Abstract
We present a methodology to identify refactoring operations that reduce the bug rate in the code. The methodology is based on comparing the bug fixing rate in certain time windows before and after the refactoring. We analyzed 61, 331 refactor commits from 1, 531 large active GitHub projects. When comparing three-month windows, the bug rate is substantially reduced in 17% of the files of analyzed refactors, compared to 12% of the files in random commits. Within this group, implementing 'todo's provides the most benefits. Certain operations like reuse, upgrade, and using enum and namespaces are also especially beneficial.
| Original language | English |
|---|---|
| Title of host publication | PROMISE 2019 - 15th International Conference on Predictive Models and Data Analytics in Software Engineering |
| Publisher | Association for Computing Machinery |
| Pages | 12-15 |
| Number of pages | 4 |
| ISBN (Electronic) | 9781450372336 |
| DOIs | |
| State | Published - 18 Sep 2019 |
| Event | 15th International Conference on Predictive Models and Data Analytics in Software Engineering, PROMISE 2019, co-located with the 13th International Symposium on Empirical Software Engineering and Measurement, ESEM 2019 - Recife, Brazil Duration: 18 Sep 2019 → … |
Publication series
| Name | ACM International Conference Proceeding Series |
|---|
Conference
| Conference | 15th International Conference on Predictive Models and Data Analytics in Software Engineering, PROMISE 2019, co-located with the 13th International Symposium on Empirical Software Engineering and Measurement, ESEM 2019 |
|---|---|
| Country/Territory | Brazil |
| City | Recife |
| Period | 18/09/19 → … |
Bibliographical note
Publisher Copyright:© 2019 Copyright held by the owner/author(s).
Keywords
- Code quality
- Machine learning
- Refactoring