On the Presence and Causes of Lingering Defects in Software : An industrial study of lingering defects

Abstract: Background. Remaining defects are a cause of significant issues in software development industry and not fixing them as soon as possible increases the risk of negative effects such as system crashes, customer dissatisfaction, and higher costs. Despite these observations, companies do, due to varying factors, not always address newly found defects. These factors, which lead to lingering defects, are not well studied and therefore warrant more research. Objectives. To identify factors that companies consider when deciding to allow defects to remain and linger. As a secondary objective, the study aims to discover how many and what types of defects that companies leave unaddressed. Methods. A two-step methodology is applied where a snowballing-based literature review is used to acquire an understanding of the research area to identify the research gap. These results are then complemented by an empirical and industrial case study with 16 interviews. The interviews were analyzed with an approach inspired by thematic coding, leading the study’s main conclusions. Results. Results of the study identifies 11 factors that developers consider when deciding to address a known software defect. Results also show that the impact of lingering defects vary but is generally low and that certain types of defects, e.g. non-functional defects, are more likely to linger. Conclusions. The study provides insights to industrial practitioners with factors to consider for deciding about if a defect shall linger or not. The study also provides evidence, contrary to the body of knowledge, that lingering defects may not be a large problem in practice. However, due to the size of the study and its interview sample, future research is required.