A Model Based on Software Quality Factors which Predicts Maintainability

Computer scientists are continually attempting to improve software system development. Systems are developed in a top-down fashion for better modularity and understandability. Performance enhancements are implemented for more speed. One area in which a great deal of effort is being devoted is software maintenance. Brooks estimates that fifty percent of the development cost of a software system is for maintenance activities [BROF82]. Since a large portion of the effort of a system is devoted to maintenance, it is reasonable to assume that driving down maintenance costs would drive down the overall cost of the system. Measuring the complexity of a software system could aid in this attempt. By lowering the complexity of the system or of subsystems within the system, it may be possible to reduce the amount of maintenance necessary. Software quality metrics were developed to measure the complexity of software systems. This study relates the complexity of the system as measured by software metrics to the amount of maintenance necessary to that system. We have developed a model which uses several software quality metrics as parameters to predict maintenance activity.