Study on Simplified Dynamic Slicing Algorithm based on Input Classification

Software failures have gradually become a serious problem in information systems. As a result, diagnosing software faults and maintaining software becomes more important and normalized. There is an urgent need for effective methods to locate faults and maintain software. The paper presents a simplified dynamic slicing algorithm based on input classification. Building the program dependency graph on definitions of control dependency and data dependency, the new algorithm decreases the conditional control sentences by input classification, which reduces the space of the traditional slicing algorithm and gets fewer sentences after slicing. A variety of dynamic slicing experiments of the typical program show that the proposed algorithm is higher efficiency and can get more accurate results, which provides an effective means for quickly locating faults.