Ripples: Utilizing Per-Contact Visualizations...

In this paper, several Microsoft researchers attempted to improve the usability of multi-touch displays by adding a simple set of visualizations to cover a variety of common tasks and common mistakes. In order to reduce visual clutter, they attempted to find the least number of visualizations that would cover common errors. Some of these errors include:

• Accidental activation (resting an elbow on the table)
• Object scaling constraints (object has reached maximum size)
• Interaction at a distance (continuing to select a scroll bar while moving away from it)
• Stolen capture (being unable to press a button that is being held down)

After researching various visualizations, the team of researchers conducted a user study testing accuracy and perceived responsiveness. The results from this user study showed that 62% preferred Ripples to be enabled, while 23% preferred it to be disabled (15% had no preference). A second test showed that Ripples consistently improved accuracy for touching small circles around the table.

While I believe that the concept of this paper is sound, there were several points that I felt could have been presented better. For starters, the fact that almost a quarter of the testers preferred Ripples to be disabled is a fairly significant statistic, but the paper mostly breezes right over it. Also, it seems like such a visualization system should be used in a more natural environment (with multiple people using the same touch table), in order to get a better subjective view of the full range of the table's visualizations.