Defesa de qualificação de Francisco Neto, 06/09, 8h

postado em 8 de ago de 2012 12:04 por Nazareno Ferreira de Andrade   [ 29 de ago de 2012 06:41 atualizado‎(s)‎ ]
Candidato: Francisco Gomes de Oliveira Neto
Título do trabalho: Test Case Selection in the context of Specification-Based Regression Testing
Orientador(es): Patrícia Duarte de Lima Machado

Data: 6/9/2012
Horário: 8h
Local: SPLab (local alterado em 29/8)

Banca: Roberta Souza Coelho (UFRN,, Richard Torkar (University of Gothenburg,, Emanuela Gadelha Cartaxo (DSC/UFCG), Tiago Massoni (DSC/UFCG).

Resumo: During software maintenance, several modifications can be performed in the specification model in order to satisfy new requirements or functionalities. The costs related to Regression Testing of the modified software are known to be high and also this activity involves a lot of effort. Test case selection, test case prioritization, test suite reduction among other methods, aim at reducing these costs by selecting or prioritizing a subset of test cases so that less time, effort and thus costs are involved in performing the regression test. In this work, the general problem of automatically selecting test cases based on modifications of the specification model is investigated. The focus is on the Model-based Testing (MBT) approach to provide automation for test case generation and selection based on the specification models of the system. To handle the problem investigated, we have proposed a strategy that combines Similarity based selection with a Value based approach for automatically identifying modifications and classifying the test cases into retestable, reusable and obsolete. The expected benefit is to select the most important test cases related to the modifications identified. Initial studies and the results obtained have shown that the strategy is able to reduce the number of test cases by approximately 70%, whilst revealing regression faults. Thus, the proposed strategy is able to select a small subset of test cases capable of revealing regression faults, and then reduce the costs for regression testing.