An Empirical Study on Object-Oriented Metrics

Mei-Huei Tang, Ming-Hung Kao and Mei-Hwa Chen


The objective of this study is to investigate the correlation between the object-oriented design metrics and the fault distribution. Such relationship can be utilized to allocate test resources, so that testing can be more effective. We conducted an empirical study on three industrial real-time systems which contain a number of natural faults reported for the past three years. The faults found in these three systems are classified into three types: object-oriented faults, object management faults and the traditional faults. An object-oriented design metrics suite proposed by Chidamber and Kemerer was assessed as the indicator for testing resources allocation. We investigated whether these object-oriented design metrics can be used to estimate fault population and fault distribution in the object-oriented programs. Our results show that existing OO metrics are not sufficient to be used to achieve this goal and we identified new metrics that can be served as a good indicator.