Mapping Instructions onto Actions: A Comprehension-Based Model of Display-Based Human-Computer Interaction