ICANN/DNSO
DNSO Mailling lists archives

[ga-full]


<<< Chronological Index >>>    <<< Thread Index >>>

Re: [ga] Voting rules, take 4


"Roeland Meyer (E-mail)" wrote:

> A few comments, if you please;
> 1) small groups beget simple rules. Polling/consensus is
> appropriate.
> 2) medium-sized groups need more complex rules. Votng is
> introduced.
> 3) large groups absolutely need formal rules, with fully
> transparent accountability.

I don't know how these metrics help us resolve the issue before us.
"Complexity" is a moving target and may be in the eyes of the
beholder.  And, we lack definition for the various categories of
"size."  Indeed, there may be an inverse relationship between "size"
and the ability of a group to deal with complexities.

Almost all groups with more than 5 members seem to be conducted
according to rules.  Formal actions generally require defined
majorities and quorums.   Most rules do not depend on the size of the
group for their logic nor function.  Diversity of interests is
probably the most significant determinant of the type of procedure
required/desired by the participants.   Furthermore, the cost/benefit
analysis applied when considering a given proposal will take into
account the effects of complexity.

I suggest we discuss specific proposals on their merits rather than
invoking generalizations which must, themselves, be explained,
validated and properly applied.

> The problem is that the group size changes. Hopefully, the group
> grows.

Absolutely.

> Another problem is credibility.

Agreed.

> Accusations of electorial fraud
> is proportional to the complexity of the voting system.

I don't know the basis of this assertion.  I suspect other factors
predominate, such as the nature/significance of the things decided,
the diversity of interests affected, the mechanisms employed for
assuring fair results and the participants' perceptions regarding the
fairness of the procedure and administration involved.

> Even simple systems need to be defensible.

Agreed.

> Weighted voting rules are
> allways susceptible to challenge.

We already know that nothing is perfect.  So, everything is subject
to challenge, including simple but wrong answers.  We are seeking a
system which is susceptible to least challenge.  Again, that requires
a case-by-case analysis of proposals on their own merits.

> ...With due respect to messr. Froomkin, his proposal is too
> complex, with inadequate audit-trail.

Is this a criticism of proportionate representation a la STV (which
does not require political parties)?  If so, it is not responsive to
Harald's motion which deals with a different case--the multiple
choice/single winner vote.  And,  I question that conclusion.  I
think we all understand ranking our preferences.  The rest is done by
the computer.  The result is immediate.  I have not heard anyone on
this list argue that those results are less fair.

The same is true regarding IRV (which does address Harald's specific
case).

The "audit trail" issue, on the other hand,  is legitimate,
significant, and exists in all votes upon secret ballot.  We may need
some mechanism for appointing poll watchers.

I make no comment on the rest of Roeland's points.

--
This message was passed to you via the ga-full@dnso.org list.
Send mail to majordomo@dnso.org to unsubscribe
("unsubscribe ga-full" in the body of the message).
Archives at http://www.dnso.org/archives.html



<<< Chronological Index >>>    <<< Thread Index >>>