Date

2014

Document Type

Dissertation

Degree

Doctor of Philosophy

Department

Computer Science

First Adviser

Blank, Glenn D.

Other advisers/committee members

Heflin, Jeff; Columba, H. Lynn

Abstract

Debugging is a necessary aspect of computer science that can be difficult for novices and experienced programmers alike. This skill is mainly self-taught and is generally gained through trial and error, perhaps with some assistance from a professor or other expert figure. Novices encountering their first software defects may have few avenues open to them depending on the environment in which they are learning to program. The evident problem here is that the potential for a student to become stuck, frustrated, and/or losing confidence in their ability to pursue computer science is great. For a student to be successful when working professionally or progressing through academia they need to be able to function independently; trusting their own knowledge on par or above that of others so that their productivity does not rely on the knowledge of someone else. In order to solve this problem an Intelligent Tutoring System for teaching debugging skills to the novice utilizing Case Based Reasoning, Static Program Slicing, and the student's preferred learning style was proposed. Case acquisition and automatic Exercise Generation were also explored. The system built for this research program was evaluated using novice students at the College and High School levels. Results of this evaluation produced statistically significant results at the p<.05 and p<.01 levels, with generated exercises exhibiting significance at the p<.01 level. These results prove that the methodology chosen is a valid approach for the problem described, that the system does in fact teach students how to debug programs, and that the system is capable of successfully generating exercises on the fly.

Share

COinS