usability issues in touchscreen mobile devices. We found that training on gesture-mapping positively impacted the user's initial learning (faster completion time and fewer gestures performed), but had a similar outcome on memorability as the users who received no training, while the gestures-without-mapping training led to a negative outcome.

It was unclear whether bend gestures can be easily learned and memorized as control schema for games. To answer this, I built a novel deformable smartphone case that detects bend gestures at its corners and sides, and created Paper Ninja, a mobile game that uses bends as input. I conducted a study comparing the effect of three pre-game training levels on learnability and memorability: no training, training of the bend gestures only, and training of both the bend gestures and their in-game action mapping. My findings suggest that users can learn bend gestures by discovery and training is not essential.