Abstract:The Qwerty soft keyboard is widely used on mobile devices. However, keyboards often consume a large portion of the touchscreen space, occluding the application view on the smartphone and requiring a separate input interface on the smartwatch. Such space consumption can affect the user experience of accessing information and the overall performance of text input. In order to free up the screen real estate, this article explores the concept of Sparse Keyboard and proposes two new ways of presenting the Qwerty soft keyboard. The idea is to use users’ spatial memory and the reference effect of spatial landmarks on the graphical interface. Our final design K3-SGK displays only three keys while L5-EYOCN displays only five line segments instead of the entire Qwerty layout. To achieve this, we employ a user-centered computational design method: first study the reference effect of a single landmark key (line segment) from empirical data, then make assumptions to generalize the effect to multiple landmarks, and finally optimize the best designs. To make the text entry function more complete, we also design and implement gestural interactions for editing operations and non-alphabetical characters’ input. User evaluation shows that participants can quickly learn how to type with K3-SGK and L5-EYOCN . After five 15-phrase typing sessions, participants achieve 88.1%--92.8% of the full Qwerty keyboard in terms of words per minute on the smartphone and 98.4%--99.1% on the smartwatch. The differences on character and word error rate between our keyboard designs and the full Qwerty keyboard are not significant. The results of out-of-vocabulary words input are also promising. In addition, participants can quickly recall the typing skills and maintain the input performance even after a few days. User feedbacks in real application contexts show that with the low occlusion keyboard, users can acquire more information and perform less scrolling on the smartphone and achieve a higher input efficiency on the smartwatch with a more fluent input experience.

TypeBoard: Identifying Unintentional Touch on Pressure-Sensitive Touchscreen Keyboards.

ResType: Invisible and Adaptive Tablet Keyboard Leveraging Resting Fingers

GhostType: the Limits of Using Contactless Electromagnetic Interference to Inject Phantom Keys into Analog Circuits of Keyboards

ForceBoard: Subtle Text Entry Leveraging Pressure.

PalmBoard: Leveraging Implicit Touch Pressure in Statistical Decoding for Indirect Text Entry.

TOAST: Ten-Finger Eyes-Free Typing on Touchable Surfaces.

Reactions, accuracy and response complexity of numerical typing on touch screens.

The Effect of Tablet Computer Configurations and Touchscreen Gestures on Human Biomechanics, Performance, and Subjective Assessment

VIPBoard - Improving Screen-Reader Keyboard for Visually Impaired People with Character-Level Auto Correction.

Hand Posture's Effect on Touch Screen Text Input Behaviors: A Touch Area Based Study

TapType: Ten-finger text entry on everyday surfaces via Bayesian inference

A Data-Driven Design Evaluation Tool for Handheld Device Soft Keyboards

Investigating Gesture Typing for Indirect Touch

Exploring Low-Occlusion Qwerty Soft Keyboard Using Spatial Landmarks

Recognizing Unintentional Touch on Interactive Tabletop.

Typing on Split Keyboards with Peripheral Vision.

A Monte Carlo Simulation Approach for Quantitatively Evaluating Keyboard Layouts for Gesture Input

Itype: Using Eye Gaze to Enhance Typing Privacy

I-Keyboard: Fully Imaginary Keyboard on Touch Devices Empowered by Deep Neural Decoder

PrinType: Text Entry via Fingerprint Recognition.

An Assessment Method and Toolkit to Evaluate Keyboard Design on Smartphones