Uma abordagem usando visualização de software como apoio à refatoração para aspectos

• Ano de defesa: 2013
• Autor(a) principal: Delfim, Fernanda Madeiral [UNESP]
• Tipo de documento: Dissertação
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Estadual Paulista (Unesp)
• Palavras-chave em Português: Computação gráfica; Visualização de software; Programação orientada a objetos (Computação); Mineração de dados (Computação); Software visualization
• Link de acesso: http://hdl.handle.net/11449/110381

The evolution of existing software systems to aspect-oriented technology has as first step the aspect mining, which aims to identify crosscutting concerns in source code to be encapsulated into aspects. Several techniques have been proposed for aspect mining, but still with shortcomings. One cause of these shortcomings pointed out in the literature is inadequate presentation of the results obtained by these techniques. Software Visualization can be used to analyze, interpret and combine results of aspect mining techniques, being the results presented with program characteristics. This work presents a visual approach of multiple coordinated views in order to provide an environment for the presentation of the results generated by aspect mining techniques, as well as to improve the understanding of the user to analyze them for future refactoring to aspects. The multiple coordinated views are used to allow the analysis: of associations based on method calls, at class and method levels, allowing visualization of the units call frequency based on fan-in metric; of the control and data dependencies between program instructions; of the program structure; of how instruction sets (slices) are composed in several classes; and of bytecode. The focus is to investigate whether visualization helps in program comprehension by the results generated using program slicing and fan-in analysis techniques, proposals for mining aspects in a complementary way. A software visualization tool, named SoftV is4CA (Software Visualization for Code Analysis), was developed to support the proposed visual approach. The preliminary study showed that the proposed coordination model supports the analysis by exploration of different levels of details