An Empirical Study on Interaction Factors Influencing Bug Reopenings

Author

Jinkun Pan ; Xiaoguang Mao

Author_Institution

Coll. of Comput., Nat. Univ. of Defense Technol., Changsha, China

Volume

2

fYear

2014

Firstpage

39

Lastpage

42

Abstract

Bugs can be reopened after they have been closed due to identification of the actual cause, previous incorrect fixing, or better reproducing, etc. Reopened bugs may increase the cost in maintenance, degrade the overall quality of the software product, reduce the trust of users, and bring unnecessary work to the already-busy developers. To minimize the occurrence of bug reopenings, the potential causes and factors should be analyzed. In this paper, we explore 24 interaction factors to study their influence on bug reopenings. The data are extracted from Mylyn logs of four open-source projects. We first verify the negative impacts of bug reopenings. Then, we identify 17 factors that significantly influence the likelihood of bug reopenings using statistic tests. In addition, we build decision trees using interaction factors to predict bug reopenings and achieve good performance.

Keywords

decision trees; program debugging; program testing; public domain software; software quality; statistical analysis; Mylyn logs; decision trees; open-source projects; reopened bugs; software product quality; statistical testing; Accuracy; Computer bugs; Decision trees; Measurement; Open source software; Software engineering; Mylyn; bug reopenings; empirical software engineering; interaction factors; software maintenance;

fLanguage

English

Publisher

ieee

Conference_Titel

Software Engineering Conference (APSEC), 2014 21st Asia-Pacific

ISSN

1530-1362

Print_ISBN

978-1-4799-7425-2

Type

conf

DOI

10.1109/APSEC.2014.90

Filename

7091213