A Cognitive Model And Gaze-Based Evaluation Of Multiple Representation Use During Program Comprehension And Debugging