IESG Response to the Appeal Against the Removal of the Co-chairs of the
GEOPRIV Working Group
Introduction
This is the IESG response to the appeal by Randall Gellens, Allison
Mankin, and Andy Newton posted at:
http://www.ietf.org/IESG/APPEALS/IESG_Appeal_20070622-final.pdf
Cullen Jennings recused from all discussion of this appeal.
The appeal raises three major points for the IESG to address:
1. The removal of the WG Chairs violates IETF process;
2. The actions taken interfered with the consensus process; and
3. There is a conflict of interest.
The appeal also proposes a remedy. This response includes some
comments about the proposed remedy.
1. The removal of the WG Chairs violates IETF process
RFC 2418 says:
Working groups require considerable care and feeding. In addition
to
general participation, successful working groups benefit from the
efforts of participants filling specific functional roles. The Area
Director must agree to the specific people performing the WG Chair,
and Working Group Consultant roles, and they serve at the
discretion
of the Area Director.
Since all WG chairs "serve at the discretion of the Area Director,"
they can be replaced at any time. The previous GEOPRIV WG co-chairs
were told about their removal in private before the public
announcement. This action was not required, but it is the most
polite way to handle the situation. Perhaps the public announcement
could have provided some rationale, but the authority to remove a WG
chair is clear.
2. The actions taken interfered with the consensus process
The appeal claims that a sequence of events, including the
replacement of the GEOPRIV WG chairs, interfered with the consensus
process. From the wording of the appeal, it is not clear which items
in the sequence are offered as background. It is clear that removal
of the GEOPRIV WG chairs is claimed to have interfered. Some of the
events in the sequence occurred more than two months prior to the
submission of the appeal, and it is recognized that these events are
not subject to appeal. Each of the items in the sequence is
addressed for completeness.
2.1. Changing of GEOPRIV WG session time at IETF 68
It appears that many events lead to the unfortunate rescheduling the
GEOPRIV session at IETF 68. Andy's broken arm, schedule conflicts, a
wedding, and airlines schedules impacted by bad weather on the East
Coast of the United States were all contributing factors. As a
result, only one of the GEOPRIV co-chairs (Randy) was at the IETF 68
meeting in Prague. The RAI ADs were faced with a situation where
either the SPEERMINT WG or the GEOPRIV WG was going to meet without
any of their chairs present in the session. With little time to make
a decision, the RAI ADs chose to adjust the schedule so that
SPEERMINT, the younger of the two WGs, would have a session where a
WG chair could attend.
Due to the schedule change, Randy was no longer able to attend the
session. The change created a conflict with LEMONADE, and Randy
needed to be in the LEMONADE session. However, the GEOPRIV WG co-
chairs had already selected Henning Schulzrinne (the author of RELO),
to co-chair the session with Randy. The decision was made that Jon
Peterson (one of the RAI ADs) would co-chair the session with
Henning.
Scheduling is a problem that faces all ADs. This was a choice of the
lesser of two evils. In hindsight, it may not have been the best
possible choice. However, a reasonable decision was made.
2.2. Changing of the GEOPRIV WG session agenda at IETF 68
Discussion of Layer 7 Location Configuration Protocol (L7-LCP) was
planned, but it is not clear from the agenda that a decision between
HELD and RELO was planned. As usual, the first agenda item was an
agenda bash. There was an agenda change that moved the discussion of
location signing to the end of the session, but this change is not
relevant to the appeal.
The original agenda is available at:
http://www3.ietf.org/proceedings/07mar/agenda/geopriv.txt.
The minutes summarize the agenda as:
A. Agenda Bashing
B. Document Status
C. Future Directions
D. L7-LCP Problem Statement & Requirements
E. L7-LCP Protocol Selection
F. Geo URI
The agenda that was published prior to the meeting does not indicate
that RELO would be discussed, and it does not include the "L7-LCP
Protocol Selection" agenda item. A participant could easily have
been surprised when these topics were added during agenda bashing.
Review of the audio stream of the agenda basing portion of the
GEOPRIV WG session at IETF 68 makes it clear that all of the
participants were made aware of the push for a decision. No one
spoke against this proposed change to the agenda, and no one hummed
against the revised agenda when asked.
2.3. Assessing the opinion at the GEOPRIV WG session at IETF 68
Discussing the L7-LCP problem statement and HELD were clearly on the
agenda that was posted prior to the meeting. It is clear that people
following the GEOPRIV WG knew HELD and RELO were the two HTTP L7-LCP
candidates. The appeal asserts that no decision should have been
made at the WG session at IETF 68. However, it was made clear
through hums, that the people in the room felt they needed to make a
protocol selection. They clearly indicated that waiting for
consensus on the requirements document was not needed.
It was recognized that location signing was still a topic of active
discussion. However, the still-under-discussion problem statement
document no longer included location signing as a requirement. The
GEOPRIV WG chose to remove it from the requirements between the -00
and -01 drafts. Draft -02 was available prior to this WG session.
The appeal takes issue with the participation of Cullen Jennings in
the counting of hands. Ted Hardie and Cullen both helped the Acting
GEOPRIV WG Session Chairs count hands. RELO and HELD each got 22
hands. Yet, hums indicated that the people in the room wanted a
decision. They were faced with a tie. To break the tie, Cullen cast
a vote for RELO. Rohan Mahy stated during the session that Cullen
had previously indicated no technical preference between RELO and
HELD. However, Cullen did indicate that he believed that a decision
had to be made. In hindsight, Cullen regrets casting a vote. In the
end, Cullen's vote made no difference since votes from the Jabber
room were used to break the tie. By totaling the hands in the room
and the Jabber room votes, HELD was declared the winner.
The appeal questions the integrity of the Jabber process. The belief
is that most Jabber participants were listening to the real-time
audio feed; therefore, they were not asked different questions.
Rather, they were prompted to organize their responses. The Acting
GEOPRIV WG Session Chairs were satisfied that a plurality of opinion
had emerged, and no one has challenged their assessment of the
outcome.
The appeal raises questions regarding RFC 3929. However, RFC 3929
was not employed. Ted Hardie, the author of RFC 3929, explained
during the session how the process used was not compatible with RFC
3929. Ted spoke quite clearly about the process that was proposed
for selecting the protocol, and based on review of the audio stream
the people in the room were in agreement with using the proposed
process.
2.4. Interfering with the GEOPRIV WG chairs' processes after IETF 68
It is clear that Cullen Jennings was pushing the GEOPRIV WG to make
progress. Over the years, the GEOPRIV WG became a divisive group.
Cullen was looking for closure on long-standing unresolved issues
that were preventing the GEOPRIV WG from fulfilling their charter.
Nearly every AD finds themselves in this situation at one time or
another.
On March 30, after the IETF 68 meeting in Prague, Cullen told the
GEOPRIV WG co-chairs that he planned to relieve them of their
positions. The appeal states that a reason given for this decision
was that the chairs should "never have allowed the HELD proposal to
remain viable." Cullen denies saying such a thing. He also said,
"I'm not sure what I said that could have been misinterpreted as
meaning this."
(Note: The IESG has no way to determine what was actually said during
this discussion between Cullen, Randy, Allison, and Andy.)
Cullen did not dictate the choice of document editor to the new
GEOPRIV WG chair. However, he readily admits that it was "obvious"
to him that a good editor would be required. According to Cullen, he
asked the previous GEOPRIV WG co-chairs if they had any suggestions.
They had none. Therefore, they engaged in a discussion of the
qualities to look for in an editor. Cullen suggested they consider
Mary Barnes. In a previous conversation with Mary, Cullen had asked
her if she would be willing to be editor if the chairs asked her to
do so, and she had indicated that she probably had time to take on
the job. Mary is not strongly connected to either of the proposals,
is very organized, and is experienced at working as an editor on
highly controversial proposals. Cullen reports that the previous
GEOPRIV WG co-chairs seemed to think Mary sounded like a good choice.
The current GEOPRIV WG chair ended up choosing Mary, and he believes
that Mary is the best person for the job.
After IETF 68, the previous GEOPRIV WG co-chairs began the process to
confirm the direction that was set during the GEOPRIV WG session.
They used a series of questions, giving mail list members a certain
period of time to respond. The previous GEOPRIV WG co-chairs were
replaced with the current GEOPRIV WG chair before this process was
complete. However, the current GEOPRIV WG chair continued the
process. At the end of the allotted time, the current GEOPRIV WG
chair assessed the mail list traffic, and determined the consensus.
No one, including the three appellants, has approached the current
GEOPRIV WG chair to indicate disagreement with the recorded
consensus. Any disagreement with this consensus call should begin
with the current GEOPRIV WG Chair, and then work its way up the chain
if satisfaction is not found. There is no evidence that the change
of WG chairs during the consensus call had any impact on its outcome.
3. There is a conflict of interest
When dealing with conflict of interest, perception is just as
important as reality.
It seems that everyone agrees that the GEOPRIV WG has become divided
into at least two camps on the HTTP L7-LCP issue. The appeal refers
to "the Cisco camp," and claims that Cullen used his IETF leadership
position to support that camp.
The appeal suggests that Cisco favors a particular L7-LCP solution.
From the GEOPRIV WG history it is clear that participants from Cisco
have advocated lower layer LCP approaches (such as DHCP), and
therefore it is quite difficult to understand how Cullen's effort to
force a choice between HELD and RELO supports the inferred Cisco
agenda. Since the core DHCP documents had long since been completed
(RFC 3825 was published in July 2004), it is the opinion of the IESG
that these actions by Cullen did not further any Cisco agenda and
were instead targeted at timely completion of the GEOPRIV WG charter.
The agenda that was posted prior to IETF 68 and the minutes reflect a
time slot for Cullen Jennings to discuss "Future Directions." The
appeal says that Cullen used this time to seek acceptance of a DHCP
extension. This extension had not been previously discussed in the
GEOPRIV WG and at the time had no associated Internet-Draft.
However, review of the audio stream from the GEOPRIV WG session shows
that very little time was spent discussing any aspect of DHCP. After
the meeting, James Polk wrote an individual submission which the
GEOPRIV WG may or may not adopt to fill the identified hole. This
draft is not a subject of this appeal.
The appeal states that the agenda slot was given to Cullen for
validation of the current milestones. Milestones were certainly not
the focus of any discussion during the GEOPRIV WG session. When
asked about updates to GEOPRIV WG milestones, Cullen indicated that
he was not prepared to address new milestones until outstanding
critical work is complete. There seems to be a serious
miscommunication on this point. The message referenced in the appeal
is about updates to the milestones in the existing GEOPRIV WG
charter. The previous GEOPRIV WG co-chairs were not proposing new
work items. The proposed milestones can be found at:
http://www.ietf.org/mail-archive/web/geopriv/current/msg02571.html
The IESG believes that Cullen's actions are consistent with an AD
that is pushing a divisive WG toward closure on chartered work. The
IESG also believes that the GEOPRIV WG milestones are seriously out
of date, and they should be updated promptly.
4. Proposed Remedy
The proposed remedy is not appropriate. Appeals are designed to
allow the IETF to reconsider a decision and to correct a mistake.
The remedy proposed in the appeal is to move the GEOPRIV WG from the
RAI Area to another area. This recommendation does not correct any
mistakes that may have been made, nor does it consider the technical
needs of the GEOPRIV WG. Rather, it is an attempt to move the
GEOPRIV WG to a place in the IETF organization where Cullen Jennings
will have less influence.
By the way, the question of transferring the GEOPRIV WG to the
Applications Area has been discussed on at least one previous
occasion. The proposal was discussed by the RAI and Applications
ADs, and they decided that the GEOPRIV WG fits well in the RAI Area.
They recognized that the GEOPRIV WG has implications for things other
than SIP and that it makes use of protocols developed in the
Applications Area. As a result of these discussions, Lisa Dusseault,
one of the Applications ADs, agreed to be a technical advisor to the
GEOPRIV WG.
The IETF has at least two mechanisms to address concerns about
recurring biased behavior of an AD. While such concerns might be
offered as rationale in an appeal, an appeal is not the best
mechanism for addressing long-standing AD behavior concerns. In the
future, if anyone has such concerns, the IESG believes that these
other mechanisms ought to be used to address them. The first and
least disruptive mechanism is input to the Nominations Committee
(NomCom) when the offending AD is being considered for a subsequent
term. The second and much more radical mechanism is the recall
process as described in RFC 3777.
5. Conclusion
For the reasons provided above, the appeal is denied.
The IESG observes that the GEOPRIV WG milestones are out of date. We
encourage the RAI ADs to work with the current GEOPRIV WG chair to
update them.
_______________________________________________
IETF-Announce@ietf.org
https://www1.ietf.org/mailman/listinfo/ietf-announce
