Re: Thoughts over the holidays - IESG procedures.

[Scott Bradner <sob@harvard.edu>]

- publish something-like-this: good/bad/pointless?
>> may be bad if folks think that the community can direct internal iesg process
>> but I expect better to publish than not

- if good: material covered: subjects to add/delete?
- proposals for changes to better describe our current procedures?

>> just including the parts where I have comments

----------
2. IESG communications

   The IESG is in constant session through its mailing list -
   iesg@ietf.org.  The IESG has a teleconference every second Thursday
   for 2 1/2 hours, which is the place where decisions get reviewed and
   consensus on IESG action is gauged.

>> this says the only consensus gaging place/time is the telechat - this
>> is not totally true - we can, and sometimes do, deal with documents 
>> via email (teh stun doc being a recient example) - these words 
>> might be tweaked to indicate that we do not have to wait for
>> a call to get things done


3. IESG decisionmaking

   The IESG attempts to make all its decisions by consensus.

   The normal behaviour expected of an AD is that he/she searches for
   consensus, raises objections if he has them, listens to the arguments
   for and against her objections, and makes an informed decision about
   whether to go along with the consensus of the group, attempt to go
   further in discussing the problem, or recusing himself from the
   action because she has insoluble problems with the issue at hand.

>> above is true but it might be clearer to put it in the context
>> of documents that the AD is not sponsoring (not a good idea to 
>> abstain from those, even if possible)

4. IESG document review procedures

   The IESG review procedure is defined by the IESG.

   The procedure consists of:

   o  An initial review by the responsible AD, assisted by whatever
      reviewers the AD wants to bring to bear

   o  Once the responsible AD is satisfied that the document is worth
      sponsoring, a review by the entire IESG
>> if a stds track doc, or in some other cases, the step after an AD
>> is satisfied is an IETF Last-Call

   o  If the IESG has questions or comments, the responsible AD takes
      the token to resolve these with the authors or WG responsible
      before taking the (possibly revised) document back to the IESG for
      re-review.

   The procedure of IESG evaluation is different for standards-track
   documents and non-standards-track documents.

4.1 IESG review of standards-track and BCP

   For this class of document, each AD is requested to state an opinion.
   One AD for the area to which the document belongs (the "shepherding
   AD") MUST register a "yes" vote to the document in order to get it on
   the agenda.  At least 2/3 of the IESG must register either "Yes" or
   "No Objection" in order for the document to go out.
>> 2/3 of those not recusing for conflict (like being authors)

   If an AD has an objection to the document, this is discussed on the
   mailing list, and on a telechat.  If the IESG comes to consensus that
   the objection is worth sustaining, the shepherding AD is tasked with
   resolving the matter with the working group.  This may involve an
>>                                            ^
>>                           or author 
>>   can be author on individual submission or if a minor wording
>> change is needed w/o changing meaning of doc
>> note - Harald's words represent the ideal, right now much
>> of the time the WG has no idea that there was any IESG pushback
>> unless they look at the tracker

   explanation of the group's choices, clarification of text in the
   document, or a reevaluation of the choices made.

   The IESG tries to get a "token holder" AD for the objection; the
   chosen holder is not necessarily the one who raised the objection,
   but one who the IESG trusts to detect whether the problem is solved
   or not.

   When a document is updated, the shepherding AD will notify the token
   holding AD, and he will review the document; it can then either be
   approved instantly (if the shepherding AD and token-holding AD agree
   that there are no further problems), or it can be put on the agenda
   for a new telechat.

   The IESG tries to get all objections to a document in a single pass;
   however, this does not always succeed, and at times, resolving some



Alvestrand                Expires June 1, 2003                  [Page 5]

Internet-Draft              An IESG charter                December 2002


   issues can cause other problems to surface.  When the working group
   takes so long to revise a document that ADs have been replaced before
   the document comes back for re-review, it is also hard to hold new
   ADs to a promise to go lightly on surfacing remaining issues.
>> "go lightly" is not quite the right feeling - maybe 'be held to
>> a review done before they were on the IESG'

>> have to run to work - more later