A THEORY OF SYSTEM COMPLEXITY

Abstract

A theory is developed, using the basic principles of information theory, which explains the error behavior in systems. The results are system and configuration independent and have applicability across many disciplines. As an example, the results may be used in understanding the defect properties of computer programs. An explicit formula is given for deriving the critical size of systems. The effect of modularity on system properties is analyzed and the optimum size of a subsystem is derived. It is shown that modularity increases the critical size and decreases the number of expected faults in the system.