Impact of class attributes on cognitive complexity

Abstract

Understanding, an internal process of human beings, is difficult to measure but not impossible. Therefore attempts have been made to measure the understandability of software system in terms of its complexity. Understandability of source code can be measured in terms of its cognitive complexity which is also called psychological complexity. This paper presents a metrics for measuring understandability of a class integral to Object Oriented Software System. The manuscript proposes metrics for measuring cognitive complexity of class due to its attributes. The proposed metrics takes into consideration the complexity introduced by data types of attributes that forms data elements of a class. The primitive, system defined and user-defined data types, used for defining the attributes have been weighted to measure the cognitive complexity of a class. Also an empirical study has been performed to gain insight on the correlation between the proposed measure and the understandability of the program. The results show the significance of measuring contribution of attributes towards cognitive complexity of a class. A metrics meant to measure cognitive complexity of a class should consider weighted measure of complexity introduced by different attributes of a class.