A 100.000 foot perspective on "what is the problem"

[James Kempf]

> Monday, December 16, 2002, 12:45:14 PM, you wrote:
> James> RFC 2850 says nothing about IAB architectural oversight of working
group
> James> activity after it is underway.
>
> By way of trying to see whether there is some practical rule that can be
> formulated, let me suggest that you try to construct a specific proposal,
> that goes beyond the reasonable, but vague "IAB provides architectural
> oversight of working group activity after it is underway".
>
> In other words, if we cannot find something pretty specific to require, then
> there is nothing we can do about this procedurally.
>

OK, at the risk of being told that this is a problem statement list and not a
solutions list, how about the following.

This is how an architectural review would be requested by the WG chair and
shepherding AD:

    The architectural review process is initiated when the Working Group
   Chairs and shepherding Area Director, in consultation, decide to
   request an architectural review of some architectural work coming
   from an IETF Working Group. The review is strictly voluntary and not
   required; indeed, unless the Working Group's charter and area of
   work results in some proposals for architectural changes, a review
   is entirely inappropriate. The motivation for the review may be a
   concern on the part of the Working Group Chairs and shepherding Area
   Director about the fit between the Working Group's concept and the
   Internet architecture at large, or it may simply result from the
   desire for an outside opinion on the architectural work. Scheduling
   of the review is typically done after the Working Group has come to
   consensus on their initial architectural concept, through the
   release of Working Group Internet Drafts with the problem statement,
   requirements, and/or framework for the subject area; however, the
   Working Group Chairs and shepherding Area Director can request a
   review earlier if the Working Group is having trouble coming to
   consensus on the architectural changes and they believe an outside
   opinion would help.

This is how it would be conducted:

   The review is always conducted at a face-to-face meeting, either at
   a regularly scheduled IETF meeting or at an interim meeting. If the
   review is conducted at a regularly scheduled IETF meeting, it is
   considered to be part of the normal Working Group agenda for that
   meeting. The IAB may send a list of questions about the existing
   architectural work that come out of the Working Group's documents or
   discussion with the IESG to the Working Group Chairs by email
   beforehand, but these questions are to be answered by the Working
   Group during the review. The length and content of the review is up
   to the Working Group to decide, but it should be concise and to the
   point.

And this is what would be done with the result:

   The shepherding AD and Working Group Chairs are under no obligation
   to utilize the review, and whether to implement the review's
   conclusions is strictly a matter for the Working Group to decide, in
   consultation with the shepherding AD. Alternatively, if the Working
   Group decides to ignore the result, the IESG may choose to enforce
   implementation as part of their management function in approving
   RFCs, if they agree with the conclusions. The review is not
   considered to be a formal part of the standardization process, but
   rather a semi-formal, advisory function. As such, it is not included
   in the formal work products of the Working Group.

There's more about consultation between shepharding AD and IESG about the
appropriateness of the review, how the IAB picks members for a review board,
preparation of the review board in consultation with the shepherding AD and
other interested IESG members.

This comes out of an ID I've written with a proposal for a semi-formal process
for conducting architectural reviews. The intent is to make an architectural
review an agreed upon procedure that a shepherding AD can use if she or he feels
it is necessary, without making it a heavyweight drag that can slow down the
process.

            jak