Abstract

Abstract

Abstract
Software testing is an important component that leads to quality software production. This paper presents the results of a framework for assessing the level of maturity in Software Testing application in the context of Small and Medium-Sized Enterprises (SMEs) based on TMMi model. Our framework includes an evaluation questionnaire based on TMMi sub-practices, support tools with examples of artifacts required to ensure that the questionnaire is thoroughly completed, as well as an automated tool support for its application, enabling SMEs to carry out self-assessment. The framework was applied in ten companies and before the results presented, it can be concluded that the companies maturity in software testing is low and that the companies positively assessed the adequacy of the framework developed for the context of SMEs.

Abstract
Context: Software testing is an important task in software product quality assurance. Alternatives approach for teaching and training students and professionals have been addressed in many studies, recently. In this article, peer testing is investigated as one alternative approach to mitigate the current lack of well-trained professionals. Objective: To verify the efficiency and efficacy of peer testing in order to create test cases to promote software testing education and training. Method: Conduct a controlled experiment in which subjects design and implement test cases for three small software units. Results and conclusions: In this study, it has been observed that peer testing was significantly more efficient than individual testing (with t-Test). On the other hand, it has not been observed significant differences in the efficacy of peer testing.

Abstract
This paper proposes combining the mutation testing technique with evolutionary computing to improve the test data applied to SELECT instructions. From a heuristic perspective, this approach uses Genetic Algorithms to optimize tuples selection from an original database. In other words, generating a smaller amount of tuples able to detect faults in the instructions. Mutants are analyzed to evaluate each set of data tests selected during the evolutionary process. Once the appropriate reduced database was found, it can be used whenever the SQL statement test is necessary. The experimental results indicate that the metaheuristics outperform random methods and reach, in average, 80.3% of the optimal value.

Abstract