Integração de técnicas para melhoria da qualidade de requisitos: uma abordagem aplicada

Abstract

Context: The faster a business can change its processes and software, the more prepared it will be to react to competitive market events and business rule changes. In this respect, requirements engineering (E.R) stands out as a branch of software engineering responsible for dealing with one of the first stages of development: understanding the customer's needs and defining these in the form of a structured set of requirements that the software must meet. Therefore, the requirements of a software project must be defined in a consistent, complete and correct manner, in order to enable the validation and verification of the software by the customers and the certainty that they were correctly attended to. Objective: This dissertation presents a proposal for improvements in the quality of requirements aimed at evaluating the consistency, completeness and correctness of requirements, registered in the requirements artifacts. Method: First, a mapping was carried out to know the existing processes and artifacts. Sequentially, the Uni-REPM evaluation model was applied to identify the maturity level of the requirements processes. A check was performed on the requirement artifact called Service Order (S.O) for a period of 60 days to identify whether the requirement units described in this artifact had consistency, completeness and correctness, for this analysis the use of the 3'C technique was used. In sequence, a new analysis was performed to identify whether the tasks and actions contained in the requirements processes classified as incomplete by the Uni-REPM evaluation model directly affected the consistency, completeness and correctness of the requirements units. With this result in hand, implementations of improvements in these tasks and actions and new collections were carried out.Conclusion: Upon completion of the three data collections, it was possible to observe that, even with an increase in: 14% in the amount of OS, 141.72 % in the UI, and still an increase in 111.95% in encoding hours, the hours used in rework were reduced by 73.23%.