A method for aspect mining using production rules, dependency graphs and two-level grammars

Adopting aspect-oriented technologies requires revisiting and restructuring the entire traditional software lifecycle in order to identify and represent occurrences of crosscutting during software requirements engineering and design, and to determine how these requirements are composed. In this research, we propose an aspect mining approach to identify and model crosscutting concerns (aspects) by restructuring the meta-level at the breakpoints of: requirements elicitation, analysis, design, and implementation activities. The main purpose of this research is twofold: (1) "isolation" of crosscutting concerns in the early steps of software life cycle (2) identification of crosscutting concerns in legacy systems. The proposal outlined in this research illustrates a method to reformulate existing object-oriented artifacts by mining and explicitly modeling crosscutting concerns. In this method of restructuring we represent the semantics of UML artifacts by using a set of production rules, which refer to object definitions and relationships metadata. The result of our proposed restructuring process is an aspect-oriented design whereby aspects may be introduced into the object-oriented design or certain classes may be restructured as aspects.