APPLYING AUTOMATIC CODE GRADING SYSTEM FOR TEACHING PROGRAMMING LANGUAGES IN UNIVERSITIES

Abstract

The rapid growth of online education platforms highlights the need for effective automatic grading systems, especially for programming courses in universities. This paper introduces an Automatic Code Grading System named AutoChecking, tailored for evaluating code submissions in various programming languages at the university level. AutoChecking combines static code analysis and dynamic execution to ensure thorough and equitable assessment of students' programming assignments. It features a sandbox environment for securely running student codes against predefined test cases, evaluating not just accuracy but also efficiency and coding style. The system is adaptable, with a flexible scoring mechanism to meet different course and instructor needs. AutoChecking integrates with existing online Learning Management Systems to streamline submission, grading, and feedback, making it easier for instructors to manage assignments and monitor student progress with detailed analytics. We evaluated the effectiveness of the AutoChecking through a semester-long deployment in introductory and advanced programming courses at three universities. The results showed that AutoChecking saves instructors' time, provides consistent feedback to students, enhancing their learning, and plays a crucial role in maintaining academic integrity by detecting code similarities.