A programming learning diagnostic system using case-based reasoning method

Abstract

The knowledge of programming consists of conceptual knowledge and strategic knowledge. Lack of problem-solving strategies is one of the major reasons why code learning is hard for many novice programmers. In this paper, we not only construct these two types of knowledge, but also construct the relationship of learning cases and e-portfolio for diagnosing the learning difficulties of each learner. Based on these constructed knowledge, we develop a learning diagnostic system using case-based reasoning method to give the learners feedback and suggestion in real time. We choose C as a first programming language for the beginners. The system is built on 3-tier client/server architecture and can run on various platforms such as Windows, Linux, OS X, and Android. There are two learning diagnostic stages to indicate novice programming mistakes about the conceptual programming knowledge and strategic programming knowledge, to help learners recognize, diagnose, and recover from programming errors.