Aspect mining using event traces

Abstract

Aspect mining tries to identify crosscutting concerns in existing systems and thus supports the adaption to an aspect-oriented design. This paper describes the first aspect mining approach that detects crosscutting concerns in legacy systems based on dynamic analysis. The analysis uses program traces that are generated in different program executions as underlying data pool. These traces are then investigated for recurring execution patterns based on different constraints, such as the requirement that the patterns have to exist in different calling contexts in the program trace. The implemented approach was evaluated in several case studies over systems with more than 80 kLoC. The tool was able to identify automatically both existing and seeded crosscutting concerns. © 2004 IEEE.