OOCQM Framework for calculating object-oriented code quality using quality factors and code metrics

Abstract

Purpose: Source code is core of Software engineering. Source code of good quality can be maintained and upgraded easily. Measurement of code quality is important from various views like from developer’s point of view, from manager’s point of view and from end user’s point of view. Main issue in measurement of quality is that quality it needs to be measured from different aspects of software. Methodology: Source code quality can be broken down in factors like maintainability, reusability, change proneness, testability and stability. These quality factors are high level representation of code quality and cannot be measured directly. In order to calculate quality factors there is need for calculations for certain aspects of software code at lower level. Research in field of source code quality has proposed enormous amount of code metrics that measure quality in different aspects like complexity, size, coupling, cohesion and abstractness etc. This study proposes a framework named Object Oriented Code Quality Meter (OOCQM) for measuring source code quality of object-oriented code using low level code metrics and high-level quality factors. Code metrics has a proven relationship with quality factors. This relationship is used to calculate a numerical value for quality factors based on metric values. It is observed that all selected metrics has negative correlation with mapped quality factor. Quality factors scores are aggregated and used to depict quality of code in numerical form. A PHP based tool is developed to validate the results. Results: Framework results are compared with Maintainability Index (MI) which is popular quality measure in terms of code maintainability. OOCQM measures code quality correctly as quality results are correct according to MI. This framework provides more detail at individual quality factors level. OOCQM is compared with few other tools developed for quality measurement. Comparison shows that this tool supports more quality factors analysis than other tools.