Design Metrics Which Predict Source Code Quality

Hartson, H. Rex and Smith, Eric C. and Henry, Sallie M. and Selig, Calvin (1987) Design Metrics Which Predict Source Code Quality. Technical Report TR-87-32, Computer Science, Virginia Polytechnic Institute and State University.

Abstract

Since the inception of software engineering, the major goal has been to control the development and maintenance of reliable software. To this end, many different design methodologies have been presented as a means to improve software quality through semantic clarity and syntactic accuracy during the specification and design phases of the software life cycle. On the other end of the life cycle, software quality metrics have been proposed to supply quantitative measures of the resultant software. This study is an attempt to unify the concepts of design methodologies and software quantity metrics by providing a means to determine the quality of a design before its implementation. By knowing (quantitatively) the quality of the design, a considerable amount of time and money can be saved by realizing design problems and being able to correct these problems at design time. All of this can be accomplished before any effort has been expended on the implementation of the software. This paper provides a means of allowing a software designer to predict the quality of the source code at design time. Actual equations for predicting source code quality from design metric values are given.