This article aims to point out some results of the implementation of a theory stemming from a research in mathematics education and we are going to discuss its contributions and limitations. A future primary teacher in his masters year at the University of Geneva decided to try to implement the notion of sets of tasks (« jeux de tâches »), developed in our PhD (Del Notaro, 2010). That concept describes the experimenter as an element of what Brousseau (1998) calls the « milieu1 » who involves his own knowledge to interact with pupils. After describing what we mean by the notion of sets of tasks and its origins, we are going to give a definition and to state the problem by exposing our main idea: the fact that the exploration of the milieu by the experimenter will interact with the exploration done by the pupils. This will demonstrate that this interaction is an interaction of knowledge. Our research methodology is also going to be presented. Finally, we are going to reveal certain effects of the didactical transposition and to analyze the interpretation the student has made. We are going to show how the transposition of a theory in class transforms and makes the knowledge evolve. As a conclusion, we are going to mention that although the exercise was quite successful in some aspects, there should be a discussion about the effects of the transposition to understand the evolution of the knowledge. It is certainly not easy to define this point, but we are going to propose some elements of thinking.