<p>This article presents a model for students’ application of algorithmic knowledge during a lesson on pattern generalization: a boundary object bridging algebraic and algorithmic thinking. Drawing on the anthropological theory of didactics, the study proposes using a task type generator for structuring pattern-based programming activities and a praxeological model for analyzing algorithmic knowledge. The experiment, conducted among students aged 11–12&#xa0;years using a program called Pixel’Art, showed a progression of the techniques the students used, from naïve programming to more elaborate strategies, including the use of loops. The didactic variables of the generator, such as the number of repetitions or the complexity of the patterns, influenced the transition towards other problem-solving techniques. Analyzing the activity traces also revealed two frequent technological elements: resorting to visual control to assess the programs and the belief that “there’s a bug,” which motivated adjustments. This field experience shows the relevance of a task type generator and a praxeological model for describing and following the development of students’ algorithmic knowledge. Finally, the article highlights the value in transposing theoretical frameworks derived from mathematics to computer science to enrich didactic practices and deepen our understanding of this emerging discipline.</p>