Empirical study on evolution of open source software using Object Oriented metrics

This paper presents an empirical study to analyze the quality of open source software JFreeChart, during its evolution. CKJM tool is used in the experiment to measure the changes in structure of the software. It gives a quantified value called as „metric‟. These metrics are used to analyze the quality of the software. Software metrics used in the analysis are Ca (Referred by), Ce (dependent on), CBO (Coupling Between Objects) and LCOM3 (Lack of Cohesion of Methods) metrics. An empirical study is conducted based on two assumptions. Firstly, as the class size increases coupling should increase but cohesion remains same. Secondly, we assume that removed classes of older version have high coupling and low cohesion than added classes in the new version. Based on the empirical study we formulated some basic hypothesis which supports our study and an experiment is conducted to verify whether the hypothesis is proved