An open multi-agent system (MAS) is a MAS where agents may be added, be removed or evolve (e.g. modify their abilities). Usually, the openness of a MAS is controlled by a middle-agent. In this paper, we tackle the problem of openness from a general point of view (considering the middle-agent approach as one of the possible solutions). Then, we propose an original approach based on the principle of distribution which requires a collective activity of the agents to manage openness. Finally, we compare these two approaches to determine the advantages and disadvantages of each.