Navegando por Data de Publicação, começando com "2015-07-13"

Filtrar resultados por ano ou mês

Agora exibindo 1 - 1 de 1

Uma contribuição para a avaliação da qualidade de modelos na fase de projeto de software
(Universidade Federal de São Carlos, 2015-07-13) Souza, Odair Moreira de; Ferrari, Fabiano Cutigi; https://lattes.cnpq.br/3154345471250570; https://lattes.cnpq.br/8447938107204065
Context: Software engineering aims to provide software development with high quality in short time and low cost. In this context, product quality assurance goes through all phases of development process, avoiding the generation of high cost in corrective maintenance in advanced phases of development. In the design phase, a mean of anticipating the detection of structural and behavioural problems and reduce rework is by applying metrics in design models. Objective: This work aimed to provide a systematic and automated process for the evaluation of software design models based on internal metrics and external quality attributes. A conceptual framework, entitled QM2, and automated tools have been proposed, which include the main activities and resources needed for assessing the quality level of models during the design phase. Method: For the definition of QM2, studies have been conducted to identify the essential activities for design evaluation, addressing quality models, internal metrics and reusable software resources. For the evaluation of the the framework and tools developed, were addressed models of a robotic em¬bedded system represented in UML. Results: Was established a sequence of steps to evaluate the quality level of software models. Was performed the automation of the QM2 framework, including tool for metrics extraction and the management of assessment projects. Conclusion: The simple definition and extraction of metrics do not ensure the quality of a software project, because the interpretation of values should be associated with external quality attributes. In that sense, in this work the assessment of design is performed by means of extraction and calculation of model metrics. The quality level of a model is obtained through the association between internal measures computed and external quality attributes.