Note-taker: Richard Barnes

0. Chairs introduced meeting and reviewed document status.  Round of applause for Gonzalo.

1. Simon Romano: CCMP
-- Current draft moves to a RESTful approach
-- HTTP single-verb transport plus CCMP (XML) body
-- Implements Create/Retrieve/Update/Delete model for managing objects

2. Simon Romano: CCMP Call Flows
-- Use case: Alice creates, configures, and joins a conference
-- Main open issue: Additional data required by the model
-- How many understand the issue?: Not many
-- M. Barnes: Oscar's comments were mainly about media, not conference stuff.  I think we need to address this.
-- Simon: You should be able to express relationships between conference objects (parent/child).  E.g., sidebars
-- Adam: We'll probably just ask Oscar to do one more rev of the data model to address this
-- Henning: Two issues:
   1. Atomicity: What happens when you add users and one fails?
      -- Have you thought about how to handle these cases?
      -- Simon: Don't know if that really belongs in the protocols
      -- Henning: There seem to be a few approaches
         -- Leave it to implementers
         -- Make a recommendations
      -- Simon: Would like to specify as much as possible in protocol
         -- But it's hard to acheive
      -- Henning: Should have hard succeed/fail
   2. What happens when you have patricide?
      -- Do children go away when parent dies?
      -- Need to clear up this case
      -- MBarnes: We've got that -- disallow killing parents
         -- That's why we need to know relationships in data model

3. Mary Barnes: Call flow examples
-- Fixed some errors to align with current protocol
-- Plan to complete details based on prototype
-- Publish as Informational
-- Adam: Do you want to add this as a milestone?
   -- Mary: Yes, that would be good
   -- Show of hands to add a milestone: Several (~7) in favor
   -- Gonzalo against: Don't see level of energy to do it
   -- Roni: If the document is ready, there's no work, no concern
   -- Lennox: Think you want to have error cases too
   -- Show of hands to adopt this document: 12 in favor, 0 against
-- Cullen: Do we have an idea when we expect these docs to be done?
   -- Adam: We linearized our work to deal with limited resources
      -- Think we're in the last straightaway
   -- Cullen: Pub-req before the next IETF?
   -- Mary: Yes.
   -- Adam: Who will do a detailed read? Roni, Theo, Lennox
   -- Rosen: As I recall, most tries at conferencing couldn't finish
      -- This progress is amazing
-- Simon: At these meetings, nobody has read the drafts
   -> Let's close this group and do DCON (!)

4. Geir Arne Sandbakken: BFCP over UDP
-- Trouble getting BFCP/TCP through RTP relays for NAT traversal
   => Translated BFCP to UDP
-- Now works with ICE, IMTC considering adopting BFCP
-- Open issues: Dynamic message sizes, security considerations
-- Adam: Not going to ask for adoption now, but wanted to present it
   -- People please read through this
-- Roni: Work is going on in IMTC because companies are meeting and working.  This work is important.
-- Elwell: This is important, especially since TURN-TCP isn't done
-- Cullen: If we limited message sizes, would it still be useful?
   -- If so, then that would be much easier than fragmenting
   -- Geir Arne: Biggest packets seem to be 40bytes
   -- Gonzalo: BFCP messages can get big, I will help with this doc
-- Lorenzo: There are packet size limitations in TCP too
-- Lorenzo: Have you considered UDP/TCP gateways?
-- Gonzalo: You would have problems with that.
-- Gonzalo: There's a scope question; do we want to do BFCP bugfixes too?

5. Adam: Large beer for Jon Peterson