An Empirical Study on the Urgent Self-admitted Technical Debt

Technical Debt (TD) refers to the phenomenon of taking shortcuts to achieve short-term gains at the cost of higher maintenance effort in the future. While Self-Admitted Technical Debt (SATD) is the TD that has been documented in code comments. Due to its impact on software quality, SATD that is of more long-term interest should be prioritized for removal. However, most current research infers the priority of SATD removal from self-defined features, and lacks a systematic understanding of it. In this paper, we define “urgent SATD”, as the SATD that have a high priority for removal. We conduct an empirical study on nearly 6500 SATD in two popular open-source projects to examine the distribution and features of “urgent SATD” with other SATD. We found that only a small fraction of removed SATD is urgent SATD (4%–16%), and urgent SATD has significant differences from other SATD in terms of text features and code size features, indicating that it has more complex text descriptions and related-code.