Simplifying Automated Assessment in CS1

Abstract

Automated assessment in CS programming courses offers timely feedback, objectivity, and consistency of the evaluation. Although it facilitates the evaluation of the assignments, creating the tests to validate the program functionality still takes a considerable amount of time. The instructor has to devise a strategy for testing and calculate the result that a correct program is expected to produce. To address these difficulties, we created a system that instruments the student code to capture the program states and verify the program’s correctness by comparing them with a reference solution. Our approach reduces the required code for testing substantially.