Effectiveness of Software Metrics for Object-oriented System

Measurement is a buzzword applied to each and every field of engineering for the development of hardware or firmware products, and software engineering is nowhere an exception. Measurement in software engineering can be best described in terms of metrics, as a quantitative measure of degree to which a system, a component or process possesses a given attribute. The urge for software metrics allows the analysts, designers, coders, testers and the managers to conceive the software development processes and assess the system. The objective of this paper is to evaluate software like ATM using available subset of metrics from traditional and object-oriented methodology. The traditional metrics such as cyclomatic complexity, size and comment percentage are used to compute the software complexity. This paper also analyses a widely used subset of object-oriented metrics such as the Chidamber and Kemerer metric suite to compute the system reliability. The metric values are evaluated for a real life application, which helps us to know the complexity and the reliability of the ATM software.