AIRDoc – An Approach to Improve Requirements Documents

Requirements models and specifications tend to be plagued by many problems such as requirements that have been abandoned and that are no longer meaningful, descriptions that are unnecessarily long and convoluted, and information that is duplicated. These problems hinder the overall understandability and reusability of software models throughout the whole development process. In this paper we propose an approach called AIRDoc that supports the identification of potential problems that may be present in requirements models. AIRDoc is based on the elaboration of goals and the definition of questions and hypotheses that will be addressed by requirements metrics. In order to improve the quality of requirements models we advocate the use of refactorings and requirements patterns. A case study demonstrates how the AIRDoc has been successfully applied in large requirements model conducted by SERPRO, a Brazilian Government software company.