A nesting level complexity measure

Abstract

Two well-publicized program complexity measures are software science and cyclomatic complexity. Three areas where these measures do not always follow our intuitive notions of complexity are: structured vs unstructured programs, nested vs sequential predicates, and the use of case statements. This paper defines a nesting level complexity measure that punishes unstructuredness, and the nesting of predicates, and rewards the use of case statements. Examples are given where the nesting level complexity agrees with intuitive rankings of program structures where software science, cyclomatic complexity, and their suggested refinements do not.