An experiment to test user validation of requirements: data-flow diagrams vs task-oriented menus

Abstract

A major concern for systems developers is the problem of user validation of system requirements. To the extent that the requirements representation elicits important feedback from the user, the representation technique can be said to be appropriate for user validation. Moran suggests that users develop a conceptual model of a system defined as the “knowledge that organizes how the system works and how it can be used to accomplish) tasks”. An experiment is conducted to determine whether a task-oriented, downward-cascading menu representation (similar to what the user would find in a prototyping environment), which closely corresponds to Moran's idea of a user's conceptual model, permits better comprehension and therefore better validation than data-flow diagrams, which have been strongly advocated as facilitating user validation. A within-group, counterbalancing technique to mitigate subject variability was used to test subjects' comprehension of data-flow diagrams vs the task-oriented menu representations of requirements. Subjects scored significantly higher on the test using the menu representation, indicating that task-oriented menus may be more effective in user validation of requirements than data-flow diagrams.