Perceptible Performance

Many if not most software applications interact with human users. The performance of such applications is defined by the users' perception. As research in human-computer interaction has shown, the key determinant of perceptible performance is the system's lag in handling user events. Thus, to understand the performance of such applications, developers have to focus on perceptible lag.

Our group has conducted a diverse set of studies and developed numerous tools related to perceptible performance.
  • LagHunter - Catching Latency Bugs in the Wild
  • LagAlyzer - Latency Profile Visualization
  • LiLa - Listener Latency Profiling
  • FlyBy - End-User Performance Bug Reporting
  • Trevis - Calling Context Tree Visualization and Comparison
  • Pounder - GUI Capture & Replay (only minor contributions to existing tool)