Coda keeps source distributions current across a set of machines. The model is that the master source tree is kept on a server host, and client hosts run coda to get themselves in sync with the server. This is different from rdist where the server "pushes" files out to its clients. Coda is not as flexible as rdist(1), but it requires less of the client machines in that they do not have to be running Berkeley networking programs or even Unix at all.
