| Solaris |
|
|
CAB Voting System
Pilot Conversation: 1/29/05 to 2/02/05
CAB Voting System
- -
- Original Message --
Subject: [Opensolaris-discuss] CAB: voting system?
Date: Sat, 29 Jan 2005 11:51:04 -0500
From: Bill Sommerfeld <sommerfeld@sun.com>
To: James Grisanzio <Jim.Grisanzio@Sun.COM>
CC: opensolaris-discuss@opensolaris.org
I don't believe I've seen any discussion of the voting method to be used.
If you list a bunch of desirable properties of a voting system, traditional voting systems used (vote for single candidate; person with most votes wins) tend to not score so well. In particular, traditional systems work very poorly when there are large numbers of similar candidates, which is most likely what we'll face in selecting the CAB.
On the other hand, there are a bewildering array of alternative systems out there (including approval voting, preferential systems such as STV, and the bizarrely named "Cloneproof Schwartz Sequential Dropping"); some of these systems take some effort to understand and that by itself is a negative.
As a perhaps extreme example of how this can hurt, my understanding is that New Zealand just changed to use a particularly complex hybrid scheme, with the result that in some precincts 10% or more of the ballots were
uncountable.
Anyhow, there are a bunch of writeups on the various systems out there; many are written by people with an axe to grind one way or another, so read skeptically.
background reading:
http://electionmethods.org/
http://en.wikipedia.org/wiki/Arrow%27s_impossibility_theorem
http://en.wikipedia.org/wiki/Category:Voting_systems
http://en.wikipedia.org/wiki/Category:Voting_theory
and, of course, there are open source software packages out there
implementing some of these algorithms:
http://stv.sourceforge.net/
- Bill
-- Original Message --
Subject: Re: [Opensolaris-discuss] CAB: voting system?
Date: Sat, 29 Jan 2005 11:48:46 -0800 (PST)
From: hostmaster@rite-online.net
To: Bill Sommerfeld <sommerfeld@sun.com>
CC: James Grisanzio <Jim.Grisanzio@Sun.COM>, opensolaris-discuss@opensolaris.org
On Sat, 29 Jan 2005, Bill Sommerfeld wrote:
> I don't believe I've seen any discussion of the voting method to be used.
I'm sure Rick Hennasee (the guy who wrote an x86 assembler in
ksh or bash) would love us to use his scripts! :-)
--
Rich Teer, SCNA, SCSA, author of "Solaris Systems Programming"
President,
Rite Online Inc.
Voice: +1 (250) 979-1638
URL: http://www.rite-group.com/rich
-- Original Message --
Subject: Re: [Opensolaris-discuss] CAB: voting system?
Date: Mon, 31 Jan 2005 15:31:11 +0100
From: Joerg Schilling <schilling@fokus.fraunhofer.de>
To: sommerfeld@sun.com, Jim.Grisanzio@Sun.COM
CC: opensolaris-discuss@opensolaris.org
Bill Sommerfeld <sommerfeld@sun.com> wrote:
> I don't believe I've seen any discussion of the voting method to be used.
>
> If you list a bunch of desirable properties of a voting system,
> traditional voting systems used (vote for single candidate; person with
> most votes wins) tend to not score so well. In particular, traditional systems work very poorly when there are large numbers of similar candidates, which is most likely what we'll face in selecting the CAB.
Let me introduce that we will have something between 70 and 100
voters for the Pilot part of the CAB. This is more than you may
easily survey without involving a formal voting process - except
when there are only 2-3 candidates.
I believe that we may really first try to find the rules.
If we would like to follow usual formal voting rules, we should
first try to create a voting commitee.
The first duty of this commite would be to find the names
of the candidates and the names of the permitted voters. Usual rules
forbid that a member of the voting commitee would become a
candidate. If we go this way, I believe that this is a godd idea
for our election too. So if we go this way, we would need some
volunteers for this job.
What we all need to find out in addition is the way we would like
to vote: open, secret, or whatever....
Of course if you believe that we will be able to find the right
way without a formal process, you are welcome to describe your ideas.
P.S. I don't like bureaucracy but it may help us to defend our election
later to non pilot members of the community.
Jörg
-- Original Message --
Subject: Re: [Opensolaris-discuss] CAB: voting system?
Date: Mon, 31 Jan 2005 15:31:11 +0100
From: Joerg Schilling <schilling@fokus.fraunhofer.de>
To: sommerfeld@sun.com, Jim.Grisanzio@Sun.COM
CC: opensolaris-discuss@opensolaris.org
Bill Sommerfeld <sommerfeld@sun.com> wrote:
> I don't believe I've seen any discussion of the voting method to be used.
>
> If you list a bunch of desirable properties of a voting system,
> traditional voting systems used (vote for single candidate; person with
> most votes wins) tend to not score so well. In particular, traditional systems work very poorly when there are large numbers of similar candidates, which is most likely what we'll face in selecting the CAB.
Let me introduce that we will have something between 70 and 100
voters for the Pilot part of the CAB. This is more than you may
easily survey without involving a formal voting process - except
when there are only 2-3 candidates.
I believe that we may really first try to find the rules.
If we would like to follow usual formal voting rules, we should
first try to create a voting commitee.
The first duty of this commite would be to find the names
of the candidates and the names of the permitted voters. Usual rules
forbid that a member of the voting commitee would become a
candidate. If we go this way, I believe that this is a godd idea
for our election too. So if we go this way, we would need some
volunteers for this job.
What we all need to find out in addition is the way we would like
to vote: open, secret, or whatever....
Of course if you believe that we will be able to find the right
way without a formal process, you are welcome to describe your ideas.
P.S. I don't like bureaucracy but it may help us to defend our election
later to non pilot members of the community.
Jörg
-- Original Message --
Subject: Re: [Opensolaris-discuss] CAB: voting system?
Date: Tue, 01 Feb 2005 12:04:12 -0800
From: Andrew Tucker <Andrew.Tucker@Sun.COM>
To: Bill Sommerfeld <sommerfeld@sun.com>
CC: James Grisanzio <Jim.Grisanzio@Sun.COM>, Andrew Tucker <tucker@trinity.sfbay.sun.com>
So, what do you recommend?
I'd be happy to let the external community decide how they're going to
select their representatives, but I don't think we can expect them
to come to a consensus on their own. We're going to need to put a
stake in the ground, at least for the initial election (presumably
the CAB can then determine rules for future elections).
Andy
On Sat, Jan 29, 2005 at 11:51:04AM -0500, Bill Sommerfeld wrote:
> I don't believe I've seen any discussion of the voting method to be used.
>
> If you list a bunch of desirable properties of a voting system,
> traditional voting systems used (vote for single candidate; person with
> most votes wins) tend to not score so well. In particular, traditional systems work very poorly when there are large numbers of similar candidates, which is most likely what we'll face in selecting the CAB.
>
> On the other hand, there are a bewildering array of alternative systems out there (including approval voting, preferential systems such as STV, and the
> bizarrely named "Cloneproof Schwartz Sequential Dropping"); some of these systems take some effort to understand and that by itself is a negative.
>
> As a perhaps extreme example of how this can hurt, my understanding is that
> New Zealand just changed to use a particularly complex hybrid scheme, with
> the result that in some precincts 10% or more of the ballots were
> uncountable.
>
> Anyhow, there are a bunch of writeups on the various systems out there;
> many are written by people with an axe to grind one way or another, so
> read skeptically.
>
> background reading:
>
> http://electionmethods.org/
> http://en.wikipedia.org/wiki/Arrow%27s_impossibility_theorem
> http://en.wikipedia.org/wiki/Category:Voting_systems
> http://en.wikipedia.org/wiki/Category:Voting_theory
>
> and, of course, there are open source software packages out there
> implementing some of these algorithms:
>
> http://stv.sourceforge.net/
>
> - Bill
>
-- Original Message --
Subject: Re: [Opensolaris-discuss] CAB: voting system?
Date: Tue, 01 Feb 2005 16:51:46 -0500
From: Bill Sommerfeld <sommerfeld@sun.com>
To: Andrew Tucker <Andrew.Tucker@sun.com>
CC: James Grisanzio <Jim.Grisanzio@Sun.COM>, Andrew Tucker <tucker@trinity.sfbay.sun.com>
On Tue, 2005-02-01 at 15:04, Andrew Tucker wrote:
> So, what do you recommend?
My preference is for approval voting.
You vote "yes" or "no" to every candidate, and add up all the "yes" votes.
Whoever gets the most "yes" votes wins (or the top two).
In the event of a tie, I'd break it randomly but allow for one to yield to the other.
Very simple to explain and understand; does not require multiple rounds of ballot redistribution or graph theory to come up with the winner.
downside is that you don't collect any information about relative preference
among the approved candidates.
you probably avoid seating someone that a significant minority hate.
(probably good for consensus building)
you may get more compromise candidates than favorites.
-- Original Message --
Subject: Re: [Opensolaris-discuss] CAB: voting system?
Date: Tue, 01 Feb 2005 14:29:48 -0800
From: Andrew Tucker <Andrew.Tucker@sun.com>
To: Bill Sommerfeld <sommerfeld@sun.com>
CC: James Grisanzio <Jim.Grisanzio@Sun.COM>, Andrew Tucker <tucker@trinity.sfbay.sun.com>
On Tue, Feb 01, 2005 at 04:51:46PM -0500, Bill Sommerfeld wrote:
> On Tue, 2005-02-01 at 15:04, Andrew Tucker wrote:
> > So, what do you recommend?
>
> My preference is for approval voting.
Sounds good to me (after a bit of research). I'll think about how
to suggest it to the pilot folks.
Andy
-- Original Message --
Subject: Re: [Opensolaris-discuss] CAB: voting system?
Date: Wed, 02 Feb 2005 14:06:15 -0800
From: Andrew Tucker <Andrew.Tucker@Sun.COM>
To: Joerg Schilling <schilling@fokus.fraunhofer.de>
CC: sommerfeld@Sun.COM, Jim.Grisanzio@Sun.COM, opensolaris-discuss@opensolaris.org
On Mon, Jan 31, 2005 at 03:31:11PM +0100, Joerg Schilling wrote:
> Bill Sommerfeld <sommerfeld@sun.com> wrote:
> > I don't believe I've seen any discussion of the voting method to be used.
>
> If we would like to follow usual formal voting rules, we should
> first try to create a voting commitee.
Although I think that's a reasonable idea, I'm worried about the
time it will take. Also you still have a bootstrapping problem -
how do you select members of the voting committee? In any case,
you can look at the initial CAB as (at least in part) being that
committee -
rules for the ongoing board, including voting rules.
Ideally, we'd let you (the pilot community) decide how you're going
to select your representatives for the initial CAB. On the other
hand, it's difficult to get consensus among 80 people. In order
to bootstrap, I'd like to suggest a few basic rules and process
steps.
Once the nomination period is over, Jim will be contacting the
nominees to make sure they're willing to serve, and asking each
to provide a short bio. The bios will go up on the pilot site
prior to the start of voting.
In order to address the concerns Bill brought up, I suggest using
approval voting. This is relatively simple, and avoids issues with
similar candidates splitting votes (for those familiar with recent
US politics, think Nader in 2000). The basic idea is that each
voter makes a yes/no selection for each nominee -the winners- Original Message -
are those who receive the most "yes" votes. Approval voting is
used by the IEEE as well as a few of the American mathematical
societies (AMS, MAA, ASA). For more details see:
http:bcn.boulder.co.us/government/approvalvote/center.html
or just google for "Approval Voting".
The votes should be kept private, with only the results announced.
We'll send out ballots (probably just a text file with yes/no boxes
to fill in unless someone has a better idea) to individual pilot
members, with the reply-to field set to an alias set up for the
purpose. Once the voting period is over, we'll count the ballots,
make sure there's only (at most) one per member, and post the
results. Only a few people within the OpenSolaris team at Sun
will see the raw ballots, and will of course keep the information
about how individuals voted confidential.
Comments? Concerns?
Andy
--
Subject: Re: [Opensolaris-discuss] CAB: voting system?
Date: Wed, 02 Feb 2005 14:06:15 -0800
From: Andrew Tucker <Andrew.Tucker@Sun.COM>
To: Joerg Schilling <schilling@fokus.fraunhofer.de>
CC: sommerfeld@Sun.COM, Jim.Grisanzio@Sun.COM, opensolaris-discuss@opensolaris.org
On Mon, Jan 31, 2005 at 03:31:11PM +0100, Joerg Schilling wrote:
> Bill Sommerfeld <sommerfeld@sun.com> wrote:
> > I don't believe I've seen any discussion of the voting method to be used.
>
> If we would like to follow usual formal voting rules, we should
> first try to create a voting commitee.
Although I think that's a reasonable idea, I'm worried about the
time it will take. Also you still have a bootstrapping problem -
how do you select members of the voting committee? In any case,
you can look at the initial CAB as (at least in part) being that
committee -part of its role will be to define the format andthe winners
rules for the ongoing board, including voting rules.
Ideally, we'd let you (the pilot community) decide how you're going
to select your representatives for the initial CAB. On the other
hand, it's difficult to get consensus among 80 people. In order
to bootstrap, I'd like to suggest a few basic rules and process
steps.
Once the nomination period is over, Jim will be contacting the
nominees to make sure they're willing to serve, and asking each
to provide a short bio. The bios will go up on the pilot site
prior to the start of voting.
In order to address the concerns Bill brought up, I suggest using
approval voting. This is relatively simple, and avoids issues with
similar candidates splitting votes (for those familiar with recent
US politics, think Nader in 2000). The basic idea is that each
voter makes a yes/no selection for each nominee -
are those who receive the most "yes" votes. Approval voting is
used by the IEEE as well as a few of the American mathematical
societies (AMS, MAA, ASA). For more details see:
http:bcn.boulder.co.us/government/approvalvote/center.html
or just google for "Approval Voting".
The votes should be kept private, with only the results announced.
We'll send out ballots (probably just a text file with yes/no boxes
to fill in unless someone has a better idea) to individual pilot
members, with the reply-to field set to an alias set up for the
purpose. Once the voting period is over, we'll count the ballots,
make sure there's only (at most) one per member, and post the
results. Only a few people within the OpenSolaris team at Sun
will see the raw ballots, and will of course keep the information
about how individuals voted confidential.
Comments? Concerns?
Andy
-- Original Message --
Subject: Re[2]: [Opensolaris-discuss] CAB: voting system?
Date: Wed, 02 Feb 2005 23:32:24 +0100
From: Robert Milkowski <rmilkowski@task.gda.pl>
Reply-To: Robert Milkowski <rmilkowski@task.gda.pl>
Organization: CI TASK http://www.task.gda.pl
To: Andrew Tucker <Andrew.Tucker@sun.com>
CC: Joerg Schilling <schilling@fokus.fraunhofer.de>, Jim.Grisanzio@Sun.COM, opensolaris-discuss@opensolaris.org, sommerfeld@sun.com
Hello Andrew,
Wednesday, February 2, 2005, 11:06:15 PM, you wrote:
AT> On Mon, Jan 31, 2005 at 03:31:11PM +0100, Joerg Schilling wrote:
>> Bill Sommerfeld <sommerfeld@sun.com> wrote:
>> > I don't believe I've seen any discussion of the voting method to be used.
>>
>> If we would like to follow usual formal voting rules, we should
>> first try to create a voting commitee.
AT> Although I think that's a reasonable idea, I'm worried about the
AT> time it will take. Also you still have a bootstrapping problem -
AT> how do you select members of the voting committee? In any case,
AT> you can look at the initial CAB as (at least in part) being that
AT> committee -
AT> rules for the ongoing board, including voting rules.
AT> Ideally, we'd let you (the pilot community) decide how you're going
AT> to select your representatives for the initial CAB. On the other
AT> hand, it's difficult to get consensus among 80 people. In order
AT> to bootstrap, I'd like to suggest a few basic rules and process
AT> steps.
AT> Once the nomination period is over, Jim will be contacting the
AT> nominees to make sure they're willing to serve, and asking each
AT> to provide a short bio. The bios will go up on the pilot site
AT> prior to the start of voting.
AT> In order to address the concerns Bill brought up, I suggest using
AT> approval voting. This is relatively simple, and avoids issues with
AT> similar candidates splitting votes (for those familiar with recent
AT> US politics, think Nader in 2000). The basic idea is that each
AT> voter makes a yes/no selection for each nominee -the winners- Original Message -
AT> are those who receive the most "yes" votes. Approval voting is
AT> used by the IEEE as well as a few of the American mathematical
AT> societies (AMS, MAA, ASA). For more details see:
AT> http://bcn.boulder.co.us/government/approvalvote/center.html
AT> or just google for "Approval Voting".
AT> The votes should be kept private, with only the results announced.
AT> We'll send out ballots (probably just a text file with yes/no boxes
AT> to fill in unless someone has a better idea) to individual pilot
AT> members, with the reply-to field set to an alias set up for the
AT> purpose. Once the voting period is over, we'll count the ballots,
AT> make sure there's only (at most) one per member, and post the
AT> results. Only a few people within the OpenSolaris team at Sun
AT> will see the raw ballots, and will of course keep the information
AT> about how individuals voted confidential.
AT> Comments? Concerns?
Easy and simple - imho it's definitely reasonable for first CAB.
But all nominated candidates should speak now if they want or not the
job. I don't see any reason why we should wait. It's because if one of
my favorite decide not to take the job I prefer to know it before
nominations end so there's still a time to potential new nomination in
that case.
--
Best regards,
Robert rmilkowski@task.gda.pl
--
Subject: Re: Re[2]: [Opensolaris-discuss] CAB: voting system?
Date: Wed, 02 Feb 2005 23:33:22 +0100
From: Joerg Schilling <schilling@fokus.fraunhofer.de>
To: rmilkowski@task.gda.pl, Andrew.Tucker@sun.com
CC: Jim.Grisanzio@Sun.COM, opensolaris-discuss@opensolaris.org
Robert Milkowski <rmilkowski@task.gda.pl> wrote:
> But all nominated candidates should speak now if they want or not the
> job. I don't see any reason why we should wait. It's because if one of
> my favorite decide not to take the job I prefer to know it before
> nominations end so there's still a time to potential new nomination in
> that case.
Did someone create a complete list?
It would be nice if this list could be mailed to us.
Jörg
- -
XWiki Enterprise 2.7.1.34853 - Documentation
Terms of Use
|
Privacy
|
Trademarks
|
Copyright Policy
|
Site Guidelines
|
Site Map
|
Help
Your use of this web site or any of its content or software indicates your agreement to be bound by these Terms of Use.
© 2012, Oracle Corporation and/or its affiliates.