Serious Games are computer games that are designed for a primary purpose other than pure entertainment and that can be used for teaching. Although they are relevant to many fields of education, their development remains very expensive and time consuming. In this thesis, we focus mainly on Learning Games (LGs), that is on Serious Games designed for educational purposes, and more specifically on LGs used for training students in engineering schools. The first part of the thesis is devoted to an analysis of the needs of those who create LGs and a state of the art in terms of methodologies and tools available. Our study shows the need to facilitate collaboration between the various actors, with complementary roles, involved in the development of a LG (domain expert, pedagogical expert, game designer ...). Our research also highlights the need for models and visual representations of the LG scenario to facilitate the design of LGs that are fun and attractive while still maintaining their educational values. To address the first need, we propose a global collaborative methodology for creating LGs in which we indentify the tasks assigned to each actor who is involved in the creation process and the tools available. To meet the second need identified by our preliminary study, we propose a LG scenario model that represents the educational structure chosen by the pedagogical expert and also the way this structure is integrated into a game scenario imagined by the game designer. To reify our proposals, we have developed an authoring environment called LEGADEE (LEarning Game DEsign Environment) that guides each designer with a "toolbar" adapted to his or her role and also provides a validation system that analyzes the ongoing creation. Lastly, we have designed an evaluation protocol to validate our authoring environment as well as the methodology and the models proposed during which we compare 24 LGs of which half were created with LEGADEE and half without. Our evaluation indicates that our tool tends to improve the quality of LGs at several levels. It also brings to light the limits of our work and provides guidance for future improvements of LEGADEE and the evaluation process itself.
