Making Robot Programming Visible through Solvable Tasks and Classroom Practice
Suyu Jiang, et al.
ICWL–SETE 2026 — Joint Conference of the 25th International Conference on Web-Based Learning and the 11th International Symposium on Emerging Technologies for Education
Abstract
A browser platform lets a beginner program a simulated hair-cutting arm with blocks, inspect the cut, and revise the program. The early tasks used geometric targets and had no demonstrated solution. The revised generator runs a reference program and uses the result as the target, so every task has a known solution. Over a path of 2,134 waypoints, the tool-tip positions of the browser and the server agree to one part in a million. Two sessions of three hours with 30 students of mixed programming experience showed two problems: cutting instruction was missing, and the students met barriers to further practice. These findings produced a new cutting lesson and changes to practice and competition. The analysis covers 1,615 log records and 13 reconstructed rounds. None of the 21 adaptive sessions reached the precision that is necessary to link practice with competitive item calibration.
Record
Status: Under review (double-blind)