Demystifying Sonar Tool Estimates in the Contexts of Familiar and Unfamiliar Software Projects: An Empirical Study with Junior Developers
Andrej Katin, Nebojša Taušan, Valentina Lenarduzzi, Vladimir Mandić · 2023
Developing a high-quality software product requires complete familiarity with the software product requirements and constraints.However, developers are often assigned to projects they are unfamiliar with, making project properties hard to understand for them.Sonar tool is one of the most popular Automated Static Analysis Tools that nowadays has become an integral part of software development process.The objective of this study is to evaluate the usefulness of the Sonar tool in the context of familiar and unfamiliar projects.More specifically to investigate the accuracy of Sonar's estimates while the remaining three assess the perceived refactoring difficulties, Sonar issue and fixing instructions as possible reasons for differences in estimates.A repeated measurements experiment in the context of a university course was designed and involved two sessions during which the subject refactored Sonar issues identified in familiar and unfamiliar projects.In total 60 students completed assignments, who in two sessions combined refactored 5,179 Sonar issues by violating 138 different Sonar rules.This study showed that estimates provided by Sonar are less accurate for familiar projects.Plausible explanations are that the refactoring and understandability of Sonar tool issue description seem to be more challenging in unfamiliar, compared to familiar projects.