[council] (Fwd) STV with Geographical Diversity

[Dennis Jennings <Dennis.Jennings@ucd.ie>]

FYI

Dennis

------- Forwarded Message Follows -------
Date sent:      	Wed, 22 Sep 1999 04:22:14 +0900 (JST)
From:           	maruyama@nic.ad.jp
Subject:        	STV with Geographical Diversity
To:             	dennis.jennings@ucd.ie
Copies to:      	intl-wg@nic.ad.jp
Send reply to:  	maruyama@nic.ad.jp

Dear Dennis,

This is Maruyama with JPNIC.

I heard from Mr. H. Hotta that you are proposing a certain modified
version of STV for the comming election of ICANN directors from DNSO
NC.  It sounds very good to me.  Here I would like to suggest you that
one should give a good name for a good idea to be widely accepted.  I
would like to suggest "Single Transferable Vote with Geographical
Diversity"("STV with GD" or "STV+GD" for short) to distinguish this
one from the original STV.

See you in LA.

Regards,

----
N. Maruyama (Vice President of JPNIC)
maruyama@nic.ad.jp

> Names Council, Berkman Centre.
> 
> Here below is the expert recommended process for the election by the Names 
> Council of the three ICANN Directors.  This process is simple (YES, it is), 
> guarantees that voters preferences are followed, ensures that no votes are 
> wasted, guarantees a geographically diverse outcome, ensures that a single 
> vote gives a result (i.e. no rounds of voting with the risk of each 
> subsequent round being influenced by the previous rounds), and is 
> internationally recognised as "sound".  (The expert advice came from Mr. 
> Simon Hearn of the Electoral Reform Society of the UK - )
> 
> The process uses a Single Transferable Vote.
> 
> The Electors (19 Names Council voters) list their candidates in order of 
> preference - 1, 2, 3 etc.
> 
> The person with the highest number of first preferences (No. 1s) gets seat 
> A, provided they get the required quota of votes (19 seats divided by (3 + 
> 1) - i.e. 5 votes).  Other candidates from the winner's geographic region 
> are then eliminated and their votes redistributed for the next stage, in 
> the usual way and so on.
> 
> If nobody reaches the quota on the first count, which is more than likely 
> if there are many candidates, the candidate with the least votes is 
> eliminated and his/her votes are redistributed, and this proceeds 
> repeatedly until someone passes the quota mark.
> 
> Once a person is elected to Seat A, their surplus of votes is redistributed 
> according to the next preferences indicated on the votes cast for them. 
>  The next person past the quota get seat B - after which candidates from 
> seat B winner's region are also eliminated for the next round and their 
> votes redistributed (always providing there are candidates from other 
> regions).  Etc.
> 
> It may sound complicated, but a simple calculation gives the results 
> immediately.
> 
> This method is fair, guarantees that the preferences of all the voters are 
> taken into account, and definitely gives a geographically diverse outcome. 
>  It is also a standard process, and can be objectively scrutinised by an 
> independent expert.  It is also understood to be sound, PROVIDED that the 
> geographically diverse elimination mechanism is well publicised in advance 
> and is clear to all the voters (in this case the NA members - so this is 
> not an issue)
> 
> To meet the 50% rule, the following is added.  On completion of the voting 
> and elections as described above, a ratification vote is held - i.e. the 
> result is voted on the by Names Council, each member having one vote.  To 
> ratify the election, the result must receive approval by 50% of the votes 
> cast.  If the result fails this test, a new election is held.  (This 
> provides a useful final confirmation (or rejection) process).
> 
> Thanks
> Dennis
-------------------------------------------------------
Dennis M. Jennings
Director, Computing Services, University College Dublin.
Address:  Daedalus Building, Belfield, Dublin 4, Ireland
E-mail:  Dennis.Jennings@ucd.ie
Telephone:  +353-(1) 706 7817
Fax:        +353-(1) 706 2362