Review of Open Software Bug Datasets

The localisation of the bug position in a source code and the prediction of which specific parts of a source code might be the cause of defects play an important role in maintaining software quality. Both approaches are based on applying information retrieval techniques and machine learning or deep learning methods. The prerequisite for using these approaches is the availability of a consistent bug dataset of sufficient size. This paper presents an overview of available public bug datasets and analyses their specific application areas. The paper also suggests possible future research directions in this field.