A Fault Injection Method for Generating Error-correction Exercises in Algorithm Learning
Ryota Itoh, Hiroyuki Nagataki, Fukuhito Ooshita, Hirotsugu Kakugawa, Toshimitsu Masuzawa · Institutional Repositories DataBase (IRDB) · 2007
In this paper we propose a method for generating errorcorrection exercises for undergraduate students in computer science who learn algorithms. Our main goal is to inject faults automatically into a correct source code that implements an algorithm to be studied. The proposed method utilizes design paradigm of the algorithm to determine effective fault types and positions in a source code. We have developed a prototype system and evaluated the appropriateness of the generated exercises to algorithm study. We carried out error-correction exercises in an algorithm class, and most students evaluated that the exercises are effective for algorithm study.