From error to Learning: Dojo, a Platform to Improve Programming Teaching

Abstract

The teaching of programming faces particular pedagogical challenges, such as the need for continuous feedback, consistent practice, and a gradual understanding of errors. In numerous heterogeneous educational contexts, traditional manual correction methods are insufficient and not scalable.
To address these challenges, this article presents Dojo, an open source virtual platform designed to accompany the teaching of programming. This system integrates automated feedback mechanisms through industrygrade code analysis tools, which enable the detection of functional and stylistic errors in the students’ solution. Dojo not only identifies problems, but also generates contextualized explanations and provides specific learning resources, allowing students to receive accurate corrections early on. A specialized pedagogical repository complements these technical analyses, translating error messages into student-friendly language. From a didactic point of view, the direct visualization of errors and their correction co tribute to the self-regulation of learning, facilitating the progressive acquisition of debugging and code improvement skills, encouraging a constructive interpretation of the error.
Preliminary results from the platform’s implementation have been encouraging, showing benefits in terms of faster feedback cycles, giving greater autonomy to students and enhancing the educator’s role, allowing them to focus on advanced conceptual aspects.
Its modular architecture allows for future expansions, consolidating its potential as a transformative educational resource, transforming errors into valuable learning opportunities. Dojo represents a bridge between academic training and professional practices, offering a scalable model for teaching programming.