Designing an Automatic Debugging Assistant for Improving the Learning of Computer Programming

Abstract

Finding bugs in programs (debugging) is a core skill for practical programmers. However, debugging programs can be difficult to novice programmers. Even worse, repetitive failures may defeat students' enthusiasm for learning. The presence of a mentor giving hints and help face-to-face with students will surely make such a learning process much more effective and enjoyable. However, this requires lots of manpower and resources. To address this problem, we seek to capitalize on the potential advantages offered by hybrid learning. We are working towards a system for providing a certain level of automatic debugging assistance to students. Instructors can identify common errors in students' programs using the system and incorporate useful debug-guiding information into it so that students will be prompted with pertinent hints when common errors are detected in their programs.