DOMtutor: Automated Autograding for Logic in Computer Science

The teaching of computer science at universities often relies on traditional, theory-oriented methods, characterized by 'transmission'-style lectures, manual grading of exercises, and delayed feedback. This approach is particularly prevalent in theoretical computer science subjects like logic. Automated autograding frameworks are identified as a potential solution to enhance teaching by providing near-instant feedback to students.

Why it matters

The adoption of automated feedback systems can significantly improve the efficiency and effectiveness of educational delivery, particularly in technical fields. This shift addresses limitations of traditional teaching methods by enabling scalable, timely student support and fostering practical skill development over purely conceptual understanding.

Key insights

• University computer science education is frequently structured classically, emphasizing theoretical concepts over practical application.
• Current teaching methods involve 'transmission'-style lectures and manual grading of exercises, resulting in delayed or absent feedback.
• Theoretical computer science subjects, such as propositional or first-order logic and automata theory, are particularly susceptible to conceptual, untestable exercise approaches.
• Autograding frameworks offer a means to automatically execute and evaluate code, providing near-instant feedback to students.

Source

arXiv — Computers and Society — https://arxiv.org/abs/2608.16899