Re: [council] Convention-Style Voting adapted to the DNSO conditions

[Amadeu Abril i Abril <Amadeu@nominalia.com>]

Dennis Jennings wrote:
> 
[,,,][
> 
> First Election to elect the Three Year ICANN Board member.
> 
> First round:  All NC's vote - 3 votes - three candidates must be voted for.
>  If any candidate has 10 votes or more, that candidate is elected.  If not,
> the candidates with zero votes are eliminated.
> ** Period of discussion within NC (I mean 30 - 60 minutes - not 30 - 60
> days)
> Second round:  All NCs vote - 3 votes - three candidates voted for.  If any
> candidate has more that 10 votes, that candidate elected.  If not
> candidates with 1 vote are eliminated.
> 
> Additional rounds -  until a candidate get more that 10 votes.

Dennis,

A) I aree with you and Patircio that the only way to elect three
candidates with the GD constraint thru the Convention-style is by
running three consecutive elctions. This was implied in my amendment
to AdrewĄ's proposal, even if I am afraid I did no t made it explicit.

I also agree that the first elected should take the three years seat,
the second the two yerts seat and tha last one the one year seat.

We also seem to agree on the fact that, at least on the first round,
the NC members should cast threee votes. But here come TWO main
differneces between your approach and mine.

A) I "object" to the fact that anyone be elected in that first round
In my view it only would serve to a) eliminate nominees with no or
marginal support and b) give all the NC an idea of the relative
support for each candidate. A poll, if you want. The reason is clear,
in my mind, whith the followng reasoning.

Under your system, a candiate haing the support of half the NC could
be elected. 10 votes or more. But as each NC member is casting THREE
votes each, the total numever of vote is 57. We could end with FIVE
candidates elected in the firts round. A real mess.

The situation is harldy bvetter trhu p`referential vote. (I prefer
preferential vote, exactly in the way you stated, but again used as a
poll, as a guidance for the real election). Here the totla of votes
cast is 114. As you assume that majority is simply "fisrt choice by
the majority of NC members, ie 30 votes or ten times being ranked
first) we still could with three candidates elected at the first round,

The conflicts regarding GD, and even helping finding a sound result
are neraly unsolvable. Short of reuiring super-supermajority votes,
hwih would also distort the sytem...and would be very uunlikely to
achive any ruesult during this first round.

In order to make things sim'ler, I would not insist on the candidates
with xx votes being "eliminated forever", just for that round. But I
insit in the benefficial effects of having at least one round with
three votes it, preferential or not (I prefer preferential) as a
helpful tool to clarify the possible outcome, the strnght of the
candidates, to helpf converge beyond the own consituency or regional
nominee and achieve sound results in as few rounds as possible.

B) Form the explanation above you might infer the second differnece: I
am  firmlyu confinced that, given the fact we are electing cnadidates
in three consecutive elections, and that we are using simple majority,
the only sound way to achive that is using ONE VOTE PER NC MEMBER form
that point on. I withdraw my susggestion of  scond "polling" round
with only two votes, if necessary for clarity reasons (and I think it
is). So, with the rsults of the poll. We start converging for the most
likley first cnadidate, easily discovering in no more thn two rounds
form whcih rgion it is, and knowing nevertheless that our initally
prefferred candidate form ohter region has not been eliminated, only
deferred to the next "election" of the second or third slot.

Combining election of one candidate at a time with three votes each
member of the elctorate, as I interpret form your mail (please correct
me if I am wrong) seems to prevent simplicity and even avaialability
of results.

With this in mind  I will review Elzbieta's combined proposal as a
further suggestion,.

Amadeu