TechREF: uma técnica de engenharia reversa orientada a features

Abstract

Reverse code engineering plays a key role in various Software Engineering activities, such as model generation from legacy code and retrieval of system features. In the context of Software Product Line (LPS), for example, a software product is composed of a set features that are constantly changed to accommodate changes in business rules. Consequently, the model (for example, UML class diagram) that represents the entire LPS needs to be modified to reflect the updates made. In this context, several tools have been proposed in the last decades, for example, Astah and ArgoUML. However, the tools (and their techniques) do not support feature-oriented reverse engineering, are imprecise in terms of the completeness of the generated diagrams, as well as requiring a high effort to update the models because they are manual or semiautomatic. To mitigate this problem, this paper proposes TechREF, a reverse engineering technique oriented to features. Automatically, TechREF captures the execution flow of code associated with a feature, identifies the structural and behavioral information of the code, and generates diagrams of UML classes, as well as persists such diagrams. TechREF was evaluated through a case study having real reverse engineering scenarios. This evaluation sought to verify the effort and correctness of the activities that will be carried out in the experiment with the use of the models oriented to features.