This study examines the comprehension and modification by novices of programs written with taskspecific and non-task-specific constructs. Teachers who write programs for use in their work studied and modified a program written in Hypercard using either task-specific or non-task-specific constructs. It was found that the task-specific group’s comprehension of the program was significantly better, although the difference was relatively small. However, the correctness of the program modifications was much greater in the task-specific group.