Résumé
GAMA, a versatile agent-based modeling platform, has been in development for over 15 years and has attracted a close community of users and developers around the world, each adding their own unique twist to the project. Its use cases range from standard multi-agent simulation, to serving as the basis for serious virtual reality games, to feeding reinforcement learning algorithms. This diversity, however, brings many challenges to the development of the software and the communication around it. The multifaceted nature of the platform has resulted in a complex code base, causing the various layers of software to interact in intricate ways, especially in complex models. This complexity poses significant difficulties in maintaining the platform, which encompasses a wide range of technical issues. It can also hinder communication within the community, leaving active members unaware of new features, outdated practices, or upcoming events. GAMA has taken steps to improve its development practices, taking advantage of the many tools developed by the open source community, and in particular, making advanced use of the features provided by the Github platform. These changes are already showing results, such as more consistent software releases with detailed explanations of the changes they bring, or an overall improvement in code quality leading to greater stability. This keynote aims to show what's going on behind the scenes by presenting an overview of the life of the GAMA community (the different events, trainings, ways of communication, etc.), as well as the different solutions implemented in recent years to improve software quality and what are the plans for the future of GAMA.