Paper Reading #7: Performance Optimizations of Virtual Keyboards for Stroke-Based Text Entry on a Touch-Based Tabletop

• Paper: Performance Optimizations of Virtual Keyboards for Stroke-Based Text Entry on a Touch-Based Tabletop
• View paper here.

• Authors:
• Jochen Rick - Published researcher for the Open University at the time of writing, currently a faculty member in the Department of Educational Technology at Saarland University.
• Source: http://home.cc.gatech.edu/je77/2

• Presented at the 23rd annual ACM symposium on User interface software and technology.

• Hypothesis: Can a stroke based keyboard be developed for a touch-activated surface that would be efficient enough to replace the standard qwerty keyboard?

• Methods: Rick had 8 participants perform a stroke gesture through four points at varying angles and destination sizes and measured the quickness at the various angles as well as segments in the strokes (beginning, middle, and end). The results were then inputted into an implementation of Fitts's Law to determine which keyboard layout would be best suited for the stroke method.

• Results: In terms of speed and of improvement of stroking over tapping, the best results were those of the keyboards with the smallest space between key, such as the ATOMIKs, GAGs, and OPTIs. The wider keyboards, such as the qwerty and DVORAK keyboards, performed among the worst. Based on this data, Rick was then able two optimize two new "OSK" keyboards that outperformed all others.

Summary

     Jochen Rick's article discusses the problems with the way we currently use keyboards on touch screen devices, namely a familiar qwerty keyboard with a touch screen interface. He argues that this is unnatural and that a more efficient solution can be found through stroking gestures. He uses a wide collection of previously created keyboard designs and calculates the most efficient keyboard layout using the methods described above. In his discussion, he brings up potential problems for implementing this keyboard interface into today's devices, such as the difficulty of getting people to adopt this new way of inputting words. He states, however, that if it is adopted, it could increase text input speeds by up to 50%.

My View

     I remember we discussed this very issue in class the other day. My whole issue is that I use an iPod Touch and I'm perfectly fine with using the qwerty keyboard. It's what I've grown up with and what I'm comfortable with and not to say my typing is tremendously fast, but I can't see myself getting too much faster with a new key system.

     Rick does note in his conclusion that getting a wide enough range of people to accept this new keyboard system seems a bit impractical, and I have to agree with that. Most people use touch screens because they are convenient and fairly simple to operate, but mostly because the motions are natural to them. If you remember a couple blogs back, I absolutely geeked out at the idea of Manual Deskterity mainly because everything seemed to be activated by a fairly natural movement. To many of us, typing on a qwerty keyboard has become a natural motion for us, which is why I don't really foresee a future in stroke typing.

     I will say, however, that I fully respect the massive amount of work that went into this research. I never realized how much math could go into something as seemingly simple as designing a keyboard. I think Rick might be on the right track towards developing a widely accepted keyboard system for touch devices, but I don't think it's quite there yet.