In this report, we survey different automatic grading systems in the context of programming courses. We seek to offload repetitive grading tasks to these systems so that teachers may instead focus on providing high-quality formative feedback. To evalu- ate which tasks automatic grading systems are suitable for, we categorize them by which SOLO-level they attain. Our findings are promising and suggest that the prior work pro- vide automatic grading methods up to the relational level. Finally, by employing the methodology of letting students design their own test cases, automatic grading can be used to reach the extended abstract level.
You may also start an advanced similarity search for this article.