Evidência empírica das abordagens composicional e anotativa para gerência de variabilidades em linhas de processo de software

Abstract

Every day the level of demand from customers with regard to the quality and complexity of software systems increases. Because of this, companies are increasingly customizing their software processes according to market needs. Thus, a systematic way of achieving this goal is the adoption of the Software Process Line (SPrL) strategy, in which each product derived from its variabilities is a specific software process. Taking this scenario in consideration, the objective of this study is to empirically evidencing the compositional and anotative approaches for managing variability in software process. To do so, an OpenUP and a SCRUM-based SPrLs were taken into account. Furthermore, the Eclipse Process Framework was chosen to represent the compositional approach, whereas SMartySPEM was considered for the annotative approach. In order to reaching the settled objective the exploratory sequential strategy was adopted, in which two empirical studies were carried out: a qualitative study analyzing specific criteria for SPrL (modularity, traceability, detection of errors, granularity, sistematic management of variability and adoption) and a quantitative study aiming to substantiate the effectiveness of the approaches with respect to the systematic management of variability. Thus, the qualitative study built a body of knowledge for the quantitative study. As a main contribution of this work, evidence was provided with regard to variability modeling in SPrL, showing that, so far, it is not possible to point out significant differences on such approaches taking the variability criterion into consideration. Further studies in industry with well qualified practitioners are needed to generaliza the initial obtained results.