Priorização de classes para testes unitários baseada em multicritérios

Abstract

Contemporary society has increasingly using systems that help in some way in performing daily tasks. This use provides a growth in the use of software systems, increasing the need of high-quality software systems. This quality can be obtained by applying proven techniques to software testing, aiming to increase the confidence that the software systems meet the customer needs and do not present a high-rate of defects. There are several studies related to software testing in academia and in industry. Such studies seek for improvements in benchmark testing with the goal of increasing the efficiency and the security of the developed software systems. Tests on object-oriented systems are among the most researched topics on software engineering because the very used programming languages are object-oriented. The prioritization of tests is critical for companies who wants to optimize their development, given that this step can be one of the most expensives in the hole software development process. This prioritization can be based on a multi-criteria basis, using a decision method for finding a good solution. Several tools support the testing activities by automating them and making them more reliable. This research presents concepts related to: testing and analysis of software metrics, how to compute them, how to evaluate them through the AHP (Analytic Hierarchic Process), a proposed process of prioritizing classes for unit testing object-oriented systems, a tool to automate this process, and a case study as a way of evaluation.