Skip to main content

Minutes for IESG at IETF-90

Meeting Minutes Internet Engineering Steering Group (iesg) IETF
Date and time 2014-07-23 21:10
Title Minutes for IESG at IETF-90
State Active
Other versions plain text
Last updated 2016-01-12

IETF 90 Administrative and Management Plenary
Minutes: Amy Vezza, IETF Secretariat

Narrative Minutes - Eliot Lear


   1. Welcome - Jari Arkko

   2. Host presentation - Gonzalo Camarillo

   3. Reporting

      - IETF chair - Jari Arkko

      - IAOC chair and IAD - Chris Griffiths and Ray Pelletier

      - IETF Trust chair - Tobias Gondrom

      - NomCom chair - Michael Richardson

   4. Postel award announcement

The award was presented to Mahabir Pun for his sustained efforts in the face of
adversity to network the country of Nepal.

   5. IAOC open mic

Wes George: We should stop using clip microphones.ÊIt creates a huge amount of
disruption and delay.ÊPeople can't be heard or they blow peoples' ears out.

Scott Bradner: Discussed matter with AMS.ÊThis crowd is incapable of using
microphones.ÊIt's amazing how inept this crowd is.ÊAnd so the discussion was to
not use the lavalier microphones, but use portable microphones.

The audience proceeded to misbehave with etherpad. No other questions for the

   6. IESG open mic

The experiment with etherpad failed.

John Klensin: I've noticed over the last year that there seem to be situations
in which other bodies seem to be doing work which contradicts, modifies,
deprecates IETF standards or IANA registries.  Specifically mentioned WHATwg. 
Relevant W3C groups seem to think that it is their job to mediate. What is the
strategy to deal with this?

Barry Lieba: Tactical answer: we have a good liaison relationship with W3C. 
The situation with the URLs/URIs is in a good place.  W3C needs some changes
related to IRIs. And they will write a spec on their side that extends RFC 3987
in the short term, but leaves 3986 alone.  The IAB is responsible for the
liaison relationships.  They can speak more to strategy.

John Klensin: concerned about other SDOs overriding our specs.

Barry Leiba: The strategy is getting healthy liaison relationships.

Jari Arkko: We don't have full power to say what others should or should not be
doing, but we can use those relationships where we have them.  It's the IETF
coming up with the right standards at the right time, and voluntary adoption by
the world of the right standards.

Alissa Cooper: And to make the work we do and our processes more amenable to
they way people do things on the web today. People go elsewhere to do things
when they think it will be easier than doing it here.

Russ Housley: We are not the Internet police. We develop standards, and they
are voluntarily implemented. If we developed timely specs we wouldn't be in
this area.

Joe Hildebrand: specifically about IRIs, we tried to do this work and failed.
We haven't shown any interest at coming back to the work. We've said we don't
want the work. It would be strange for us to be upset about this at this point.
 If we want to do real standards around IRIs, we need to find the people to do
the work and complete it in a timely fashion.

Joel Jaeggli: There isn't the stomach for doing that.

Scott Bradner: the 3GPP relationship evolved to a good place because we showed
that we could listen to requirements.  Maybe we need a way to do that in a more
general way.

Ted Hardie: A couple of comments on this topic. I don't think we're in a good
place yet. I think we have a plan to get to a good place in the version after
HTML5 (  In addition to the contextualization that was given before,
some of this is actually marketing.  We had a disagreement with the WHATwg
outside of the internationalization questions.  They really wanted to do
something that was fundamentally different.  They wanted to describe input
strings that could become URLs, rather than the wire format.  They did not want
to define the activity as defining input strings, but rather they wanted to
call those things URLs.  We ended up in a situation very early on saying that
it's great if you want to define the heuristics, but don't call those things
URLs.  We have to recognize that we do have a fork today because of these
fundamental differences.  The W3C is committed to this the IAB is committed to
it, but it will take some work.

Michael Richardson:  On the topic of mailing lists. We have a situation due to
SPF/DKIM/DMARC, we don't have NxN communication among the NOMCOM.  We are
experiencing pain because of this.  The IETF has been very VERY VERY remiss
about not setting a policy about what goes through our mailing lists.  We
should have been rejecting mail and 77 character limit and wrapping as well. 
We have been lax and negligent.  I don't care what one does in their
enterprise, but here it has gotten to a pain point.  This DMARC P-reject
problem is something we need to do something about.  And I have little
confidence that we will come up with a specification that will fix this absent
running code. Any solution is better than what we're doing now.

Pete Resnick:  I agree with you.  By doing what you suggest, certain classes of
users will not be able to participate if we follow this approach, but that may
be an ok thing.  If the IETF if the IETF is willing to put up with that bit of
closedness it might be doable.

Michael Richardson: I repy to email simply so everyone cans see them. Its

Randy Bush: Being liberal in what you accept is not something one should be. 
But I came to the mic about moderation.  If you choose no matter which way to
moderate the list, how will you measure the success or failure or the chilling
effect or whether people feel free to participate?  How will you judge the

Ted Lemon: The good news is that we have a database to plunder.  We have
metrics that we think are important.

Barry Leiba:  It's hard to measure chilling effects both in terms of moderating
and in terms of how the list is operating now.

Ted Lemon:  On an average conversation right now, how many participants are
there now?  If moderation is working we should see more participants and fewer
messages per participant.

Randy Bush: Thomas Narten does some measurement for us already. Whats missing
is a note that being high up on this list is not a prize.

Joel Jaeggli: I am relatively arbitrary and capricious, but I exercise
self-restraint, perhaps not enough.  I'd like to be better.  I would hope that
others would want to do better as well.

Keith Moore: I believe our society has too much preoccupation with measuring
and far to little about understanding what is going on.  It's hard to
characterize when posts were constructive.  Different people have different
biases.  Figure out what good means before you try to measure it.  There is
bias against frequent posting.  Frequent posting could be a good thing if the
conversation is constructive.

Spencer Dawkins:  The IETF discussion list is where we ask about WG formation
and about discussions we're about to approve. I don't get a whole lot of
feedback on the stuff I put out for IETF last call. But people think the IETF
list is a swamp. People are not willing to engage on that list. That's a
problem. There may be some soft measurements.

Jari: How many are actively involved in the IETF mailing list
Including sending some and reading some. Relatively small number. For those who
are not participating on discussions:  large majority of the room.

Stephen Farrell: And that's from the sample of people who were willing to stay
this late.

Pete Resnick: Spencer's experience is not unique. Several of us have the
experience of many participants not being willing to engage because of the heat
of some of the participants. That's a loss to me.

Jari: Quantity of the discussions is a problem, and that is also natural due to
the number of documents.

John Klensin: I measured audiences and effectiveness.  We have a tendency to
quantitative too early.  Counting postings without biasing creates an
interesting number, but it may not be meaningful.  He will spare us the details.

Stephen: Do you think over time the Narten number would have significance?

John: I doubt it. Randy didn't ask for quantitative measures but answers.
[Randy: bingo!] The issue with the Narten number is that it may shame people
from posting when really we would want them to. Separately: is the IESG making
a serious attempt to track newcomer retention three and five meetings out? What
are those retention figures?

Alissa: Why would we measure newcomer retention?  I don't think that's the
right measure.

Spencer: Three interesting things to look at as long as we use attendee fees
for ongoing work.

Kathleen Moriarty: There has been a concerted effort to help newcomers, through
diversity efforts, education, and other. I overheard a lot of conversations to
help and welcome newcomers.

Spencer: We spent more time talking about this a year ago with diversity.

Pete: The newcomers participating on lists have learned interesting styles from
others. I don't know if the aggressive posters are as successful.

Matt Mathis: I stopped reading the IETF list years ago, because I don't like
being scolded by people who don't have a clue. Although our official position
is that work gets done on mailing lists, people who don't show up to the
meetings really have great difficulty interacting with others. And I don't have
the time to deal with that.  I wish we changed the name of the list at the end
of every IETF so that people who didn't participate in the meeting, wouldn't
know the name of the list.

Lee Howard: Perspectives from other (younger) possible participants. Mailing
lists are archaic to them. Let's have a feedback mechanism on whose comments
are useful.

Randy Bush:  Thank you John for clarifying for me: think about what the goals
are.  It's far more important to retain Matt Mathis on the list than to get 35
newbies.  I confess I am the most severe user of procmail in the IETF.

Keith Drage: We have to demonstrate to newcomers how we can be useful.

Phil Hallam-Baker:  You have to search back through the chain of posts to find
out what the conversation is, and to understand whether they are talking about.
And so without that a 2 person flame war becomes 3. And that 3rd person comes
off the worse. I was told I was wrong about [a lot of things]. But now, years
later some people think I'm right about those things. I have an extremely thick
skin. And it is extremely difficult to discouraging me from talking. The NOMCOM
mechanism filters out dissent.

Keith Moore: We need people with thick skins.

Joel Jaeggli: Some of your arguments are incredibly well structured and change
my opinion. But terribly high volume makes them less useful to me if I've
already absorbed the input. More wood behind fewer arrows, please.

Keith Moore: Different participants are looking for different things from the
mailing list. IESG needs to get one thing. Others might need to get another.
Being attacked from all sides, not defending yourself would be perceived as a
sign of weakness.

Jari:  I get it. But still we should all try to do the same thing and make our
arguments as concisely as we can, and we don't have to respond every time. And
the issue is that we don't have a working group chair for the IETF mailing
list. And I don't have the time to do what a WG chair does.

Pete Resnick:  There are folks who will attack, and people respond concerned
that the technical point will be lost. We need to capture the technical point.
WG chairs do that, its a dynamic process, and we need managers for that process.

Ted Lemon: You are not actually being perceived as weak if you stop. Part of
the job of a moderator is to SOURCE_QUENCH. We have little monsters in the
backs of our minds that need to be source quenched.

Jared Mauch from NTT: We have a variety of technologies. We can also use the
telephone for communication. We need to leverage other technologies as well.
Also, not every message must be immediately replied to. We need to use all the
technologies available to help move forward.

David Black: Not being involved in the IETF list, I've moved about 50 docs to
RFC. It is not necessary to use the IETF list as a means for review. A lot
review takes place in the work group list or area-wide review lists. And so
maybe now is a good time to review the purpose for list?

Pete Resnick: Should different uses be split out? Last Calls, Charters, is this
what you are suggesting?

David Black: Of all the things that go on that list (ietf@ietf), when and for
what purposes is that the primary forum?

Ted Lemon:  I've thought about this quite a bit.  Perhaps there is a tooling
solution.  LCs go to IETF-announce.  Maybe there is a single use mailing list
for that last call.  Frustrating to not see last call comments.

David Black: Focusing LC review elsewhere is a good idea. We have running code.
Maybe not single use lists.  WG lists are more effective.

Joel Jaeggli: IESG members are pre-moderated on all lists. Upshot is the ADs
looking for reviews from specific groups and they can target questions to
groups we think need to review for specific technology.

Keith: Raised discussion about how to judge consensus through tallies.

Leslie Daigle: Need to catch yourself in defending your pride.  But it is hard
to know when your technical point has been heard.  If we create an environment
where people where people back off where people don't know they've been heard,
We'll lose new people and old people, we will do damage. Also, claiming there
are people with clue and not clue is not helpful, and is offensive.

Richard Barnes: Well said. We are here to reach consensus, not to defend or
attack. It's important to recognize when one is in the ruff at the end of a

Hui Deng: We might have a different understanding on network neutrality. What
is IETF guidance on network neutrality here?

Barry: We had an IAB plenary on network neutrality in Maastricht, I think.
Perhaps go back to the proceedings that will give you good background on the
discussion about it here.

Joel Jaeggli: Operational perspective: have been a network operator in China.
Worked for Nokia and they were a content provider in the region. Your
experience is not remarkably different from elsewhere. They are not easy to
solve. I prefer moving packets more easily and having less settlement, but the
world is not that different than what you experience.

Hui Deng: Would like the IAB to publish what we mean by network neutrality.
Would be very helpful.

Joel: There are a lot of business decisions. We tend not to get involved in how
people run their businesses. In North America if two providers choose to run
their connections too hot. IETF doesn't want to weigh in on those decisions.
We're also really bad economists.

Matt Mathis: Everybody likes the idea of network neutrality, but speaking as
someone who has done work in that space, I don't know what that means.

Doug Otis: Using VoIP a lot. Why can't the IETF do more with the IETF using
VoIP? Is that crazy?

Richard Barnes: Not a crazy idea. WEBRTC community is trying to making it
easier. Working with IAOC on this.

Jim Gettys: A lot of the perceived problems is not insufficient bandwidth but
buffer bloat.  Go read my blog.

Jari adjourned the meeting.

Other Resources:

NOC Report:

IANA Report:

RFC Editor Report:

Secretariat Report: