Experiments for Linking the Complexity of the Business UML Class Diagram to the Quality of the Associated Code

• Published in: Computational Science and Its Applications – ICCSA 2021 pp. 95 - 110
• Main Authors: Paolone, Gaetanino, Marinelli, Martina, Paesani, Romolo, Di Felice, Paolino
• Format: Book Chapter
• Language: English
• Published: Cham Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Association complexity; Attribute complexity; Class complexity; Code quality; Complexity metrics; Method complexity; UML Class diagram
• ISBN: 9783030869694; 3030869695
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-030-86970-0_8

A relevant goal of software engineering is to assure the quality of object oriented software from the conceptual modeling phase. UML Class diagrams constitute a key artifact at that stage. Among existing models for iterative and incremental development of software systems, Model Driven Architecture (MDA) has reached a leadership position. MDA enables model-driven software development which treats models as primary development artifacts. The present empirical study answers the following three research questions: (RQ1) are there available in the literature class complexity metrics that can be adopted at the business level? (RQ2) Is it possible to adopt those metrics (if any) to predict the quality of the code returned by xGenerator? The latter is a Java technology platform for the creation of MVC Web applications, which implements the model-driven approach. (RQ3) Is it possible to identify a threshold for the adopted metrics (if any) that might suggest when a business Class diagram should be refactored?