Empirical evaluation of code smells in open-source software (OSS) using Best Worst Method (BWM) and TOPSIS approach. Issue 3 (24th August 2021)

• Record Type: Journal Article
• Authors: Tandon, Stuti; Kumar, Vijay; Singh, V.B.
• Is Part Of: International journal of quality & reliability management. Volume 39:Issue 3(2022)
• Page Start: 815
• Page End: 835
• Publication Date: 2021-08-24
• Subjects: Software code smells -- Multi-criteria decision-making (MCDM) -- Best Worst Method (BWM) -- TOPSIS -- Apache Tomcat; Quality control -- Periodicals; 658.562
• DOI: 10.1108/IJQRM-02-2021-0045
• Languages: English
• ISSNs: 0265-671X

Abstract

Abstract : Purpose: Code smells indicate deep software issues. They have been studied by researchers with different perspectives. The need to study code smells was felt from the perspective of software industry. The authors aim to evaluate the code smells on the basis of their scope of impact on widely used open-source software (OSS) projects. Design/methodology/approach: The authors have proposed a methodology to identify and rank the smells in the source code of 16 versions of Apache Tomcat Software. Further, the authors have analyzed the categorized smells by calculating the weight of the smells using constant weights as well as Best Worst Method (BWM). Consequently, the authors have used Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) to determine the rank of versions using constant weights as well as BWM. Findings: Version 1 of Apache Tomcat has least smell, and version 8 is reported to contain the maximum code smells. Notable differences in both the cases during the trend analysis are reported by the study. The findings also show that increase is observed in the number of code smells with the release of newer versions. This increment is observed till version 8, followed by a subtle marked depreciation in the number of code smells in further releases. Originality/value: The focus is to analyze smells and rank several versions of Apache Tomcat, one of the most widely used software for code smell study. This study will act as a significant one for the researchers as it prioritizes the versions and will help in narrowing down the options of the software used to study code smell.