Re: [EAI] Please - Don't state preferences yet!

[John C Klensin <klensin at jck.com>]

--On Tuesday, 24 July, 2007 07:34 -0500 Harald Tveit Alvestrand
<harald at alvestrand.no> wrote:

> 
> 
> --On 24. juli 2007 12:34 +0100 Charles Lindsey
> <chl at clerew.man.ac.uk> wrote:
> 
>> On Mon, 23 Jul 2007 15:23:36 +0100, Harald Tveit Alvestrand
>> <harald at alvestrand.no> wrote:
>> 
>>> Note - my message was intended to start the discussion on
>>> the subject on how to decide on a MIME type. I am NOT
>>>  recording people's position until   we've decided what
>>> mechanism to use to decide!
>> 
>> Ah! The methods you have proposed is known as "Alternative
>> Vote", or "AV". It usually works fine, but can sometimes go
>> completely wrong.
> .....
> 
>> 
>> The ideal mathod, designed by the Marquis de Condorcet, copes
>> better with such situations, since it examines the complete
>> matrix of how many people prefer each possible option over
>> each of the others.
> 
> Thanks for the pointer -
> <http://en.wikipedia.org/wiki/Condorcet_method> gives more
> detail of Concordet voting mechanisms.

Since I used to do election analysis for a living, long ago, the
reality is that any of these methods can produce results that
are inconsistent with the desires/ intent of those responding,
and differ largely by the assumptions one makes about the intent
one wants to optimize.

Alternate vote optimizes first choices; while there are edge
cases (including possibilities for ties) something basically has
to be someone's first choice in order to win.

Condorcet optimizes one particular preference model.  In
particular, as Charles's model illustrates very well, it assumes
that an option that is everyone's second choice is better than
any option that is the first choices of a minority of the
voters.  As such, one might think of it as one version of the
model that a good choice is the one that leaves everyone more or
less equally unhappy.

It is important to note that any of the systems, including the
ones discussed below, get a lot more complicated, with more ways
to get into strange cases, as the number of options rises above
three.   And any of them can be gamed by colluding voters,
although some are more easily gamed than others.

But there are many other options.   For example, in a situation
like the one we have, there is a strong argument that one should
first examine "I completely hate this" votes, eliminating any
options that are intensely disliked by a non-trivial number of
people, and only then applying a model based on affirmative
preferences to what is left.  Some feel that works better in
environments in which there is a lot of indifference (as long as
an answer is found) but a desire to eliminate really bad choices
before one is approved based on weak preferences.   This sort of
system is often considered bad for elections but good for some
other types of preference choices.

And then there are a whole series of "weight and scale"
techniques, some of them involving explicit elucidation of
strength of preferences and giving more weight to those who
care.  

Again, none of these is "right"; they just balance the factors
and considerations differently.

     john







_______________________________________________
IMA mailing list
IMA at ietf.org
https://www1.ietf.org/mailman/listinfo/ima