Using Stack Overflow to Assess Technical Debt Identification on Software Projects

  • Eliakim Gama UECE
  • Sávio Freire IFCE / UFBA
  • Manoel Mendonça UFBA
  • Rodrigo O. Spínola UNIFACS / UNEB
  • Matheus Paixao UNIFOR
  • Mariela I. Cortés UECE


Context. The accumulation of technical debt (TD) items can lead to risks in software projects, such a gradual decrease in product quality, difficulties in their maintenance, and ultimately the cancellation of the project. To mitigate these risks, developers need means to identify TD items, which enable better documentation and improvements in TD management. Recent literature has proposed different indicator-based strategies for TD identification. However, there is limited empirical evidence to support that developers use these indicators to identify TD in practice. In this context, data from Q&A websites, such as Stack Overflow (SO), have been extensively leveraged in recent studies to investigate software engineering practices from a developers' point of view. Goal. This paper seeks to investigate, from the point of view of practitioners, how developers commonly identify TD items in their projects. Method. We mined, curated, and selected a total of 140 TD-related discussions on SO, from which we performed both quantitative and qualitative analyses. Results. We found that SO's practitioners commonly discuss TD identification, revealing 29 different low-level indicators for recognizing TD items on code, infrastructure, architecture, and tests. We grouped low-level indicators based on their themes, producing an aggregated set of 13 distinct high-level indicators. We then classified all low- and high-level indicators into three different categories according to which type of debt each of them is meant to identify. Conclusions. We organize the empirical evidence on the low- and high-level indicators and their relationship to types of TD in a conceptual framework, which may assist developers and serve as guidance for future research, shedding new light on TD identification state-of-practice.
Palavras-chave: Indicators, Technical Debt, Stack Overflow, Mining Software Repositories
Como Citar

Selecione um Formato
GAMA, Eliakim; FREIRE, Sávio; MENDONÇA, Manoel; SPÍNOLA, Rodrigo O.; PAIXAO, Matheus; CORTÉS, Mariela I.. Using Stack Overflow to Assess Technical Debt Identification on Software Projects. In: SIMPÓSIO BRASILEIRO DE ENGENHARIA DE SOFTWARE (SBES), 34. , 2020, Natal. Anais [...]. Porto Alegre: Sociedade Brasileira de Computação, 2020 .