[ICANN-EU] Re: Is it possible to change one's vote or see the ongoing results?

["Andrew Bloch" <abloch@alum.mit.edu>]

From: "Andrew McLaughlin" <mclaughlin@pobox.com>
Sent: Tuesday, October 03, 2000 12:18 AM
> Andy:
>
> Thanks for the note.  I'll be sure to raise your security questions with
> election.com.
>
>
> [ During the at-large member nominations phase, voters could change their
> [ endorsements and the overall results were made public and updated
several
> [ times a day.  Will this be done during the elections?
> [
>
>
> No.  A vote cannot be changed:  votes are recorded anonymously, so it is
> not possible to determine which person cast which vote;  therefore, it is
not
> possible to change votes once they have been submitted.  Just as with any
> other ballot in the physical world.
>
> Also, there will be no interim results posted.  Such a practice tends to
distort
> the voting process.  At Large votes should carry equal weight, regardless
of
> whether they are cast early or late in the voting period.

Thank you for the answers.  As usual, I appreciate you taking the time to
answer my questions.  Here are my responses.  (Obviously, I don't expect any
changes in the current election, but these are things that should be
considered in the future.  That is why I've cc'ed the NA and EU forums.)

  o Allowing vote changes

That votes cannot be changed should be made clear (if it's not already) on
both ICANN's site and the election.com voting pages.  I think this is
unfortunate, given the number of candidates.  I would have liked to make my
initial vote now, and have a chance to refine it as I learned more about the
candidates.

A well designed system utilizing a one-way hashing function could allow
votes to be reliably changed without making it feasible to determine how
individual voters voted.  Obviously, you already can determine whether
someone has already submitted a vote.  The same techniques that could keep
the decision to vote anonymous could also make the vote itself anonymous.
At the same time, it makes the election results more verifiable by making it
easier to uncover voter fraud.

  o Posting interim results

Yes, clearly, posting interim results will distort the voting.  But so does
hiding interim results, in an election like this where voters are ranking 7
candidates.  Many voters will not bother seriously to consider the order of
their lower ranked candidates.  (This is a problem with instant-runoff
voting.  In traditional run-off voting, voters have a chance to refocus
their preferences on the final 2 choices.)  Yet, it may turn out that the
lower rankings are the only rankings that matter.  Posting interim results
allow voters to be more efficient.  At first, they only need to consider who
their top ranked candidates are.  After a certain amount of voting, it may
be clear that they need to consider their rankings of less-favored
candidates.  I think a combined system would work best-- a system where
interim votes are made public during the election, but, for the first day or
more, results are kept hidden.  The interim voting status could even be
progressive--at first, only totals of 1st place votes would be disclosed,
progressing to full or almost full disclosure as the election progresses.
(Note that such a system would make a larger ballot practicable.  For
example, debates such as the ones yesterday at Harvard Law School can still
take place because they can be limited to the top N vote getters, with N
decreasing as the end of the election nears.)

I believe that posting interim results worked very well in the endorsement
period.  Without it, almost certainly, there would have been fewer nominees,
in at least one region.  If the threshold had been much higher than 2%, such
as the original 10% proposal, it would have been essential for even one
member nomination in most regions.  (You can see this clearly looking at
http://www.icannnot.org/icannel.cgi , which I've recently fixed to tabulate
the final results correctly.)

I also believe that the endorsements were a little distorted by the early
reporting of interim results--before many members even knew that the
endorsement period had begun, a day or so before the first email was sent
out to members announcing that voting had begun.  (It's completely
understandable that the first email wasn't sent immediately, in case the
first few endorsements uncovered bugs in the system.)  (I admit that my page
at http://www.icannnot.org/icannel.cgi might have had a little to do with
the distortions, but it's hits were very low initially, especially among
non-candidates, and I didn't promote it much because of the experimental
nature of the whole election process.)

  o Such decisions should be made by the board in advance

Such apparently minor "technical" choices as these can, and do, greatly
effect election results.  These are types of choices that should be in
initial voting proposals, subjected to public comment, and ultimately,
decided upon by the ICANN board.  (Decisions like these matter a lot more in
the long run than many other decisions the ICANN board considered, such as
what to do in the unlikely event of a tie for the last open slot in a
region.)  At a minimum, the decisions should be made clear well before the
start of the election.

  o Think outside the [ballot] box

    "We seem to be stuck thinking inside the box of traditional, paper-based
elections.
    This is an electronic, web-based election.  We are not limited by the
same constraints
    as traditional paper-based elections."

These were the first few sentences of a public comment I submitted on the
election proposal that, for some unknown reason, never was made public.
[The full comment still exists on the ICANN website at
http://www.icann.org/cgi-bin/mbx/rpgmessage.cgi?election;3967B6AA00000015#ms
gthread , but it was never listed on the index page,
http://www.icann.org/mbx/election/ .  I sent a couple of emails to ICANN
but, surprisingly, I never got a response.]

It might be that a practice such as not allowing votes to be changed is
universal in the physical world.  However, even if that were true, practices
in the "physical world" should never be used as a justification for
decisions involving the virtual world of the Internet, even if such
practices are universal in the physical world.  At best, physical-world
practices are a starting point for discussion or an example which may or may
not be a good analogy for a virtual world, and this problem with
physic-world practices should always be acknowledged when discussing the
application of such practices in a virtual world.

Practices in the physical world are based on the severe constraints or
limitations of the physical world.  Practices in the physical world
shouldn't be applied to a virtual world, for several important reasons.
First, because many of the limitations of the physical world are lifted in
the virtual world, we should be able to devise a better practice in the
virtual world.  Second, once the limitations of the physical world are
removed from the actors within a virtual world, activities can take place
that the physical-world practices were not designed to deal with.
Physical-world practices may be severely suboptimal, or even
counter-productive, when applied to a virtual world.  Likewise, new
constraints of a virtual world can harm the effectiveness of a
physical-world practice, by impeding its implementation or by constraining
the actors in an incompatible way.

Attempting to help justify a decision by pointing out that it's done the
same way in the physical world proves nothing to me, nor anyone that
considers the possibilities of the Internet.  Instead, I am left with the
impression that little critical thought went into the decision-making
process, making me even more critical of the results.

I've taken the time to write this not to be overly critical of the
overworked and underfunded ICANN staff, but in the hope that after this
election, before making decisions, ICANN will engage in more critical
thought -- thought which goes outside the conventional boxes of the physical
world and takes into account the special nature of the Internet.

Andy Bloch