On Feb 7, 2011, at 2:01 PM, Adam Roach wrote: On 2/7/11 12:44 PM, Cullen Jennings wrote: I was somewhat surprised to see this back in LC. I am still not aware of any use case where this actually helps. I searched the IETF and WG lists for email with the subject draft-ietf-sipcore-199 and I do not see a single email that suggests there is support for this draft or the changes in it since the previous LC.
This draft has no use that I understand how it helps - it is at best a very limited optimization. The SIP standards is already too complicated with too many extensions. I believe this draft makes SIP worse.Thought the draft mandates that systems need to work even when the 199 are lost, I do not think that is how the proponents of the work intent to use. I could be very wrong but I presume that people intent to use to control middle boxes that control media gates. It's broken for that but given that is not discussed in the draft, it's hard to discuss how it is broken and what would be needed to fix it.
I do not support publishing this draft as standards track without actual WG discussion on what the problem is this draft solves and if there is WG consensus that problem is worth solving.
Cullen: Speaking as a chair of SIPCORE, I would like to clarify a minor point. Although it may not have been your intention, your message can be read as implying that the SIPCORE chairs and/or the RAI area directors have decided to move this document forward without community input.
It was really just sloppy overly fast writing on my part - I did not mean to imply that. The total time I'm willing to spend on this draft is pretty minimal.
For the benefit of those on this mailing list who were not involved in the earlier stages of this document's progress: the initial community input on whether to pursue the problem solved by this draft was taken at the face-to-face SIPPING meeting during IETF 69. In particular, the minutes for that meeting reflect: "Hum... in support of working on this problem."
In April of 2008, the requirements document was handed over from SIPPING to the SIP working group for working on actual protocol mechanism. (The SIP change process at the time limited core protocol changes to the SIP working group, requiring a change in venue).
Within the SIP working group, there were at least 316 message in 38 different threads on the protocol document over the one year period spanning April 2008 to April 2009. This period included a two-week long Working Group Last Call period in November of 2008.
In April of 2009, the RAI area underwent a reorganization, which resulted in the conclusion of the SIP working group. As part of the chartering of the SIPCORE working group, this document (and its associated milestone) was transferred to the SIPCORE WG.
In January of 2010, the document entered IETF last call for the first time. During IETF balloting, certain issues were identified by the IESG, and balloted as a "DISCUSS" position. Over the course of 2010, the document's author worked with the RAI ADs to address these issues to the satisfaction of the currently seated IESG.
I thought we were still in middle on conversation about what to do about any this stuff - thus my surprise it was in LC again. Last email I had was you in May of asking Christer if he was going to reply. He eventual did but it did resolve what this was use for. It did point out that no one was complaining about it. That caused me to send an email in June that said
"Silence is not consent - the WG does not care about this and there is not consensus that it is useful, what it is useful for, or that it should be published. I'd be happy to shown to be wrong as I, like I suspect many others, don't care too much one way or the other. "
After that, other than people saying the IETF had to do this because 3GPP had already made it mandatory, I had not heard any more about it.
As a result of these conversations, a technical issue involving interaction with another SIP extension mechanism was identified, and brought back to the SIPCORE working group. A resolution to the issue was identified in early December 2010, and a revised version of the document was produced to reflect this resolution.
Uh, you mean the one where the total traffic seems to be
Chair (Paul) proposed the WG should see if the WG was OK with it Christer (Author) said he was OK with even though he did not really like it Keith proposed some text without saying if he actually liked any of it or not Christer (Author) said he was OK with it
And no one else said anything. Hopefully I failed to find the email, or lost it, or am just confused, but I have a hard time imaging how that looks like consensus in one of the largest WG in the IETF.
At every step of this process, the IETF, RAI, and SIP community has opportunity for involvement. The volume of discussion demonstrates a non-trivial interest in this mechanism.
Hmm ... I 100% agree there has been opportunity for involvement. I am talking about the actually involvement since the previous LC over a year ago. In that year, there looks to be mostly silence about any of the issues or changes to the draft. I did not go do a good search but when I searched for the draft name, I'm not seeing much discussion. If you tell me 10 people were actively engaged in the last year and support these changes, I'm happy to agree I am in the rough in the consensus and ignore the whole thing. But if you are telling me that no one objected, then I do not believe silence is the same as consensus. Anyways - I this is very much on my low priority list - I have not gone back and looked carefully at the previous discussions.
|