A Survey Paper for Bug Localization

Bug localization is the task to locate the source code entities which are relevant from the bug report. Manual bug localization is a time consuming and labor consuming task as developers has to go through thousands of source code entities to locate the relevant one. Current research provides methods as various IR techniques, classifiers, combination of classifiers to improve bug localization. To make bug localization partial or maximum automated some tools are available but mostly they are based on simple query, IR techniques or eliminating unnecessary stack traces or use basis of previous bug reports and its changed files and based on that history relevant source files will be identified. There are automated path generation for software fault localization. But none of all these tools/techniques achieves highest efficiency as they work only on one area or combination of areas. To make automation more successful and efficient there is a need for finding hybrid approaches. This paper provides the literature survey about what is done regarding bug localization so far and what is the future scope for automation in bug localization. It also points how efficient automated bug localization may help maintain the software cost.