13. A Metric Suite for Predicting Software Maintainability in Data Intensive Applications

Software maintainability is the vital aspect of software quality and defined as the ease with which modifications can be made once the software is delivered. Tracking the maintenance behaviour of a software product is very complex that is widely acknowledged by the researchers. Many research studies have empirically validated that the prediction of object oriented software maintainability can be achieved before actual operation of the software using design metrics proposed by Chidamber and Kemerer (C&K). However, the framework and reference architecture in which the software systems are being currently developed have changed dramatically in recent times due to the emergence of data warehouse and data mining field. In the prevailing scenario, certain deficiencies were discovered when C&K metric suite was evaluated for data intensive applications. In this study, we propose a new metric suite to overcome these deficiencies and redefine the relationship between design metrics with maintainability. The proposed metric suite is evaluated, analyzed and empirically validated using five proprietary software systems. The results show that the proposed metric suite is very effective for maintainability prediction of all software systems in general and for data intensive software systems in particular. The proposed metric suite may be significantly helpful to the developers in analyzing the maintainability of data intensive software systems before deploying them.