High-impact Refactoring Based on Architecture Violations

TLDR

Software refactoring is a key technique for maintaining and evolving object‑oriented systems, and high‑impact refactorings—those that strongly affect architectural quality—are especially valuable; bad smells and code metrics are commonly used to identify refactoring needs. The authors aim to enhance existing refactoring identification techniques by incorporating architectural violation analysis to spot opportunities for high‑impact refactorings. They applied a combination of tools and techniques to a mid‑sized Java enterprise telecom application that required radical functional extension, in order to identify high‑impact refactoring opportunities. The study reports on the resulting architecture, the refactoring process, tool support, and related experiences from applying these methods.

Abstract

Software refactoring has been identified as a key technique for the maintenance and evolution of object-oriented systems. Most interesting are high-impact refactorings, that is, refactorings that have a strong impact on the quality of the system's architecture. "Bad smells " and code metrics have been suggested as means for identifying refactoring needs. According to our experience these techniques are useful yet, in order to spot opportunities for high-impact refactorings, they should be complemented with the analysis of architectural violations. The subject of this report is a mid-sized Java enterprise application from the telecommunications domain whose functionality had to be radically extended We show how we combined several tools and techniques to identify' opportunities for high-impact refactorings, and discuss the resulting architecture, the refactoring process, tool support as well as related experiences