Assessing quantitatively a programming course

Abstract

The focus on assessment and measurement represents the main distinction between programming course and software engineering courses in computer curricula. We introduced testing as an essential asset of a programming course. It allows precise measurement of the achievements of the students and allows an objective assessment of the teaching itself. We measured the size and evolution of the programs developed by the students and correlated these metrics with the grades. We plan to collect progressively a large baseline. We compared the productivity and defect density of the program developed by the students during the exam to industrial data and similar academic experiences. We found that the productivity of our students is very high even compared to industrial settings. Our defect density (before rework) is higher than the industrial, which includes rework.