Hello,
I'll reply to several messages below to reduce ietf@ mail traffic.
At 03:03 27-06-2013, Eggert, Lars wrote:
Section 2 says:
RFC 3777 [RFC3777], Section 5, "Nominating Committee Operation",
Paragraph 1 of Rule 14, is replaced as follows:
Members of the IETF community must have attended at least 3 of
last 5 IETF meetings remotely or in person including at least 1 of
the 5 last IETF meetings in person in order to volunteer.
A few questions:
(1) How do you define "remote attendance"?
(2) How does the secretariat determine whether someone has remotely
attended? (Based on whatever definition of remote attendance you have in mind.)
I prefer not to get into a definition of "remote attendance" for
now. For what it is worth the current system only tells us that a
person has paid the registration fee. That person could have gone
shopping, fallen asleep during the WG sessions, or sitting in a
corner as he or she does not have a dot and is not considered as important.
Question (2) is based on the assumption that the IETF Secretariat has
to make that determination. I am leaving question (1) open as I
would like to listen to what you and everyone else has to say
first. Question (2) could be addressed after that.
At 03:33 27-06-2013, Yoav Nir wrote:
When we started having one day passes, the question was raised about
whether attending on a one day pass "counts" for NomCom eligibility.
If I remember correctly the answer was no. So we couldn't let
listening in on one session count, right?
Yes.
I guess you can prove attendance by Jabber log, but that's not
really helpful IMO.
Agreed.
At 03:50 27-06-2013, Stephen Farrell wrote:
I think physically attending 1 meeting might not be enough
esp. if a nomcom were selected that had a lot of folks on
who'd only ever been to one meeting. But I'm sympathetic
to the goal and am sure some qualification rule (*) could
be worked out. (Probably after about 1000 messages;-)
The MySQL field reserved only has two characters for the message
count. As such the limit is 99 messages. :-)
Please note that I am not ignoring the view expressed above.
However, before getting into that I'd like to hear from
folks who've been on or chaired nomcoms. I know a lot of
it is done remotely, but how important is the f2f part
that happens during meetings? Would it really be ok if
say 5 voting members could never come to a meeting
whilst serving? (And I think that'd not be an unlikely
outcome.)
I'll leave to folks who have been on NomCom to comment.
(*) Like I said, too early to get into it, but the nomcom
selection process could also require that the voting
members collectively have been to N meetings, with each
voting member able to contribute at most M to that total.
Say with N=30 and M=4 or something, and keep running the
random selection until you get 10 voting members that
satisfy that. And if we went there, we could also require
that nomcom as a group have written a number of RFCs
perhaps or even have some folks with Jari's h-index>3
or something. We could have lots of fun with all that:-)
My first response to the above would be no. However, there is
something in the above which is interesting. Once I get a sense of
the high level picture I may get back to it.
At 04:06 27-06-2013, Olafur Gudmundsson wrote:
I read the draft, I think there might be some merit to this proposal
but I think the threshold issue should be clarified.
Ok.
What does "one of the last five mean" during an IETF meeting?
"The 5 meetings are the five most recent meetings that ended prior
to the date on which the solicitation for nominating committee
volunteers was submitted for distribution to the IETF community."
It's any one meeting out of that.
I think the threshold of having attended one meeting is too low, I
would relax the rule to say something like this:
"must have attended at least 5 meetings of the last 15 and including
one of the last 5".
15 meetings is 5 years, I know that is a long time, t this will
allow people that that have been involved for a long time but have limited
resources to attend to participate in Nomcom/recall processes.
I tried that in another draft. I dropped the proposal because of
other NomCom-related issues.
Q: do you want to limit how many "infrequent" attendees can be on
Nomcom just like the number of people from a single organization can sign a
recall ?
The quick answer is yes. One of the issues is the assumption behind
"attended meetings". There is also the question of fairness. I will
have to reconsider my answer based on the views that are expressed.
At 05:36 27-06-2013, John Curran wrote:
I have read the draft, and believe that it moves the qualification to
serve on the NomCom in the right direction. Long-term, it would be ideal
if remote IETF participation was equivalent (both as an experience and as
a NomCom qualification) to in-person IETF participation.
My preference is to look for both short term and long term results.
Noting agreement in the direction, the reality of remote participation
today is somewhat different. In recent years, I have been a frequent
remote participant and occasional on-site participant, and while it is
possible to effectively contribute to working group efforts remotely,
such success is predicated on knowing quite a bit about IETF processes
and workflow, and it not clear to me that a remote participant picks
up the necessary background at anywhere near the same rate as on-site
participants. As a result, I am concerned that the proposed language
in draft wouldn't necessarily provide for experienced IETF participants
in the NomCom, and/or those who have well-informed insight into what
makes for good IAB/IESG/IAOC members.
Reality is a dream that comes true when there is agreement and
commitment. I have participated remotely. In my opinion, it is not
possible to contribute effectively unless one has read the draft
being discussed. Being on-site does not mean that a person has done that.
The reality is that some working groups are not working and it has
been like that since several meetings. The reality is that one or
more directorates have not been working since several years and
nothing was done about that.
What, among other things, make a good IAOC member is someone who
reads the contracts before approving them. I leave it to experienced
IETF participants to state whether they consider that as well-informed or not.
There are three IAB members from Europe and one from China. There are
two IESG members from Europe. There is one IAOC member from
Europe. Everyone else is from the United States. It has always been
that way and that's how it will be.
I chatted with a student from Greece. It was the middle of the night
at his location. He has never attended the IETF. He writes code and
participates in a working group. He offered to explain to me the
work being done in that working group. In my opinion he has an
understanding of the IETF. My opinion is based on something he told me.
Note also that the proposed language also increases the possibility of
"capture" (i.e. the ability of an single organization to inappropriately
skew the outcome of the process) in that a relatively large pool of
remote participants could quickly be made NomCom-eligible by having them
attend the very next IETF meeting, and then all volunteered to serve on
the NomCom. While this is not a particularly likely course for a party
not happy with the IETF, it is an aspect to be considered in the NomCom
processes.
There is a limit on affiliation. Any attempt to "capture" will be visible.
With an view towards finding a middle ground, would it be possible to
change your proposed text from:
"Members of the IETF community must have attended at least 3 of
last 5 IETF meetings remotely or in person including at least 1
of the 5 last IETF meetings in person in order to volunteer."
to this:
"Members of the IETF community must have attended at least 3 of
last 5 IETF meetings remotely or in person including at least _2_
of the 5 last IETF meetings in person in order to volunteer."
Please do not read this as antagonistic. I will say no. I am
actually saying no to the interpretation of "attended".
The change from 1 to 2 meetings being in-person significantly reduces the
potential risk of capture while also increasing the exposure level of
NomCom volunteers to dynamics that occur in the hallways and between the
formal IETF working group sessions. The net result recognizes the value
of remote participation, moves in the right direction, but does so at a
more moderate pace than you originally propose.
There is a potential risk of doing nothing. Bernard explained
something to me. I think I understood what he wrote. I would be
lacking empathy if I blamed him.
I have one word for the proposal, fairness. I would like to read the
explanation of the IETF Chair if the potential risk of capture means
that the IETF cannot be fair.
Disclaimers: My views alone. NomCom '95 Chair (back before any NomCom
procedures existed... :-)
Was that during the clergy days? :-)
At 06:18 27-06-2013, Michael Richardson wrote:
I have not read the thread yet, on purpose.
Ok.
I would like to thank you for going out of your way to do something I asked.
As a person who has done significant remote participation myself, and has also
observed the difficulty new people have in understanding how things fit
together, I can not support your specific proposal, but I support the idea.
I would suggest:
2. Updated Text from RFC 3777
RFC 3777 [RFC3777], Section 5, "Nominating Committee Operation",
Paragraph 1 of Rule 14, is replaced as follows:
Members of the IETF community become eligible for the NomCom by
having attended at least 3 of the last 7 IETF meetings in person.
Once a person has become eligible for NomCom, they retain their
elibility to NomCom by attending at least 1 of the last 4 IETF meetings
in person, and at least 3 of the last 5 meetings in person or remotely.
Should a person lose eligibility for NomCom, they return to
not-eligible.
(We could, true to form, describe this as a state machine with three states,
or even a simpler to write in Verilog one with 7-8 states)
I agree that new people can have difficulty understanding how things
fit together.
My quick response to the suggestion is no. The reason is because of
the way "attended" is being interpreted and because I don't think
that it is good to have a high barrier of entry.
I have lowered the bar to remain eligible such that a person who not travel
for 12 months (such as someone on maternity/paternity leave. Civilized
countries get at least 1 year..) could remain eligible.
Thanks for the explanation. There was an unrelated discussion about
parents finding it difficult to attend IETF meetings.
At 06:24 27-06-2013, Michael Richardson wrote:
2) for the November meeting, the nomcom *itself* must be present.
I think it unrealistic to think that the nomcom itself could
be remote for that meeting. For the summer and march meeting,
the nomcom could be anywhere.
When is it summer? :-)
At 06:48 27-06-2013, John C Klensin wrote:
While it risks taking us into a statistical rathole, I think
that notions like the above may be the sort of thing we should
look at.
I left the door open for that.
More broadly, I think we may need to try to figure out what we
really want and need on or from the Nomcom in this decade
(remembering that the system was designed for the rather
different times and IETF composition of the early 1990s and has
been tuned in minor ways but not carefully and openly reviewed
since) and then try to devise criteria to match. It seems to
me that may require looking at separate aspects of things rather
than trying to come up with a single surrogate for everything.
Yes.
We might want to look at whether some collections of
participants should be guaranteed representation or weighted
more heavily in the selection calculations (whether voting or
otherwise). That raises the risk that SM identifies of pushing
us toward a Nomcom as a representative body of constituencies
demanding slots, but the advantages seem very strong for Dave
Crocker's proposal to guarantee a certain level of expertise and
some ideas to be sure that the perspective of remote
participants or other underrepresented populations are heard.
The question is how to find the right balance and then reach
sufficient consensus around the justification that we can hold
the line. Not easy, but, at least IMO, probably worth the
investment it would take.
I need to get back to something Dave Crocker wrote. There was a
discussion (not on an IETF mailing list) which pointed to
constituencies and representativity. I don't think that excluding
people is a good idea. The last sentence, in a way, sums up what the
proposal will have to achieve.
Similarly, I'm pretty sure that "groks the IETF" [1] is an
important and useful criterion. I don't think "3 of last 5" is
a valid exclusive surrogate. Perhaps what is needed is a list
Yes.
If we separated the "IETF culture" requirement, I still think
that some level of participation, even face to face
participation is important. I don't think that 3 face to face
meetings in 5 is needed for people who already understand the
culture; maybe some combination of remote participation and less
frequent attendance should be equally acceptable.
The "IETF culture" requirement is ambiguous. Arturo asked a good
question. I would describe the side discussions as being about
understanding why the IETF is doing X.
"Participation" is similar. If we think it is important, then
someone who is actively contributing to mailing lists and
document reviews and who is showing up in meeting Jabber logs
with useful comments is, IMO, a more appropriate Nomcom member
than someone whose company pays registration fees and travel
expenses and who then shows up at meetings and either goes to
the beach or sits in a few WG meetings reading email. I don't
know how to eliminate the second (perhaps others have ideas) but
I can think of ways to identify the former as long as they are
not the exclusive "minimum participation" admission criteria.
My focus is "participation".
I would just hope that we don't fall into the trap of focusing
on what is easy to measure and quantify rather than what is
important and a good measure of what we are looking for. It is
I'll add that the trap is also focusing on the details prematurely.
Regards,
S. Moonesamy