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

• Title:
• Performance Optimizations of Virtual Keyboards for Stroke-Based Text Entry on a Touch-Based Tabletop
• Reference Information:
• Jochen Rick. 2010. Performance optimizations of virtual keyboards for stroke-based text entry on a touch-based tabletop.  In <em>Proceedings of the 23nd annual ACM symposium on User interface software and technology</em> (UIST '10). ACM, New York, NY, USA,  77-86. DOI=10.1145/1866029.1866043 http://doi.acm.org/10.1145/1866029.1866043
• UIST 2010 New York, New York.
• Author Bios:
• Jochen "Jeff" Rick is the sole author of this paper.  He received his Ph.D. from Georgia Tech.  From 2007 to 2010 Rick was a research fellow at the Opera University.  He is currently a professor at Saarland University.
• Summary
• Hypothesis:
• A stroke-based virtual keyboard will be more efficient and more accurate for input on a touch device.
• There are more efficient layouts for a stroke-based device than the qwerty keyboard.
• Methods
• The author wanted to create a mathematical model which he could then use to compare the keyboard layouts.  To create the mathematical model he performed a user study to collect information about the speed of stroking in different directions and of turning times.  Once the author collected this information he analyzed it and created visual representation of it.  He used that data to figure out constants to plug into equations, one of them being Fitt's law equation for tap input.  Using these created a comparison for the speed, and therefore efficiency, of the different layouts.
• Results
• As predicted, compact layouts worked well for stroke-based text entry.  The reason for this is that there is a smaller distance that the finger is required to move between each character entered.  Similarly, wide keyboards which are efficient for tap-based text entry, such as Qwerty and Dvorak, performed poorly.  Also as predicted, stroke-based entry methods benefit from not requiring the space key be entered between words.  Finally, this study found a small contradiction with a previous study in that the OPTI II layout outperformed the Metropolis layout in tapping.  
• Contents
•   This paper presents a model for comparing the efficiency of different keyboard layouts for both tapping-based text entry as well as stroke-based.  This allows various layouts to be evaluated at a much higher rate of speed.  One of the largest barriers to introducing a new traditional (tap-based physical) keyboard is the strong support and familiarity with the current qwerty board.  The introduction of new entry methods (such as touch based devices) present an opportunity to introduce a new layout since users are already acquiring new input skills.
• Discussion
• While the work done in this study does not seem ground breaking to me, but valuable nonetheless.  The idea that a more compact keyboard would be a better fit for stroke-based input seems completely plausible.  The most interesting aspect about how this study was completed for me was that any hexagonal key layout was converted into a square key layout.  This seems to remove some of the advantage of designing a hexagonal key layout.  In particular, the hexagonal keys are designed to aid in the speed at which a key can be reached from any direction which, in turn, is an advantage to the more compact layout in the first place.  I'd be interested to see if the hexagonal key actually made any measurable difference in performance over square keys.
  
  
  

  
  

  
  

Picture Source: " Performance Optimizations of Virtual Keyboards for Stroke-Based Text Entry on a Touch-Based Tabletop"