Abstract:

Typing in Indian scripts using virtual keyboards on mobile devices poses several challenges to the users such as lower typing speeds, typing errors and slow learning. The inherent nature of the alphasyllabary style script and complex character set of Indian language scripts and also that of several popular languages of the world like Chinese, Japanese and Arabic makes the keyboard interface difficult to use. In comparison to the research on English text input methods, very minimal research has been done to solve these issues. There being no de facto standard like QWERTY is for English, many commercial Indian language soft keyboard solutions are designed with different layouts and uses different touch screen interactions like tap, long press, slide, pop-up, pie menu, swipe, etc. The question remains that what is the influence of these factors or combinations thereof on the usability of keyboards. There isn’t any framework to design these keyboards to be more effective.

The research involved conducting a series of longitudinal usability evaluations on various keyboards to collect human performance data while typing and deriving empirical models through those. Extensive amount of data was collected from users of various typing expertise levels for a sustained period of keyboard usage through controlled experimental protocol. Modelling approach involved breaking down the complete cognitive process of typing into micro cognitive human activities like searching a target, finger movements, finger touch down, thinking, etc. and deriving models for performance, accuracy, memorability and learning. These models also accommodate the deeper human behavioral aspects of typing such as; strategy of searching targets, sequence of lifts and touch down intervals and strategy of planning subsequent targets.

This research attempts to provide a tool in form of an extensive set of empirical models to predict the typing efficiency and accuracy on any keyboard layout using various combinations of touch interactions. This would enable the designers of keyboards to take optimal decisions and thereby improving the design process. Though the models are currently applied for Indic languages but the research contributes a systematic modeling approach independent of language and can be extended and generalized for many other languages of the world. In a country like India where a significant population of smart phone users are semi-literate and who prefer to communicate using Indic languages, having efficient solutions for Indic language keyboards would touch a huge user base.

Publications:

• Sanjay Ghosh, Anirudha Joshi  , Manjiri Joshi, Nagraj Emmadi, Girish Dalvi, Shashank Ahire and Swati Rangale, Shift+Tap or Tap+LongPress? The Upper Bound of Typing Speed on InScript, 39th Intl. Conference on Computer Human Interaction, (CHI), Denver, May, 2017.
• Sanjay Ghosh, Multimodal Approaches for Text Entry in Indian Language on Mobile, 15th IFIP TC.13 Intl. Conference on HCI (INTERACT), Bamberg, Germany, Sept, 2015
• Sanjay Ghosh and Anirudha Joshi, Text Entry in Indian Languages on Mobile: User Perspectives, 6th India HCI Conference on Human Computer Interaction, New Delhi, India, Dec 2014 
• Sanjay Ghosh and Anirudha Joshi, , Exploration of Usage Patterns for Multimodal Input Interactions, 6th India HCI Conference on Human Computer Interaction, New Delhi, India, Dec, 2014 
• Sanjay Ghosh and Anirudha Joshi, Exploration of Multimodal Input Interaction Based on Goals, 11th Asia Pacific Conference on Computer Human Interaction (APCHI), Bangalore, India, Sept, 2013
• Sanjay Ghosh, Anirudha Joshi, Sanjay Tripathi, Empirical Evaluation of Multimodal Input Interactions, 15th Intl. Conference on Human-Computer Interaction (HCII), Las Vegas, Aug, 2013