Complexity measures for object-oriented program based on the entropy

Abstract

f i e software metrics are widely advocated as a fundamental elements of an engineering approach to planning and controlling software developnent. n e y are o f especial importance with the objectoriented program. In this paper, we propose the new software cmplexity model which consists of the class cmplexity, the inter-object cmplexity, and the total cmplexity for the objectoriented progran. We measure the cmplexities based on the entropy concept. The class cmplexity for a class measures the information flows in a class based on the information passing relationship an-mmber data and member functions. The inter-object cmplexity for a program measures the information flows between objects. The total cmplexity for a program i s measured by the class canplexity and the inter-object cmplexity. W e evaluate the proposed metrics by the cmplexity properties proposed by Weyuh-er. Experimental results of C++ classes show the effectiveness o f the proposed metrics. W e assert that the class canplexity and the inter-object cmplexity are considerately correlated to other metrics for object-oriented concepts.

Improvements to a complexity metric: CB measure

“…As a result, over the past several years, various complexity metrics have been introduced based on various attributes of software, such as control flow [3], operator and operand count [7], irlformation flow [11], data access [16], data flow [20], branching complexity [30], logical complexity [21], spatial complexity [22], [23], object orientation [12], [1], entropy [24], [25], category learning [10], and cognitive complexity [27], [28], [29] with the intention of controlling and minimizirlg the complexity associated with software.…”

Performance based analysis on MALCOM — A software metric

“…Weyuker's suggested these attributes to evaluate the complexity measure when only traditional programming languages were in use. These properties were also used by some researchers for evaluation of popular object oriented metrics for example Chidamber's metrics [2] and Kapsu's metric [5], although the object-oriented features are entirely different in nature. It is not mandatory that all types of metrics satisfy all the weyuker's properties.…”