Sender: info-hol-request@edu.uidaho.cs.ted
Errors-To: info-hol-request@edu.uidaho.cs.ted
To: Phil Windley <windley@edu.uidaho.cs.panther>
Cc: info-hol@edu.uidaho.cs.ted, claesen@be.imec, Jim.Grundy@uk.ac.cam.cl
Subject: Re: CALL FOR VOTES - CALL FOR VOTES - CALL FOR VOTES - CALL FOR VOTES
In-Reply-To: Your message of Mon, 09 Mar 92 09:55:29 -0800. <199203091755.AA10839@panther.cs.uidaho.edu>
Date: Mon, 09 Mar 92 19:06:29 +0000
From: Jim.Grundy@uk.ac.cam.cl
Message-Id: <"swan.cl.ca.040:09.02.92.19.06.50"@cl.cam.ac.uk>


On Mon, 09 Mar 92 13:58:54 GMT, claesen@imec.be wrote:
+------------
| The deadline of 8 March 1992 for proposals to hold the HOL-93 workshop
| has passed. In chronological order I received the following proposals:
|
|    proposer:                  -  Place:
|
| 1) Juanito Camilleri          -  University of Malta, Msida, Malta.
| 2) Jeff Joyce                 -  Univ. Brit. Columbia, Vancouver, Canada.
| 3) Ramayya Kumar              -  Univ. Karlsruhe, Karlsruhe, Germany.
| 4) Phil Windley                       -  Coeur D'Alene Resort, Idaho, USA.

Then Phill Windle said:
> If we are going back to a free for all, Idaho respectfully withdraws its
> offer to host HOL93 so that we can cast our votes for UBC and not risk
> splitting the vote.

The issue of alternating between either side of the atlantic aside,
I would wish for Phill not to have to withdraw his popoasal in
the event of a free for all.
I think what we need here is a better syle of voting than the
over simplistic first past the post method, which is only ever
any good for deciding between 2 options.

Try this - its called preferential voting - and is used, amoung
other things, for political elections in Australia.

1. Everyone votes, not by marking the cadidate they most want to win,
   but by ranking the candidates in order of preference.

2. We repeatedly eliminate the candidate with the least votes,
   and redistribute that candidates preferences to the other
   candidates.
   That is:

       Put all the votes in piles according to their 1st preference.
       WHILE no candidate's pile has more than 50% of the votes DO
           Find the candidate whose pile is the smallest.
           Remove each vote from that candidate's pile and put it
               on the pile of the candidate the vote wanted next best
               to the candidate whose pile it was in.
       ENDWHILE
       Now the biggest pile has more than 50% of the vote and is
           therefore the winner.

Consider the following hypothetical vote:
60% of us want HOL93 in North America and
40% want HOL93 in Europe.

Suppose the 1st preference vote breaks down like this:
- Malta: 35%
- Karlsruhe: 5% (no offence)
- Vancouver: 30%
- Idaho: 30%

With preferential voting we will get the result that most people
wanted (either Vancouver or Idaho, depending on how the preferences
flow), with first past the post we wont (again no offence to Malta).
This example cuts both ways, perhaps most of us want HOL93 in
Europe but we split the Eurovote between Malta and Karlsruhe and
loose out to those crafty North Americans who organised to only
have one proposal.

Its not much extra hassle to administer, how about it?

Jim