A Novel Method to Measure Comprehensive Complexity of Software Based on the Metrics Statistical Model

摘要

Calculating software complexity is one of the most challenging problems in the Software Engineering due to using them in estimating errors, having a landscape of software reliability, approximating costs of software implementation and maintenance, and delivering software with better quality. Most of the recent researches on calculating the software’s complexity focus on special directions and goals. This paper presents a novel method for measuring comprehensive complexity of software based on Statistical model evaluation of the existing complexity metrics through modules. To reach this purpose, the amount of comprehensive complexity is achieved for every module by identifying statistical distribution of complexity metric quantities, normalization and their combination. Afterward, the comprehensive complexity of the software is calculated by composition of the module's complexity amounts. This method is applied on some samples of the "NASA Software Engineering laboratory" and some of its positive results are presented.