ReCodEx: Universal Automated Assessment for Modern CS Education
Abstract
Over the past two decades, automated assessment systems have become integral to programming courses and coding competitions, enabling scalability, reducing teachers' workload, and providing faster feedback to students. In this report, we present our experience with ReCodEx, an open-source system for automated evaluation developed and extensively used at our institution for 20 years. Unlike many similar systems, ReCodEx is designed with an emphasis on flexibility and versatility to reach far beyond the CS1 level. It supports 42 courses, including advanced and specialized topics like machine learning, web application development, and compiler principles, while processing >170k submissions annually. We detail how ReCodEx evaluates assignments in these diverse contexts, summarize insights from interviews with lecturers, and share our lessons learned from its development and utilization. Our findings highlight the system's strengths, challenges, and transformative impact on teaching and learning in higher education.