The complexity measurement of software through program decomposition

A new methodology of measuring software complexity is proposed based on pro gram decomposition mechanisms. The purpose of our research is to improve th e quality and reliability of software by using a good measurement of the co mplexity of the software. Thus software development cost can be reduced due to the fewer errors introduced. We present a complete survey of Various me trics in the literature. Six types of metrics or measurement techniques are discussed, including some earlier research results of the authors. We then propose a program decomposition mechanism based on the operational semanti cs of several languages constructs in most procedural languages. An algorit hm along with some examples is also given to show the feasibility of our me chanism. Our approach, relies on the path complexity of a program, is more accurate and easier to be realized. The algorithm also points out that how many individual complete paths of a program need to be tested.