Automated Evaluation System for Software Development Assignments

As software and IT systems have increased, also their complexity grew rapidly and even more boosted through AI integration. This heightens the need for skilled software engineers to avoid reliance on opaque black-box or AI-driven solutions. However, rising student numbers in programming courses exceed teaching capacity, creating larger groups. Improving teaching efficiency without reducing quality is more important than ever. In this paper, we present an evaluation system that is used to generate fast and accurate feedback for student submission in, but not limited to, software development courses. It is the evolution and extension of the system called “EsPrESSO-IT”, which can be used to get a first feedback to outline the students how far they got in their implementation and where problems exist in their code. The prototypes allow the lecturers to return the tool's feedback to the students to give them the opportunity to analyze and correct their errors. Provided in a feedback loop also the students are supported in their learning process. We present the current implementation of the system and analyze its deployment in a software development lecture. The analysis was done by means of surveys of the lecturers and students as well as by a statistical analysis of the student submission results. The results show that lecturers and students made use of the automated feedback and even would prefer to have these features available instantly and in other engineering lectures as well.