Spectrum-Based Bug Localization using Statement Weight

Abstract

Locating bugs in a program is generally a complex and time-consuming process in software engineering. Many spectrum-based approaches have been proposed to assist programmers to identify locations of program bugs. This thesis presents a spectrum-based bug localization using statement weight. An average weight of each statement is computed to determine the suspiciousness of the statements. Statements with higher suspiciousness are more likely to contain bugs and should be examined before those with lower suspiciousness. The proposed statement weight method is compared with Tarantula on one of the Siemens suites. The experimental results show that the propsed approach is more effective than Tarantula in average score. We also present the architecture of an Eclipse plug-in named JBugLight that helps programmers locate bugs by reporting the sorted list of statements according to statement weight.