Ah, thanks. My mistake was assuming you could distribute new K as a normal message, encrypted with old K. Clearly (now) this wont work since new K will be distr

Ah, thanks. My mistake was assuming you could distribute new K as a normal message, encrypted with old K. Clearly (now) this wont work since new K will be distributed to everyone, including people who need to be omitted.