A Variability-Driven Analysis Method for Automatic Extraction of Domain Behaviors

Abstract

Domain engineering focuses on modeling knowledge in an application domain for supporting systematic reuse in the context of complex and constantly evolving systems. Automatically supporting this task is challenging; most existing methods assume high similarity of variants which limits reuse of the generated domain artifacts, or provide very low-level features rather than actual domain features. As a result, these methods are limited in handling common scenarios such as similarly behaving systems developed by different teams, or merging existing products. To address this gap, we propose a method for extracting domain knowledge in the form of domain behaviors, building on a previously developed framework for behavior-based variability analysis among class operations. Machine learning techniques are applied for identifying clusters of operations that can potentially form domain behaviors. The approach is evaluated on a set of open-source video games, named apo-games.