Heuristics for automatic localization of software faults

Developing effective debugging strategies to guarantee the reliability of software is important. By analyzing the debugging process used by experienced programmers, four distinct tasks are found to be consistently performed: (1) determining statements involved in program failures, (2) selecting suspicious statements that might contain faults, (3) making hypotheses about suspicious faults (variables and locations), and (4) restoring program state to a specific statement for verification. If all four tasks could be performed with direct assistance from a debugging tool, the debugging effort would become much easier. We have built a prototype debugging tool, Spyder, to assist users in conducting the first and last tasks. Spyder executes the first task by using dynamic program slicing and the fourth task by backward execution. This research focuses on the second task, reducing the search domain containing faults, referred to as fault localization. Several heuristics are presented here based on dynamic program slices and information obtained from testing. A family tree of the heuristics is constructed to study effective application of the heuristics. The relationships among the heuristics and the potential order of using them are also explored. A preliminar