Analyzing Variability of Software Product Lines Using Semantic and Ontological Considerations.

Software Product Line Engineering (SPLE) is an approach to systematically reuse software-related artifacts among different, yet similar, software products. Previewing requirements as drivers of different development methods and activities, several studies have suggested using requirements specifications to identify and analyze commonality and variability of software products. These studies mainly employ semantic text similarity techniques. As a result, they might be limited in their ability to analyze the variability of the expected behaviors of software systems as perceived from an external point of view. Such a view is important when reaching different reuse decisions. In this paper we propose to introduce considerations which reflect the behavior of software products as manifested in requirement statements. To model these behavioral aspects of software requirements we use terms adapted from Bunge's ontological model. The suggested approach automatically extracts the initial state, external events, and final state of software behavior. Then, variability is analyzed based on that view.