Program slicing spectrum-based software fault localization

Spectrum-based fault localization technique mainly utilizes testing coverage information to calculate the suspiciousness of each program element to find the faulty element. However, this technique does not fully take consideration of dependences between program elements , thus its capacity for efficient fault localization is limited. This paper combines program slicing with program spectrum technique, and proposes a program slicing spectrumbased software fault localization (PSS-SFL) technique. Firstly, PSS-SFL analyzes dependences between program elements, and deletes some elements unrelated to the failed test outputs; then it builds the program slicing spectrum model and defines a novel suspiciousness metric for each slice element; finally, the faulty element is located according to the suspiciousness metric results. Experimental results show that PSS-SFL can be effective and more precise to locate the fault than program spectrum-based Tarantula technique. Keywords-Fault localization, program slicing spectrum, program slicing, program spectrum