SURVEY OF SYSTEM COMPLEXITY METRICS.

Abstract

Measures of the structured design of software systems are called system complexity metrics. Two particularly promising system complexity metrics are described in this paper: B. H. Yin and J. W. Winchester's metric, which is derived from a system's structured design charts; and S. Henry and D. Kafura's metric, which is derived from a system's information flow. The values computed by both are available after the end of the design phase. Consequently, they are useful in the entire software development life cycle, from the design phase on. The definition, utility, interpretation and advantages of each metric are described. Validation studies and their results are also reported for each metric. It is noted that Yin and Winchester's metric is quite successfully used at Hughes Aircraft Company but that there is no published report of the use of an information flow metric by any software organization.