UCL - SST/ICTM/INGI - Pôle en ingénierie informatique, Lozano Rodriguez, Angela, Mens, Kim, Portugal, Jawira, International Conference on Software Analysis, Evolution and Reengineering, UCL - SST/ICTM/INGI - Pôle en ingénierie informatique, Lozano Rodriguez, Angela, Mens, Kim, Portugal, Jawira, and International Conference on Software Analysis, Evolution and Reengineering
This paper reports on evidence found of five possible relations (Plain Support, Mutual Support, Rejection, Common Refactoring, and Inclusion) among four bad smells (God Class, Long Method, Feature Envy, and Type Checking). We analyzed several releases of three open-source applications (16 for Log4j, 34 for Jmol, and 45 for JFreeChart) using four direct and two indirect metrics. This analysis uncovered correlations between three of these bad smells, namely, Feature Envy, Long Method, and God Class. The strongest correlation discovered was between Feature Envy and Long Method, followed by a mild correlation between Long Method and God Class, and between Feature Envy and God Class. These findings seem to provide initial evidence of the co-existence of bad smells and therefore, the need for bad smell removal plans to take into account these correlations in order to minimize code improvement efforts.