Examining the qualitative gains of mediating human interruptions during hci

Recent trends in software development directed toward intelligence, distribution, and mobility have brought sophisticated software artifacts that often come with some unwanted side effects; frequent interruptions, for instance. In general, people are less effective when exposed to interruptions. We have created a framework that helps in selecting the most appropriate timing for interruption as a way to mediate human interruptions by the computer. The proposed framework is based on a new Interruption Taxonomy and uses Bayesian Belief Networks for selecting the best timing when mediating interruptions. An empirical study was conducted to empirically measure the qualitative gains of mediating interruptions compared to condition with no interruption control i.e., interruptions were presented immediately at random-generated points. The experimental results suggest that mediated coordination of interruption was effective in decreasing some disruptive effects of interruptions on a socioemotional level, such as: the feelings of frustration, distraction, annoyance and workload.