On 2007-05-22 23:20, Lakshminath Dondeti wrote:
I am beginning to wonder about the IETF consensus processes and so
figured I will ask for advice on this list. First, a summary of some of
my observations over the past few months (I have real examples of these,
but we don't quite need to get into those):
1. Some people (WG chairs) seem to (want to) ask for consensus on
everything and that too many times:
1a. Ask for consensus on whether to adopt a document as a WG
whether something is/should
be in the charter
who should be an editor
should the WG meet
Those are all separate questions. What is the harm in a chair wanting
to test consensus, even on matters where the chair has power of decision?
1c. Ask again and again
That certainly happens when the reply to a question is silence, or just
a couple of messages, or irrelevant replies. A chair who really needs
to know whether a draft has WG support, for example, sometimes needs
to ask several times to hear from anyone except the authors. Again,
what is the harm?
2. Some people do it in multiple settings and in many different
2a. Ask on the mailing list
2b. Ask at a f2f meeting (sometimes, after having asked a
question on the mailing list)
2c. And sometimes ask again on the mailing list (sometimes,
after 2a even)
Yes. Results can be ambiguous and people can change their minds after
hearing additional arguments. As I understand it, the philosophy is to
get the best possible technical spec, not to meet an arbitrary deadline.
And that can mean looping over the same items.
3. Some people ask questions at a face to face meeting and declare
consensus one way or another
You can declare consensus of the people in the room. But as you say,
the rule is that the consensus that counts is the consensus on the list.
That's exactly why you will see the sequence 2a 2b 2c; 2c is to confirm
the consensus in the room reached in 2b. The discussion in 2b may change
some opinions given in 2a.
From 2418 and 4677, my understanding has been that
i) there are certain things such as appointing editors which are up to
the chairs' discretion and appointing chairs which is up to the ADs'
discretion (they may seek open or selective input, but are not
necessarily required to do so).
Exactly. So what is the problem with a chair choosing to seek WG input?
ii) asking for consensus once is sufficient and the process may be
repeated only if things were unclear.
And in my experience, they are often unclear, and new technical
issues are often raised.
iii) asking a question on the mailing list is sufficient to declare
consensus; if opinions were sought at a face to face meeting, the
chair/AD must ask on the mailing list before declaring consensus.
Yet I see ADs declaring consensus after getting the sense of the room at
a face to face meeting
But they must always add "to be confirmed on the list" and must always
do so - if they don't, it's a clear process violation and can be
and so I am curious whether our process documents
are out of date or whether I am reading them out of context.
They are out of date on many details but IMHO not on these
NEW: Preferred email for non-IBM matters:
Ietf mailing list