A Framework for Improving the Concept of Cyclomatic Complexity in Object-Oriented Programming

Abstract

Zuse defines software complexity as “the software which is difficult to understand and change in future. Mainly three metrics LOC, Halstead’s and cyclomatic complexity are used to measure the complexity of software. There are some problems with LOC and Halstead’s metric. So, to overcome problems, McCabe’s Cyclomatic Complexity was introduced which is the strongest metric among all. But, there was no concept of object oriented at the time when Cyclomatic complexity was introduced. Because of ignorance of this concept in Cyclomatic complexity it was not considered as a strong metric to measure the complexity of programming language like OOPs. Cyclomatic complexity is not considering or measuring the exact software complexity means if there is interaction between two or three object classes in software then it does not calculates that complexity. There is a need to consider that object coupling complexity in McCabe cyclomatic complexity. Firstly, there is a need to consider the concept of coupling. That coupling complexity also needed to be in consideration because a change in one affects other. So, in thesis work an algorithm is purposed to measure the interaction between object classes based on all the unique external class references and along with an approach is proposed to improve the concept of cyclomatic complexity.