From molecular dynamics in virtual reality environments to fundamental questions about human-computer interaction

By nature, the molecular systems that are of interest to structural biologists exhibit high spatial and temporal complexity, in part due to the sheer density of objects and their constant motion. Virtual reality is a very useful tool to manage this complexity and enable interactive simulations that would otherwise be impractical. However, current solutions provide no adequate way to select very small objects whose movements are fast and unpredictable, and theoretical tools such as Fitts' law are of no help with unpredictable motion. To guide the design of appropriate selection techniques, notably with haptic assistance, we studied how subjects performed when attempting to select such objects in a simpler, 2D environment. We propose a new theoretical framework to model selection performance as a function of the predictability of the moving targets.