If it is still possible, here are some minutes for PUFI: Minutes of the PUFI BOF as reported by Pete Resnick (chair) Jabber scribes: Eliot Lear and Spencer Dawkins (transcript available at ) The chair started the meeting by explaining motivations and introducing some ground rules. The BOF was initially formed to deal with discussions surrounding draft-carpenter-2026-changes, but many people decided that a broader discussion was necessary. A list of suggested process changes was collected by the chair from draft-carpenter-2026-changes as well as many other documents. The chair explained that we would have to gauge consensus based on (1) the amount of pain the change itself might cause (noting that the victim of the pain is an important consideration) and (2) the motivation to make the change. The ground rules presented for the discussion were: 1. In the room, IAB and IESG members have no different standing than any other individual. 2. Caveat to 1: The IETF chair (who called the meeting) will judge the consensus of the desire to go ahead with any effort and make the decision on whether work should go ahead. 3. Caveat to 2: If the room is motivated to go ahead but the IETF chair "judges" not to, that is itself an appealable item, so the room (and the IETF writ large) is really the final arbiter of whether to go ahead with some process change effort. It was noted that it would be undesirable to get to such an appeal. The community also has to be on board with the changes. Several expressed hope that the current IESG seemed willing to work toward something successful. Some discussion of the history of NewTrk took place. Concerns were expressed regarding the success of this work, and in particular that trying to solve some small point problems might lead to a worse result than the status quo. Scott Bradner noted that having a delta (as against a full rewrite) of 2026 would be unreadable. John Klensin said that some of the problems are worth taking on if 2026 is rewritten, but not otherwise. The group then looked at the prepared issues list taken from draft-carpenter-2026-changes, some NewTrk documents, and elsewhere. Discussion started around whether it would be desirable to open 2026 for some of the document handling issues. Paul Hoffman and Andy Malis mention that some of the issues which seem minor might become major depending on how they are addressed. At first, discussion centered around whether any of the issues could be addressed individually or needed instead to be taken as a group. There was not consensus in the room one way or the other that any particular issue was either (a) worth opening 2026 for or (b) able to be addressed as an individual issue. However, the discussion quickly moved moved away from the individual issues and into whether many of the issues could be addressed without significant change to the process for handling them. The issue list was abandoned and discussion moved to how to address process change in general. There was significant minority in the room (though by no means rough consensus) that felt a change to how process issues are addressed was necessary (approximately 1/3 of the room). About half the room indicated they weren't sure. One of the reasons given is that, even though certain changes might not require a different process, taken together they might. A side discussion took place regarding whether we could simply align the process documents with current reality. The objection was that, because some feel that the current reality is not desirable, but rather the currently document processes are, documenting current practices would in some way legitimize them. The room was asked whether they felt the IETF community would be willing to accept process changes (including those that would change the structures of process change). There was strong consensus in the room that the IETF would accept such changes if presented, but that we need to avoid an approach where "death of 1000 criticisms" kills any changes. Some consensus formed around creating two "design team" groups to have discussions: One would document the differences between our documented procedures and current reality, along with some evaluation about which of those states is desirable to have. The other would undertake a design of a "process change process". These groups would be run independently. -- Pete Resnick Qualcomm Incorporated - Direct phone: (858)651-4478, Fax: (858)651-1102