This study investigates the extent to which open access (OA) publishing models affect French researchers' attitudes. Research questions were: What place does OA have in attitudes of French researchers in mathematics, biology, and computer science? Are French researchers aware of new publishing models? Do they publish in these new outlets? What funds do they use? What kind of feedback and satisfaction can we observe? We used a quantitative approach to identify structural trends. From April to May 2015, the authors gathered empirical data (interviews, training, and meetings) on researchers' opinions and understanding of OA publishing models to define the variables and build up the questionnaire that was sent out in June 2015.The findings of the survey demonstrated that, when choosing journals to publish their articles, researchers did not rate OA as a criterion but rather impact, prestige, and symbolic recognition. Publishing OA articles is only in its first stages in the publishing practices of researchers. Lack of policy, prescriptions, and lack of rewards explain why OA is not a factor that determines the publishing strategies of French researchers. The researchers who take the risk of OA publishing are biologists, the older and more experienced researchers, and those who are members of editorial teams.

Au cours de ces dix dernières années, il y a un engagement croissant de l’Union européenne en faveur de l’innovation ouverte, le libre accès et la science ouverte. Notre objectif au sein de cet article est de s’interroger sur les origines de ces politiques et d’en retracer les évolutions et les limites. L’objectif de cette analyse est également de mettre en avant les injonctions contradictoires que subissent aujourd’hui les chercheurs en matière de publication et de diffusion de l’information scientifique et technique à travers entre autres les problématiques et questionnements liés à la brevetabilité des résultats de recherche financés sur des fonds publics.

With the continued development of open access policies, it is important to promote consensus‐building projects with the various stakeholders. This article gives an account of how such collaboration has facilitated the construction of the project Héloïse. This project is a French information service dedicated to describing the policies of French publishers on the self‐archiving of scientific publications. Héloïse represents a real tool of mediation whose development involved much debate between publishers and research stakeholders. This article seeks to demonstrate that the development of trust between the actors involved in the project was a major component of its success.