Revertに着目した不確かさに関する実証的分析

Translated title of the contribution: Empirical Study of Uncertainty Focusing on Revert

村岡 北斗, 鵜林 尚靖, 亀井 靖高, 佐藤 亮介

Research output: Contribution to journalArticlepeer-review

Abstract

Uncertainty is a problem that occurs during software development due to lack of the developer's knowledge. In this paper, we investigate two research questions: RQ1) How is uncertainty related to revert, RQ2) What kinds of category is uncertainty divided into? The result of the study is as follows. For RQ1, since the cause of the bug is uncertain in revert commits, there are many cases where revert is performed as a response. In addition, in the case of reverted commits, there are many cases where a commit that has been provisionally processed is later reverted. On the other hand, for RQ2, uncertainty in OSS projects can be divided into the following five categories: uncertainty in the content of an exception, uncertainty associated with bugs, uncertainty about future changes, uncertainty related to changes, and uncertainty about the code. In addition, we have found that uncertainty related to bugs tends to appear frequently. Uncertainty to be fixed with high priority is uncertainty in which the cause of the bug is unclear or uncertainty dealt with exceptionally.
Translated title of the contributionEmpirical Study of Uncertainty Focusing on Revert
Original languageJapanese
Pages (from-to)918-929
Number of pages12
Journal情報処理学会論文誌
Volume61
Issue number4
Publication statusPublished - Apr 15 2020

Fingerprint Dive into the research topics of 'Empirical Study of Uncertainty Focusing on Revert'. Together they form a unique fingerprint.

Cite this