Grading student programs - a software testing approach

Abstract

This paper describes an experience of automating the grading of student programs. A testing framework provides guidance for developing the assignment specification and the grading program. Automation saves time and improves grading consistency and feedback to students. After an adjustment period, student programs improved. Although the instructor invests more time writing a testable assignment specification and developing the grading program, these costs are expected to be amortized over multiple courses and assignments.