Improving software maintenance size metrics A case study: Automated report generation system for particle monitoring in Hard Disk Drive Industry

Abstract

A software maintenance size can be used to predict maintenance effort or maintenance time. However, the traditional maintenance size metric only relies on line of source code (LOC), which hardly is suitable for object-oriented software. This research proposed four new maintenance size metrics based on number of classes, number of methods, average number of methods per class, and weighted methods per class. An automated report generation system in HDD industry (ARGS_PMS) was used as a case study to measure the performance of each proposed metric. We found that, for enhanced software maintenance tasks, maintenance size based on an average number of methods per class (MS-MC) gave the best performance in predicting maintenance time. For a refactoring maintenance task, maintenance size based on line of code (MS-LOC), weighted methods per class (MS-WMC) and average methods per class (MS-MC) were the best, better and good estimators for predicting maintenance time, respectively. On average, the maintenance size based on weighted methods per class gave the best performance.