A proposed software maintenance metric for the object oriented programming paradigm

Abstract

Software complexity metrics have long been used in the procedural programming paradigm. Existing OOP metrics do not address how a class's relation to other classes in the inheritance hierarchy affect it's maintenance needs. This research proposes a new OOP metric, Hierarchy Corrected Metric (HCM) that does address a class's relation to others in the hierarchy. HCM uses the values of the standard OOP metrics for each class plus knowledge of the class inheritance hierarchy to anive at a revised complexity measurement for each class. That measurement reflects not only internal complexity for the class but also the influence of other classes in the hierarchy to the complexity of the class. Using data from a commercial system HCM was shown to correlate with maintenance better than existing OOP metrics.