Avaliação de código-fonte orientado a objetos usando requisitos não­-funcionais, métricas e lógica fuzzy

Abstract

The software evaluation to ensure that the product has a quality level and it can be done by automatizated tools that are able to generate quantitatives values for the metrics of software like: number of parameters, number of line of code, cyclomatic complexity. In the literature was proposed the use of these metrics in conjunction with fuzzy logic in order to evaluate the level of code’s quality. The evaluation process is not automated and is stablished a relashinship between the metrics that are necessary to mesure a quality requirement (i.e., the metric coupling between object classes can be used to mesure the quality of the coulpling external requeriment). This work developted a evaluation of source­code using external requirements and creating a relation between metrics and requirements. The quality’s level of source­code is mesured using fuzzy logic for treatment of uncertainty. The automated evaluation allow the developer the create a Project, select the requirements that will be evaluated, create the relation between metrics and external requirements, obtain the numeric values of each metrics in using CKJM extended tools and apply the fuzzy logics phases are done using jFuzzyLogic. The results are show to developer in a grahical form. The advantages of using the automated evaluation is the possibility to integrate the use of metrics with external requiremets and fuzzy logic.