Estudos empíricos sobre abordagens para identificação de interesses transversais na engenharia de requisitos

Abstract

Requirements describe the functions to be offered by the software and possible restrictions on its operation. A group of requirements with the same purpose is called interest. These interests may be scattered and tangled in the system, in which case they are called Crosscutting Concerns (CCs). Identifying and separating these interests is critical to ensuring better software maintainability and evolution. Aspect-Oriented Requirements Engineering (AORE) studies the identification, representation and composition of these interests, also known as aspects, during the requirements engineering. There are several papers that propose approaches to facilitate the identification and separation of Crosscutting Concerns during requirements engineering. However, empirical studies comparing and analyzing these approaches in detail are still scarce. This paper has as main objective to carry out studies on approaches to identify interests of software in documents of requirements. Thereunto, two empirical studies, a Systematic Review of Literature (SLR), were conducted to investigate the state of the art on the subject, and a Experiment with Subjects to compare the effectiveness of two approaches to identify Crosscutting Concerns in requirements documents. The most relevant approaches were identified through an systematic literature review, and by a experiment two of the selected approaches were compared for their effectiveness analysis. With the accomplishment of the studies, the most relevant approaches to AORE were verified in detail, as well as two of them were analyzed by an experimental study that compared their effectiveness. The review made it possible to select the main works of the study area and the results obtained through the experiment indicate that, for the context and scenario predicted, there are significant differences between the approaches to identify concerns under study. With the accomplishment of this paper it was possible to obtain a catalog of the main and most relevant approaches for the identification of Crosscutting Concerns in Requirements Engineering, compiling the knowledge about these approaches. Empirical evidence has been obtained that makes it possible to differentiate the ObasCId and Theme/Doc approaches with respect to their effectiveness.