On the relations among object-oriented software metrics: A network-based approach

Abstract

Recent years have witnessed a great revolution in software applications. The quality of the software is important insofar as it contributes to providing better services for users. Software metrics are mainly used to obtain feedback on the quality of software design. These metrics enable developers to identify the potential weaknesses of their designs. Furthermore, software metrics may have correlations with each other and impact the outcome of each other. This specific case may cause a misleading interpretation of the design, which eventually affects the quality of software and waste the time and effort during the design phase. Therefore, selecting the appropriate metric during this phase should be carefully performed. In this work, we use network science concepts for deeply investigating the relations among object-oriented software metrics. The analysis approach is based on network visualization and network measurements. The dataset of this work was collected from accredited references in the field of Software Engineering. This study involves the main 104 metrics that are basically used during software design. The findings demonstrated interesting facts on the relations among metrics. Besides, this work is considered as a comprehensive analysis and assessment that takes into account different dimensions of the relations among software metrics. We believe that the analysis and the results can make it easy for developers in selecting the appropriate metrics during the design phase.