> Do you think such an addition should be made to the privacy list spec?
Yes, if it's not there already. Btw, the simple communications blocking XEP already has a clause that requires the server to send presence of unblocked contacts. > Although perhaps a trivial implementation of *always* doing a probe when the > privacy list changes isn't a horrible idea. Right. And if you let the client do it, then you rule out the server's possibility of doing optimizations (since it doesn't know why you are doing a probe). Moreover, the client can't do any optimizations on itself, because it doesn't know the privacy list logic (and it shouldn't try to either, because the server might have its own logic for doing that). cheers, Remko
