Evaluating the efficiency of using the Autonomy Ratio Metric for assessing ArgoUML architecture

Niculescu, Mihnea ; Dugerdil, Philippe (Dir.)

Mémoire de bachelor : Haute école de gestion de Genève, 2013 ; TDIG 101.

Metrics in software engineering are used to evaluate quantitatively and qualitatively various attributes of (usually large) systems. These figures help synthetizing information such as size, quality or complexity of various element of the analyzed software. In the past few years, Professor Philippe Dugerdil has developed, at the Geneva School of Business Administration, a new metric, called the... More

Add to personal list
    Summary
    Metrics in software engineering are used to evaluate quantitatively and qualitatively various attributes of (usually large) systems. These figures help synthetizing information such as size, quality or complexity of various element of the analyzed software. In the past few years, Professor Philippe Dugerdil has developed, at the Geneva School of Business Administration, a new metric, called the Autonomy Ratio, along with an analysis method and related software tools. The AR metric helps measuring the “functional structuring” of a system and indicates how easy is to understand the analyzed system. System understanding is very important for maintenance, which is usually the most expensive task in software engineering. Up to this day, the analysis method has been tested by Mr. Dugerdil and some other students at Geneva School of Business Administration on one large industrial system. It has revealed weakness in the architecture of this application that has been confirmed by the development team of the application. However, in order to validate the metric, more systems have to be analyzed. In the work presented here, I used the Autonomy Ration metric and the tools developed in the school to asses ArgoUML, a well-designed, mid-size open source application. I evaluated the efficacy of the method, enhanced the existing tools and proposed some improvements.