Year: 2025
Type: conference-paper
Abstract: Candli is a web-based educational game creation app in which children acquire STEAM skills by creating video games based on their own drawings. Candli allows children to program games using an accessible visual programming language inspired by established research traditions such as Scratch and Thymio VPL. While such languages are highly expressive, enabling rapid creation of dynamic and engaging agent behaviors, this very efficiency can pose challenges in classroom settings. In particular, children often encounter non-trivial problems quickly, leading to teacher overload and hindering learning. To mitigate this issue, Candli integrates technical scaffolding features designed to support children’s autonomous learning. In this workshop, we explore two such features: linting and adaptive program completion hints. The linting feature draws inspiration from software engineering practices, providing immediate feedback to students about potential errors in their code – a feature rarely utilized in educational contexts. Adaptive program completion hints employ classical AI reasoning to incrementally guide students based on their existing code structure. This method involves comparing the current student code with predefined Prolog-like solution rules and suggesting minimal, targeted corrections that do not compromise other program functionalities. These features assist learners in developing accurate notional machines for understanding code execution while preserving opportunities for free exploration and playful tinkering. Collectively, these tools offer a concrete realization of constructionist learning principles.
Author: Stéphane Magnenat
Language: English
Subfield: Computer Science Applications
Field: Computer Science
Domain: Physical Sciences
SDG: Quality education
Open Access status: hybrid
OpenAlex ID: W4411614433