Developing A Knowledge Based System To Support Debugging Computer Program Source Code

Abstract

To write programming is an essential skill for students of computing and informatics. However, learning programming skills in finding errors and correcting them to make the program run has been recognized as a great challenge for novice programmers. Most of the errors currently thrown by the compiler don‟t automatically point the novice programmers to the right direction since some of the messages need deep understanding and expertise. This finally leads students to suffer and discouraged hope of solving the problem on their own which results in programming phobia. Even worse, repetitive failures may defeat students‟ enthusiasm for learning. This study therefore investigates a number of different compilers, logical and run time errors that novice programmer‟s encounter and the associated debugging behaviors to assists them in writing error free code so that the primary objective of the study is to develop a knowledge based system to assist novice programmers in debugging computer program source code. This research has followed a Knowledge engineering process for knowledge acquisition, modeling, representation and prototype development and testing, The conceptual model of the knowledge based system is designed by using a decision tree structure which is easy to understand and interpret the causes involved in the program error. Based on the conceptual model, the knowledge is represented using „if – then„rules. The developed prototype infers the rules by backward chaining and provides appropriate suggestions as per the users query. The prototype was evaluated for its usability in which it registers 87.27% user‟s acceptance. In addition, the performance of the system was evaluated with twenty five test cases. The results of the validation test indicate that the prototype registers on the average of 80% accuracy. Most of the errors handled in this study are compile time errors. Our observation shows that there is a great challenge in understanding logical errors in writing source codes which left as a future research direction