A Fault Localization Technique for Online Programming Learning
Wei Zheng, Jinsheng Chen, Zhiwei Lu, Fengyu Yang, Peng Xiao, Xin Fan · 2023
With the development of Internet technology, online programming learning is becoming increasingly popular. Program debugging is an important step in the learning process of programming. Learners can usually only locate faults through manual troubleshooting, which is inefficient. Spectrum-based fault localization (SBFL) methods can help developers locate faults, but SBFL methods are typically used for medium-sized or large programs, with low accuracy on small programs. Due to the particularity of online programming learning, programs are relatively short and their functions are relatively simple. At the same time, the history programs of learners and the answer scripts of other people to the same question will be recorded, and these data can help to locate faults. This paper proposes a fault localization technique named FLSISS (Fault Localization based on Spectrum Information of Similar Scripts) for online programming learning. Firstly, the technique uses the suspiciousness value calculation formula of the typical SBFL to initially calculate and sort the suspiciousness value of each statement node of the script, then improves the suspiciousness value and ranking of each statement node by integrating the spectrum information of similar answer scripts, thereby the accuracy of fault localization is improved. During the process of online programming learning, compared with the traditional SBFL methods, FLSISS can locate program faults more effectively and improve the learning efficiency of learners.