| Summary: | 碩士 === 元智大學 === 資訊管理學系 === 105 === Since the rapid development of software recently, more the source codes of software have been freely shared on the internet. Many people provide their effort to upgrade the software and to debug voluntarily for improving the software quality. Although bugs are inevitable in software development, the late fixed bugs decrease the software quality and downgrade user experience.
Most literature in bugs for the open software focused on how to identify proper developers for fixing certain bugs or how to allocate which files contain the bugs. Very few analyzes which bugs need more time to being fixed and turn to long-life bugs. The long-life bugs decrease the stability and quality of the software.
To reducing the long-life bugs, this search provides a method to predict which bugs might be long life at an early stage. To create prediction models for long-life bugs, we use bug reports for Mozilla products as our input data and test data. The analysis results of the models reveal that three major factors significantly influence the time of bug fixing, including the utilization of the patches, the bug reporter, and the bug developer. The proposed research methods can be applied to help developers to predict the long-time fixed bugs so that software developer can pay more attention to them on time.
|
|---|
