A Pedagogical Evaluation and Discussion about the Lack of Cohesion in Method (LCOM) Metric Using Field Experiment

TL;DR

This study evaluates the effectiveness of the LCOM metric in measuring class cohesion across three industrial systems, revealing varying levels of cohesion and supporting its use for class design assessment.

Contribution

It provides a pedagogical analysis of LCOM's appropriateness in real-world systems using empirical data from three different industrial software systems.

Findings

System 1: 78.8% classes were cohesive

System 2: 54% classes were cohesive

System 3: 30% classes were cohesive

Abstract

Chidamber and Kemerer first defined a cohesion measure for object-oriented software - the Lack of Cohesion in Methods (LCOM) metric. This paper presents a pedagogic evaluation and discussion about the LCOM metric using field data from three industrial systems. System 1 has 34 classes, System 2 has 383 classes and System 3 has 1055 classes. The main objectives of the study were to determine if the LCOM metric was appropriate in the measurement of class cohesion and the determination of properly and improperly designed classes in the studied systems. Chidamber and Kemerer's suite of metric was used as metric tool. Descriptive statistics was used to analyze results. The result of the study showed that in System 1, 78.8% (26 classes) were cohesive; System 2 54% (207 classes) were cohesive; System 3 30% (317 classes) were cohesive. We suggest that the LCOM metric measures class cohesiveness and was appropriate in the determination of properly and improperly designed classes in the studied system.