Theoretical Analysis of Entropy Distance Metric for UML Class Diagrams

摘要

Although the entropy distance metric for UML (Unified modeling language) class diagrams overcomes some limitations of previous metrics by reason that it considers relationships among classes, it ignores complexity of classes themselves. This paper deeply analyzes and evaluates the essence and behaviors of the metric from theoretic aspects. Furthermore, some problems are pointed out, such as the structure complexity of a class diagram cannot be computed or the result is incorrect, and the structure complexities of two class diagrams cannot be distinguished, and so on.