Automatic Analysis and Evaluation of Student Source Codes

摘要

An essential part of education is the evaluation of students and the facilitation of students' studies through feedback. During evaluation and feedback, it is important to ensure that it applies equally to all students, is fair and objective. For simple (e.g., multiple choice) tests, their evaluation can be easily automated; however, source code evaluation is still done manually and often incorrectly. In our article, we present the model we used evaluating solutions for OE-NIK SzTF2 student homework written in C#. The program based on this model is able to both syntactically examine the internal structure of the code and to run the code and compare the input-output pairs. In addition, different metrics are defined for each source code examined, based on which instructors can make additional suggestions to students to improve their programming skills. The consequence is that with the help of the proposed system it is possible not only to decide whether a code is correct or not, but also due to a deeper analysis, further curious details are revealed.