A Survey of Concern-Oriented Development Approaches

Concern-oriented development has been of significant interest to the software engineering community since the early formulation of its importance by Dijkstra and Parnas. Many different approaches have emerged since then, and continue to emerge, which support developers when they need to reason about and make changes to code in terms of the concerns that govern the system. While initially much work focused on modularization of concerns, relying on programmatic constructs to encapsulate concerns in modules, new types of approaches have more recently emerged that address the problem from quite different directions and lead to quite different forms of support. This survey is inclusive of the breadth of approaches to date. Its main contributions are a set of definitions of concerns that precisely delineate the field, an evaluation framework that characterizes a set of capabilities that any approach to concern-oriented development ideally should offer, an articulation of how four, previously disparate categories of approaches closely relate, and a set of nine observations that take stock of the field as it stands today and provide guidance as to where it should head in the future.