Application of software quality metrics to a relational data base system
It is well known that the cost of large-scale software systems has become unacceptably high. Software metrics by giving a quantitative view of software and its development would prove invaluable to both software designers and project managers. Although several software quality metrics have been developed to assess the psychological complexity of programming tasks, many of these metrics were not validated on any software system of significant size.
This thesis reports on an effort to validate seven different software quality metrics on a medium size data base system. Three different versions of the data base system that evolved over a period of three years were analyzed in this study. A redesign of the data base system, while still in its design phase was also analyzed.
The results indicate the power of software metrics in identifying the high complexity modules in the system and also improper integration of enhancements made to an existing system. The complexity values of the system components as indicated by the metrics, conform well to an intuitive understanding of the system by people familiar with the system. An analysis of the redesigned version of the data base system showed the usefulness of software metrics in the design phase by revealing a poorly structured component of the system.