Software Complexity Analysis Based on Shannon Entropy Theory and C&K Metrics

This work presents an analysis of software complexity based on Shannon Communication Theory using Chidamber and Kemerer (C and K) Metrics. The complexity affects effort and cost of software development and maintenance. The main software artifact is its architecture which depends on good balance of structural decisions based on coupling and cohesion concepts. The entropy concept can help us to measure the disorder level of the system. This work proposes using entropy as reference to control software design based on C&K metrics. I also define thresholds for some metrics and calculate the entropy of jUnit software used as an example. The results are good that suggest next steps in software complexity.