Today I braved the treacherous ice-flows of Cambridge to attend (along
with just 10 or 12 other audience members, unfortunately) the
roundtable held by Common Cause and CDT on ICANN At-Large membership
(http://www.commoncause.org/icann/). The elections expertise
represented in the room was incredible, and I think ICANN will find
itself forced to heed the most widely-endorsed results of the
discussion. I would summarize them as follows:
1. Participants unanimously denigrated indirect elections and the
At Large Council. They want direct elections, if any are to be
held.
2. There also appeared to be unanimous agreement that ICANN must
choose between acting as a narrowly technical body (if that's
possible, which many participants doubted) and deciding a wide
range of policy decisions related to names and numbers. Some
participants seemed to see a spectrum where efficiency lies at one
end and representativeness (for the sake of legitimacy,
accountability, and democracy) at the other. Efficiency is the
more important goal for a narrow technical body, and
representativeness for a policy-making body. This choice could also
determine how broadly one draws the line around those allowed to be
members.
3. Many participants also stressed the importance of authenticating
the electorate; they did not find an email address and postal
address to be satisfactory.
There's also a news flash from participant Steven Hill, who talked to
Greg Crew and claimed the Board has decided on a voting process, which
Hill calls "instant runoff voting": the whole membership votes
separately on each of the At Large Council slots. Hill pointed out
that this "majority of a majority" system means as little as
one-quarter of the membership could choose all nine of their Board
representatives. The system violates the recommendation of the
membership advisory committee for proportional representation (a
method Hill endorses).
Since the discussion was loose and ranged back and forth, I'll focus
below on particular topics, organizing quotes and ideas under each
one. I haven't checked the quotes against the cybercast
(http://www.commoncause.org/icann/webcast.htm) but will try to do so
soon--in the mean time, please recognize that the following is my own
scribbles and summary; it hasn't been checked against any of the
participants. Go to the cybercast for the real thing.
Also remember that the roundtable was open-ended and speculative; the
suggestions were thrown out for the purposes of argument and should
not be taken as firm positions by the speakers. The list of speakers
is at http://www.commoncause.org/icann/participants.htm and should be
consulted so you understand their background.
Role of ICANN (technical or policy?)
------------------------------------
Harshbarger: "ICANN is a first-of-its-kind oversight body for the
Internet."
Bok: We want to make sure technology "serves human
ends rather than some sort of private exploitation."
Berman: ICANN is moving from the administration of names to a
broader mission to "decide who has a name."
Zittrain: ICANN has a "pat thing they say, that they're not
governance."
Bok: A "note of pessimism": it's hard to elect people to a
Board whose function is ill-defined.
Berman: You can adopt one of two assumptions. ICANN can run
the domain-name and number system ("making the trains
run") without governing the world, or "ICANN governs
the world." As a civil libertarian, would prefer a
clear limit on ICANN's mandate so it has to stick to a
technical role. Otherwise, you get a situation where
every constituency, such as abortion rights activists,
must be represented. I'd rather keep all these privacy
and civil rights issues from becoming ICANN's
business.
Harshbarger: But if it's so technical, would we "bore the hell out
of everybody" except the MCIs of the world and
therefore let them make all the decisions?
Hill: As domain names become more profitable and their role
becomes more complex, it's harder to stick to just a
technical role. "The cat is already out of the bag,
and walking in a direction where the domain-name
system needs some soft of governance board to prevent
capture by commercial interests." Using the WTO as an
example, "there's no way for international bodies to
avoid the big issues of who has power, first world
versus third world, and so on."
Davidson: The mandate of ICANN is narrow and technical but it
"really matters." The only way forward is to involve
the broad public.
Zittrain: Traditionally, technical bodies try to stay
"disinterested." But many modern bodies go precisely
for the "interested" forces, a "stakeholder model."
Issues at this stage of the membership process
----------------------------------------------
Zittrain: Three main questions come up after the membership
advisory committee report:
"Is membership really the way to go, and what does
that mean?" Not many people are likely to care enough
to study and vote on the issues.
Capture is too easy when you expect such low interest.
Nominations are a related question--who chooses who
can run?
Due process: How can ICANN "replicate the real-world
processes that we expect from agencies meant to
represent the public interest, without it getting too
complicated?" One must identify voters and prevent
ballot-box stuffing.
Davidson: Many open questions remain, including rules for
campaigning and conflicts of interest. At CDT we
request comments on:
How to minimize the danger of capture.
Rules for authentication, conflict of interest, and
financing of campaigns.
Better voting systems, to "get away from the
majoritarian box" (that is, letting the majority
decide everything).
Fischer: ICANN is trying to set up the election process all by
itself. This is not how it should be done: an
independent, objective administrator is required. We
are also concentrating too much on details before
setting down the prerequisites such as the function of
ICANN and the Board. The process of determining
elections is fragmented; ICANN needs to put together a
"seamless vision."
Indirect elections
------------------
The voting process chosen by ICANN (as reported by Hill) was described
above.
Bok: The At Large Council is "not likely to work." It's "an
odd assignment" to join a council whose only function
is to elect people to another body who's function
isn't clear. The Council will attract only people very
focused on achieving a particular goal.
Still: Using an intermediate Council risks screening out
several minorities that collectively could add up to a
substantial part of the membership. It's a perfect way
"to make people feel their vote doesn't count."
Costello: The At Large Council is "unwise and unworkable."
Accountability and capture
--------------------------
Hill: Current Board members come from large commercial
interests like MCI/Worldcom. "The organization has
already been captured by a combination of commercial
and technical interests." But "that's not necessarily
an evil thing. Still, it makes sense to graft on to
this more seats for people to watch the Board."
Bok: It's too difficult to get a Board that has so much
power and is still accountable. Perhaps try a
referendum system: let the whole community vote when
issues go beyond a narrow technical mandate.
Alternative: let members put issues to a referendum
vote when they feel the Board doesn't represent the
membership, but make it hard to get an issue on a
ballot. There should be "some way for the electorate
to rise up against the Board." Key issue: are you
willing to put votes up to a membership, or do you
just want the "trappings of accountability"?
Molnar: You have to worry about capture not only by special
interests, but by "fanatic extremists." Find ways to
limit ICANN's power; "protect us from ourselves."
Perhaps have two boards, one fully elected and another
chosen by a more accountable body such as the UN.
Perhaps have a form of appeal. "One Board is just too
much power in one place."
Potter and Costello agreed with Molnar that an something like an
advisory board makes a lot of sense. Zittrain pointed out that the
current proposed 18-member arrangement is a "bicameral structure" and
that there are provisions for reconsideration and review in the
by-laws.
Engstrom: The current set-up "downplays legitimacy." Ten seats
are set aside, elections are indirect, and staggered
elections reduce the effect of the members at each
election ("district magnitude"), reinforcing the
majoritarian system.
Hill: We seem to have a "fear of the masses." We don't have
to worry about capture through elections; people like
Berman in the CDT can go online and drum up a couple
hundred people to take their position.
Pantic: If things really go wrong, ICANN can be dissolved and
we can start another body.
Hill: Consider a sunset clause: after five years, for
instance, the Board or the members have to vote on
whether to keep the system.
Molnar: Nominees can be held to certain standards; for
instance, we could require respect for human
rights. Joerg Haider could not run for the ICANN
Board.
Nye: "Legitimacy doesn't come just from majoritarianism."
"Representativeness matters, effectiveness matters,
and accountability." The European Parliament is not
particularly successful at being representative and
accountable, and most other international bodies have
similar problems. The WTO is effective on narrow
grounds, but demonstrates problems when its decisions
spill out.
If ICANN is a narrow technical body, focus on
effectiveness. If it has a broader impact, a broader
mandate is important.
Scope of membership and choosing voters
---------------------------------------
Potter: Just as people outside the U.S. often wish they could
choose our government (because its policies affect
them), people not currently on the Internet are
affected by ICANN.
Nye: Reinforced Potter's observation, and pointed out that
nine slots of the Board are already reserved for
"technical people." Perhaps require members to
"inform themselves" as "the price of entry." This
could be bad, like a poll tax, if the mission of ICANN
is broad, but it's appropriate if ICANN is restricted
to a narrow technical function.
Costello: (Several quick suggestions, not necessarily
consistent.) Instead of limiting membership as Nye
suggested, aggregate the voters according to their
interests. Focus on the nominating process. Put
forward good, knowledgeable candidates whom "you can
trust" and "pre-screen the candidate list." You need a
"fair, representative nominating process." Like Gould,
thinks elections are premature.
Cranor: In real life there are political parties, so people
have a sense who they support. On the Internet there
are no political parties, and we don't even have a
good sense of what the issues are.
Gould: Membership is "premature"; proposes a "two-stage
process." The current Board should determine criteria
for the next board, perhaps choose another nine, even
run the organization for three years and then decide
whether to have elections. Start "from the top down":
who figures out who is a candidate? How can candidates
campaign?
Still: It's hard to verify identity from an email address,
and hard to verify geographic location.
Pantic: Email users are not a good population, because most
don't understand the domain-name system. Choose a
narrow group that understands the issues.
Cranor: In both the IETF and the W3C, power tends to go to the
engineers who invest the "sweat" in understanding the
issues. This involvement requires both technical
education and time (often given by the participant's
employer). People are happy with this so long as
decisions are truly technical. As soon as decisions
spill out into non-technical areas, "people see
trouble." Then they look at the participants see
they're all "American white men; that's not very
representative." While she wants a "broad electorate,"
voting online is very difficult to ensure technically.
Hill: Perhaps institute a small registration fee, decreasing
for poorer countries.
Trevor and Fischer said some form of registration for voters is
necessary.
There was also a wide-ranging discussion of the idea of restricting
membership to domain-name holders. Several people pointed out the
potential skews: domain-name holders don't have the same interests as
users, some are huge (aol.com) while others are tiny, many hold
multiple domain names (give speculators a big voice in the Board!),
there are costs in most places for holding domain names, and the
U.S. has many more domain names than other countries.
----------------------------------------------------------------------
Andy Oram O'Reilly & Associates, Inc. email: [EMAIL PROTECTED]
Editor 90 Sherman Street phone: (617) 499-7479
Cambridge, MA 02140-3233 fax: (617) 661-1116
USA http://www.oreilly.com/~andyo/
Stories at Web site:
The Bug in the Seven Modules Code the Obscure The Disconnected
----------------------------------------------------------------------
