Uma Proposta de MÃtricas para Avaliar Modelos i*
AUTOR(ES)
Emanuel Batista dos Santos
DATA DE PUBLICAÇÃO
2008
RESUMO
The Goal-Oriented Requirement Engineering has shown a promising way to describe software systems. It provides a natural mechanism to structure complex requirement documents through goals providing a mechanism to justify the existence of the requirements and facilitate the management of conflicting requirements. In this context, different approaches that use Goal abstraction have emerged, such as KAOS, NFR, V-Graph and i*. The demand for quality software requires that all artifacts produced during the software development should also be of high quality. Artifacts and documents of requirements, executable code and the system must be free of errors and failures, because problems identified in early stage are cheaper to correct. Thus, identify problems in requirements stage reduces the need and costs of correction in future stages. Similar to traditional techniques, Goal-Oriented approaches also demand mechanisms to insure the quality of their artifacts. In this dissertation we present metrics to assess the quality of i* models used for early requirement stage. The metrics matches qualities that we desired of requirement documents to basic i* constructions. In doing so we provide some effective mechanisms to identify problems in i* models. The metrics are grouped to address the following issues: typical errors, level of detail, ambiguity and complexity. In order to facilitate interpretation of values obtained by the use of metrics we presented how they can be used in connection with the GQM Method (Goal-Question-Metric). Moreover, we also provide a prototype of toll o support the automatic collection of metrics
ASSUNTO(S)
requirements engineering engenharia de requisitos quality evaluation mÃtricas para modelos i* metrics for i* models ciencia da computacao avaliaÃÃo de qualidade
