Comprehending parametric CAD models

In this study, we experimentally evaluated two GUI prototypes (named "split" and "integrated") equivalent to those used in the domain of parametric CAD modeling. Participants in the study were asked to perform a number of 3D model comprehension tasks, using both interfaces. The tasks themselves were classified into three classes: parameterization, topological and geometrical tasks. We measured the task completion times, error rates, and user satisfaction for both interfaces. The experimental results showed that task completion times are significantly shorter when the "split" interface is used, in all cases of interest: 1) tasks taken as a whole and 2) tasks viewed by task type. There was no significant difference in error rates; however, error rate was significantly higher in the case of parameterization tasks for both interfaces. User satisfaction was significantly higher for the "split" interface. The study gave us a better understanding of the human performance when perceiving and comprehending parametric CAD models, and offered insight into the usability aspects of the two studied interfaces; we also believe that the knowledge obtained could be of practical utility to implementers of parametric CAD modeling packages.