Representing knowledge in requirements engineering: an empirical study of what software engineers need to know

The selection and structuring of requirements engineering data necessitates an empirical prediction about the future utility of the selected data and of the retrieval paths available in the structure. In order to investigate these predictions, the questions asked by software engineers at software requirements and preliminary design meetings are analyzed. It is assumed that the kinds of information elicited by these questions are good indicators of what data would be most useful. Results lend support to systems that focus on the developing artifact itself, and indicate the importance of user scenarios