An Empirical Study on Measuring the Impact of Code Smells on Web Applications

Abstract

Code smells are a surface symptom that generally indicates a more serious problem in the software system therefore it is kind of hard to detect them manually so many automated tools are required to get the job done. The study focuses on proposing an empirical study that has been applied on web applications in order to detect the large classes manually based on Michele and Marinescu large classes’ detection strategy as shown in Figure 1. The result of the study was the failure of the first project group due to the inability and agreement to identify large classes, while the second project group succeeded, as the participants agreed to define large classes more precisely, in addition to that the study showed that the process of manually identifying large classes is a very difficult process therefore an automated tools (AI tools) are required to make it easier and more accurate.