Abstract
This paper presents results from the first of two empirical studies which examine the effectiveness of guidelines for use case authoring. The ESPRIT 21.903 CREWS long-term research project has developed style and content guidelines for authoring use cases for requirements acquisition and validation. The effectiveness of these guidelines has been evaluated under different conditions. Results indicate that : i. the authoring guidelines improve the overall quality of the use case prose, ii the different guidelines work differently and with different levels of efficiency, and iii use cases are never entirely correctly written ; thus, they can be systematically corrected. The paper details a qualitative and quantitative comparison between guided and non-guided use case authoring. It outlines lessons learned and implications for the CREWS software tools design.