Although high-quality software is key for a safe society, beginners still struggle to understand programming basics, whereas experienced programmers consider testing code as an unnecessary, boring chore. The situation will improve when, instead of asking students to write hypothetical code or test cases, one challenges them to find and fix bugs deliberately injected in real, executable code. Since 2020, we have been developing educational artifacts needed to test this radically new hypothesis, including a web-based bug-hunting game for standalone code, and small-scale cyber-physical systems (a self-driving car and a smart home), controlled by fault-seeded embedded software. We deployed the proposed approach and artifacts in the following settings: a software testing course for 300 computer science students, an exam for 250 economics students enrolled in an introductory Python course, and a software testing refresher workshop for 80 alumni of an IT retraining programme. Evaluations based on surveys with Likert-scale and open questions showed that a fault-based, active-learning approach (1) increases excitement in learning and (2) offers a more adequate way to assess students' skills, compared to the traditional pen-and-paper or MC exams. Future work includes equalizing the difficulty level of injected bugs and adding other gaming elements, such as badges, scoreboards, automated grading and adaptive feedback. While we realize that more rigorous qualitative and quantitative evaluations will be needed to confirm the generalizability of our approach, we are confident in its potential to contribute to making future professionals better prepared to engineer the high-quality, safe software we all can rely on.

Especially in the case of large cohorts, it is difficult for teachers to offer interesting research topics for each student's graduation project. As a team of CS students and educators, we addressed this challenge with the VU-Rover, a newly designed, small-scale autonomous vehicle that features a Linux-based single-board computer and widely extensible sensing and actuation capabilities. The Rover is pre-installed with an integrated software framework that allows students to use any programming language. Proving its capabilities after winning the NXP international student race competition in 2024, the design was replicated to multiple identical Rovers and deployed in an innovative undergraduate CS capstone project focused on autonomous driving. In its first offering, 13 undergraduate students used the VU-Rover to research topics related to perception, positioning, path planning, safety analysis, and traffic jams. We will present details on the design of the initiative, coordinated by the Autonomous Systems Engineering (ASE) group, where students and teachers work together to gather knowledge and transfer it across generations. We will also shortly zoom into one of the graduation projects and share our lessons learned. Future work includes involving more supervisors from various domains and expanding the range of research topics towards security, operating systems, green IT, and IoT.

Teaching software systems engineering is neither effective, nor inspiring if students cannot practice the conveyed theory. We report on a hands-on approach that closes the gap by means of off-the-shelf and in-house realistic miniature models of autonomous vehicles. It is innovative that the infrastructure extends beyond the widely-used solutions, to digitally controlled cars and trains, with open hardware and extendable sensory/actuating capabilities (cameras, opto-sensors, proximity sensors, LEDs, displays) operating in realistic mini-environments, including line-marked roads, railways, traffic-lights and -signs, tunnels and railroad switches. For many years already, we have been using these vehicles to structurally teach physical computing (300 undergraduates yearly) and systems testing (12 graduates yearly), and for incidental individual research projects. Recently, we started two initiatives that stimulate a more dynamic know-how transfer across generations. The first is a better scalable undergraduate capstone project, where groups of students work on a 'hot' research topic in autonomous driving. The second motivates student teams to go the extra mile in an international intelligent car race competition. Evaluations showed that although unusual for a non-engineering curriculum, 'playing' with autonomous vehicles is an excellent strategy to discover the interaction of software with hardware and environment, to better consolidate existing knowledge on programming and testing, and to explore new fields, such as AI-based computer vision, navigation and safety, adding in all cases more fun and motivation. We share the design of the scaffolding and teaching initiatives, together with lessons learned which will hopefully inspire other educators in shaping engaging and future-proof CS curricula.

Although software testing is key to a safe society, the process itself is often perceived by students as boring and stressful. Therefore, only few consider a career in testing. The adverse effect is sub-optimally tested code, with dangerous bugs left undetected. A better understanding of what testers "feel" when learning the skill in class can remedy this situation, by means of personalized, motivating bio-feedback. In order to test our hypothesis, we propose an innovative approach that uses physiological wearable sensors (cardiac activity, respiration, and skin conductance) to monitor in real-time the affective state of testers engaged in a bug-hunting game. This is a work in progress. We present the envisioned methodology and the results of two feasibility experiments. The first experiment created a training dataset, by recording bio-signals and self-reports from eleven participants involved in a mood-induction session followed by a bug-hunting task. The second experiment showed that it is possible to use deep-learning to recognize emotions from a large set of labelled multimodal (ECG, EDA and ICG) physiological data. The classification accuracy using a binary (positive-negative) emotions model was 85%, higher than the accuracy obtained using a four-emotions (anxious, down, enthusiastic and relaxed) model (57%). Future work includes optimizing the sensory system, improving the accuracy of automated emotions recognition, increasing the validity of ground-truth emotions labelling, and investigating ways to provide individualized and formative (instead of summative) bio-feedback. The proposed approach can contribute to a more sentiment-aware education, and a more objective evaluation of the effect of teaching interventions.