Meetings are an important vehicle for human interaction. The Meeting Scheduling problem (MS) considers several agents, each holding a personal calendar, and a number of meetings which have to be scheduled among subsets of agents. MS is a naturally distributed problem with a clear motivation to avoid centralization: agents desire to keep their personal calendars as private as possible during resolution. MS can be formulated as Distributed CSP, but due to the form of its constraints the PKC model does not bring any benefit here. We take entropy as a measure for privacy, and evaluate several distributed algorithms for MS according to efficiency and privacy loss. Experiments show interesting results with respect to the kind of tested algorithms.