If we opt for this type of problem, the method to understand the challenges faced by students and teachers
changes dramatically. Students must realise that, because of their formative and pedagogical goal, there is no point
in attempting to memorise solutions or copy those of their neighbour. Preparing for an exam should not be a
repetition of problems that have previously been solved by the teacher, a classmate or in the book, thus trying to
memorise the mechanisms of resolution without understanding them. Teachers are faced with a major challenge:
inventing new problems whose topic represents the desired lesson and that students can interpret (at least partially)
using the knowledge they previously possess. In this manner, with the teacher's help, classmates' help or help from
the guide to the specific problem, students may apply this knowledge when answering questions raised by the
situation. Moreover, problems aimed at learning must teach students how to explicitly solve them, particularly in a
heuristic manner. Problems must also link the theoretical models specific to the subject instead of only resorting to
routine mechanisms that prevent reflection. How do we develop these learning-oriented problems? How do we
know if a problem is good for learning?