On Wed Jul 11 21:44:21 UTC 2012, Gunnar Helstrom < [email protected]> wrote:
> Dave, Sounds good. Can you convert your conclusion to modification proposals for chapter 5 and 6.2? Please hold off until Version 0.4; I already made changes and like comments in the next review cycle. v0.4 will only have about 1/10th as many changes as v0.3. The flux is going down. It will make it easier to re-review sections such as section 5 for further refinements (if needed). -- The new wording I already did, is cleaner and I think, more acceptable. -- Instead of "blank rtt", I'm more specific as to it being <rtt event='init'/> (Because it is a conveniently universal first element that's the same with and without disco). If the protocol is followed, it is harmless to send the optional <rtt event='init'/> (start RTT session) practically instantly before the first <rtt event='new'/> (start new RTT message). -- The spirit remains the same as chat state fallback: Chat states use the fallback of that single blank <active/> element transmitted and then watching for incoming chat states -- permitted by XEP-0085 chat states section 5.1. This is exactly the same "fallback detection" technique as a single blank <rtt event='init'/> and then waiting for response. -- Yes, I'm still trying to figure out if I can still also incorporate XEP-0115 (though I want to be consistent with the same method XEP-0085 plans to implement XEP-0115) -- to be determined. So please wait and review the v0.4 of the spec coming in a few days. Mark Rejhon
