An Empirical Analysis of Software Systems for Measurement of Design Quality Level Based on Design Patterns

Author(s): Abul Khaer, M.M.A. Hashem, and Raihan Masud
Venue: Computer and information technology, iccit 10th international conference on...
Date: 2007

Type of Experiement: Case Study



This paper proposes a new simple approach to specify the design level of object oriented software systems. The analysis method of this paper uses GoF (Gang of Four) design patters as the assessment criteria. The method is then empirically studied using several open source project and found to be a good alternative to OO metric, software fault proneness, visualization and anti-pattern based approaches. The empirical study searched open source projects and counted the number of design patterns found per ten classes (this was to minimize the effects of large classes vs small classes). Their hypothesis was confirmed that the higher the design patterns the better the code is. The quality of the project is then computed and compared to the design quality that would be found by using Weighted Method per Class (WMC), Coupling between Objects (CBO) and lines of code (LOC). The results found that the quality found using design patterns matched closely to the results found using the standard OO metrics. Therefore the authors propose combining the standard metrics with a design pattern counting mechanism.