--On Thursday, March 21, 2013 17:23 +0100 Martin Rex
<mrex(_at_)sap(_dot_)com> wrote:
Keith Moore wrote:
...
IESG is the review body of last resort. When WGs do a poor
job of review, especially cross-area review, the burden
falls on IESG to take up the slack.
As I understand and see it, the IESG is running IETF
processes, is mentoring IETF processes (towards WG Chairs,
BOFs, individuals with complaints/appeals), and is trying to
keep an eye on the overall architecture, and put togethe the
pieces from reviews they obtain from their trusted reviewers,
such as directorates.
Not only does that not match closely what is specified in the
various BCPs, but there is much to quibble about it in practice.
Again, I strongly suggest that actual experience with how things
work would be a lot better than suggesting changes on the basis
of theorizing.
...
I also don't see how _more_ reviews could make things worse.
Actually, we have worked examples of that too. One problem
typically arises when someone reads a document, doesn't
understand it, but, having done the work, bogs document
processing down with typographic and editorial issues that did
not create ambiguity and that could easily be resolved by the
RFC Editor.
Had you said "more competent, focused, and substantive reviews"
I would have agreed.
I believe it would be naive to expect IESG to perform reviews
all by their own, either not asking for or ignoring all other
input and then VOTE in committe style.
But this was exactly the expectation some years ago and
continues to be the expectation of an ADs who have not
established their own review support mechanisms.
The way the IETF positions are defined and filled, biases of
various ways are _inevitable_. They solution to this is to
set up processes in a fashion that will produce good results
even where there is strong bias of various kinds -- aka "lack
of diversity" -- by distributing the work to other IETF
leadership positions besides IESG and by putting in place
controls that will likely notice and object when IESG decisions
seem to exhibit bias, and procedures to deal with this.
We more or less started with an IESG that was strictly a
steering and management body with standards approval elsewhere
(in the IAB of the time). We did away with that, putting
document final review and approval in the IESG as well. The
community has been extremely resistant to suggestions to change
that. I agree with you that it would solve a number of problems
but we might be the only two people who believe that making the
change would be desirable on balance.
Conversely, if you are convinced that there is real bias that
led to particular unfair and incorrect decisions, the appeals
process actually works very well.
But once you structure processes&controls and distribute work
in a fashion that makes it resilient to bias in I* positions,
the whole issue of diversity will be much less of an issue for
those positions.
As indicated above, I don't think that restructuring is going to
happen. Even if it did, it would merely reduce the
possibilities for deliberate abuse and bias to distort the
system. It wouldn't eliminate any of the other arguments for
diversity, most of which apply even if everyone is operating
completely in the open and in good faith.
john
given the brokenness of IETF's structure.
Brokenness usually suggests defects that could have reasonably
been avoided. While there are certainly a number of features
that each come at a cost, I'm not aware of an actual
brokenness of the IETF's structure, i.e. something that could
have been reasonably been avoided without loosing any benefits.
-Martin