A framework based approach for unified detection of Aspect Weaving Problems

In Aspect-Oriented Programming (AOP), a crosscutting concern is extracted from a base program and is modularized into a new construct called aspect. The aspect is weaved back to the program to incorporate the crosscutting behaviors. Many semantic problems, referred to as Aspect Weaving Problems (AWPs) in this paper, could be introduced in the aspect weaving process. The AWPs have been independently defined using different modeling approaches for their detections. In practice, it is difficult for AOP developers to detect these problems, because they have to switch between different tools/methods. This is a time consuming and error-prone process. Consequently, it could increase the production costs of Aspect-Oriented development. This paper proposes a taxonomy of the AWPs to clearly define and classify the problems. Based on the taxonomy and a unified modeling approach to detect the AWPs, a conceptual analysis framework for the AWPs is also proposed. The framework aims at the integration of the unified AWPs detection capabilities into legacy aspect weavers by reusing (adapting) the modeling approach for various AOP languages.