Re: Election results

The election for board members and the results are as follows:

The following members are appointed to the SPI board:
Bdale Garbee
David Graham
Luk Claes
Joerg Jaspert
Joshua D. Drake
Martin Zobel-Helas

The following members are thanked for running:
Ian Jackson
Christel Dahlskjaer
Rico Gloeckner
Richard Hartmann
Martin 'Joey' Schulze
Robert Brockway
MJ Ray

95 votes were received.

For the detailed results, please see:
http://www.spi-inc.org/secretary/votes/vote6/results.html
--
Neil McGovern
Secretary, Software in the Public Interest, Inc.

On Sun, Jul 29, 2007 at 01:20:53AM +0100, Neil McGovern wrote:
> For the detailed results, please see:
> http://www.spi-inc.org/secretary/votes/vote6/results.html

The results there seem a bit odd -- in particular, Martin and Ian
had a pairwise tie (exactly as many people ranked Martin above Ian as
vice-versa), so if there's a tie for position four between Joerg and
Joshua, afaics there should be a tie for position six between Martin
and Ian.

Running the numbers myself gave the following results after eliminating
Bdale, David, and Luk:

] Joerg (C)
] 3 beats 6: 42 40 Joerg beats Josh
] 3 beats 8: 51 32 Joerg beats Ian
] Josh (F)
] 6 beats 1: 50 34 Josh beats Martin
] 6 beats 8: 49 32 Josh beats Ian
] Martin (A)
] 1 beats 3: 44 41 Martin beats Joerg
] Ian (H)
]
] (each of the above defeats all other candidates)

That's a circular tie between Joerg, Josh and Martin for first place. The
strengths of the strongest beatpaths are:

Joerg -> Josh 42 votes
Joerg -> Josh -> Martin 42 votes

Josh -> Martin 50 votes
Josh -> Martin -> Joerg 44 votes

Martin -> Joerg 44 votes
Martin -> Joerg -> Josh 42 votes

Josh disqualifies Martin (50 vs 42) and Joerg (44 vs 42), breaking the
tie and leaving Josh as the winner of fourth place. This is using the
description from [0] as the definition of Cloneproof SSD -- which is
equivalent to the Debian method.

Eliminating Josh, then gives

] Joerg (C)
] 3 beats 8: 51 32 Joerg beats Ian
] Martin (A)
] 1 beats 3: 44 41 Martin beats Joerg
] Ian (H)

with Martin taking fifth place and Joerg sixth, with a complete ordering
for seventh to thirteenth place.

Cheers,
aj

[0] http://en.wikipedia.org/wiki/Schulze_method

On Sun, Jul 29, 2007 at 01:03:10PM +1000, Anthony Towns wrote:
> Running the numbers myself gave the following results [...]

Oh, I tried it with Single Transferable Vote (which is a multi-winner
counting method used for the Australian Senate and a bunch of other
places). The outcome in that case is:

Bdale elected ( 1st)
Joshua and David elected (2nd and 3rd)
Rico eliminated (13th)
Richard eliminated (12th)
Robert eliminated (11th)
Martin Zobel eliminated (10th)
MJ eliminated ( 9th)
Christel eliminated ( 8th)
Luk elected ( 4th)
Ian elected ( 5th)
Joerg elected ( 6th)
Joey not elected ( 7th)

The difference between STV and the system we're using is that it tries to
make each vote only count for one slot -- so all the people who elected
Bdale in the first round have their votes weighted down.

For SPI's case, we probably have a majority of people who're familiar
with Debian, so if they all rank the Debian folks they're familiar with
first, no one else will get a chance, because they're a majority for
the first slot and remain a majority all the way through to the last slot.

That the only practical difference is between Martin and Ian getting
elected probably says good things about the general level of consistency
amongst the electors, particularly given they had a pairwise tie anyway,.

Cheers,
aj

Neil McGovern wrote: [...]
> The following members are appointed to the SPI board:
> Bdale Garbee
> David Graham
> Luk Claes
> Joerg Jaspert
> Joshua D. Drake
> Martin Zobel-Helas

Well done to the new members. When do you meet for the first time?

However, it's interesting that I think both first-past-the-post and
alternative vote (also known as instant run-off voting, reportedly
recommended by Robert's Rules for election-by-mail) would have
resulted in this same board:

1. Bdale Garbee
2. David Graham
3. Joshua D. Drake
4. Martin 'Joey' Schulze
5. Luk Claes
6. MJ Ray

More generally, is SPI's type of Condorcet ever likely to elect
someone who polarises views, or who many inexplicably dislike? What
does this say for any plan to use a Condorcet for debian's social
committee? Could majorities always prevent minority reps? For SPI,
does this mean that candidates from smaller projects are at the mercy
of the debian project?

(More commentary and navel-gazing illustration at
http://mjr.towers.org.uk/blog/2007/spi#resultelections
if you want it.)

> 95 votes were received.

Finally, as I understand it, turn-out was 25% of voting members (not
the 25% of SPI members that some press reported). Why was turn-out so
low (at least, I think it's low for a vote-by-mail in this type of
organisation) and what steps will SPI be taking to increase it?

Thanks and best wishes,
--
MJ Ray - see/vidu http://mjr.towers.org.uk/email.html
Experienced webmaster-developers for hire http://www.ttllp.co.uk/
Also: statistician, sysadmin, online shop builder, workers co-op.
Writing on koha, debian, sat TV, Kewstoke http://mjr.towers.org.uk/

On Tue, Aug 07, 2007 at 11:58:27AM +0100, MJ Ray wrote:
> More generally, is SPI's type of Condorcet ever likely to elect
> someone who polarises views, or who many inexplicably dislike? What
> does this say for any plan to use a Condorcet for debian's social
> committee? Could majorities always prevent minority reps? For SPI,
> does this mean that candidates from smaller projects are at the mercy
> of the debian project?

As I'm sure you know - as it was discussed during the soc ctte debates
at Debian - the election method used here fails proportionality.

--
Antti-Juhani Kaijanaho, Jyväskylä
http://antti-juhani.kaijanaho.fi/newblog/
http://www.flickr.com/photos/antti-juhani/

On Tue, Aug 07, 2007 at 11:58:27AM +0100, MJ Ray wrote:
> However, it's interesting that I think both first-past-the-post and
> alternative vote (also known as instant run-off voting, reportedly
> recommended by Robert's Rules for election-by-mail) would have
> resulted in this same board:
>

I've done a few others (Borda, Borda Elimination, Minmax, Nanson, Ranked
Pairs, Condorcet (SPI), Condorcet (Debian), Bucklin, IRV, Pluralty) at
http://blog.halon.org.uk/geek/spi/2007-election-results-calcs.html
for those interested

> > 95 votes were received.
>
> Finally, as I understand it, turn-out was 25% of voting members (not
> the 25% of SPI members that some press reported). Why was turn-out so
> low (at least, I think it's low for a vote-by-mail in this type of
> organisation) and what steps will SPI be taking to increase it?
>

A review of who is a contributing member will be underway shortly.

Neil
--
Neil McGovern
Secretary, Software in the Public Interest, Inc.

On Tue, Aug 07, 2007 at 02:29:48PM +0100, Neil McGovern wrote:
> I've done a few others (Borda, Borda Elimination, Minmax, Nanson, Ranked
> Pairs, Condorcet (SPI), Condorcet (Debian), Bucklin, IRV, Pluralty) at
> http://blog.halon.org.uk/geek/spi/2007-election-results-calcs.html
> for those interested

I'd be interested in how Schultze's STV method (a generalisation of the
Debian method for a multiple winner situation) handles this election.

See http://home.versanet.de/~chris1-schulze/schulze2.pdf for a
description and discussion of the method and
http://home.versanet.de/~chris1-schulze/schulze3.zip for some source
code (which I tried to get to work but lost patience with).

--
Antti-Juhani Kaijanaho, Jyväskylä
http://antti-juhani.kaijanaho.fi/newblog/
http://www.flickr.com/photos/antti-juhani/

Antti-Juhani Kaijanaho <antti-juhani(at)kaijanaho(dot)fi> wrote:
> As I'm sure you know - as it was discussed during the soc ctte debates
> at Debian - the election method used here fails proportionality.

Why be sure? Assumption is the mother of all ... To be clear, I
didn't know Condorcet-SPI was already discussed - I'm not very good at
reading political intricacies (I'm much more interested in the effects
than the theory, as long as effects don't include obvious bugs), so
maybe I didn't recognise it. Also, I'd forget my head if it weren't
screwed on, I seldom read all of those debates and I don't remember
seeing a summary. (note 1)

Where's that discussion, please?

Anyway, should/how could we try to fix this proportionality failure?

Also, anyone want to try something like Principal Components Analysis
on common aspects of the platforms?

(note 1: Searching a while found the email archived at
http://lists.debian.org/debian-project/2007/06/msg00318.html
which mentions "another mail" but nothing more to find it!
Other mails that month don't look obviously about SPI's method, like
http://lists.debian.org/debian-project/2007/06/msg00261.html
A search for mails about SPI from author kaijanaho found a 1999 one.)

Regards,
--
MJ Ray - see/vidu http://mjr.towers.org.uk/email.html
Experienced webmaster-developers for hire http://www.ttllp.co.uk/
Also: statistician, sysadmin, online shop builder, workers co-op.
Writing on koha, debian, sat TV, Kewstoke http://mjr.towers.org.uk/

Neil McGovern <neilm(at)spi-inc(dot)org> wrote:
> I've done a few others (Borda, Borda Elimination, Minmax, Nanson, Ranked
> Pairs, Condorcet (SPI), Condorcet (Debian), Bucklin, IRV, Pluralty) at
> http://blog.halon.org.uk/geek/spi/2007-election-results-calcs.html
> for those interested

Thanks. I only did the ones that were obvious/easy how to do
manually. Could you link the software you used, please?

> > > 95 votes were received.
> >
> > Finally, as I understand it, turn-out was 25% of voting members (not
> > the 25% of SPI members that some press reported). Why was turn-out so
> > low (at least, I think it's low for a vote-by-mail in this type of
> > organisation) and what steps will SPI be taking to increase it?
>
> A review of who is a contributing member will be underway shortly.

Does that mean you suspect the turn-out was low because there are too
many voters and SPI will try to remove/reclassify voting members to
improve the turn-out proportion?

Thanks,
--
MJ Ray - see/vidu http://mjr.towers.org.uk/email.html
Experienced webmaster-developers for hire http://www.ttllp.co.uk/
Also: statistician, sysadmin, online shop builder, workers co-op.
Writing on koha, debian, sat TV, Kewstoke http://mjr.towers.org.uk/

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

MJ Ray wrote:
> Neil McGovern wrote: [...]

> More generally, is SPI's type of Condorcet ever likely to elect
> someone who polarises views, or who many inexplicably dislike? What
> does this say for any plan to use a Condorcet for debian's social
> committee? Could majorities always prevent minority reps? For SPI,
> does this mean that candidates from smaller projects are at the mercy
> of the debian project?

No I believe they are at the mercy of what the majority feel are
reasonable views.

Note that I have zero, nadda, zip to do with the Debian project, yet I
was still elected.

Sincerely,

Joshua D. Drake

- --

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGuIueATb/zqfZUUQRAoUrAJ924VcFuTgZL67BCF1DKZGmd7xCcwCdHUHL
xt4dS7VGGEUdZKTCyDGK7vc=
=VvOD
-----END PGP SIGNATURE-----

On Tue, Aug 07, 2007 at 02:57:46PM +0100, MJ Ray wrote:
> Neil McGovern <neilm(at)spi-inc(dot)org> wrote:
> > I've done a few others (Borda, Borda Elimination, Minmax, Nanson, Ranked
> > Pairs, Condorcet (SPI), Condorcet (Debian), Bucklin, IRV, Pluralty) at
> > http://blog.halon.org.uk/geek/spi/2007-election-results-calcs.html
> > for those interested
>
> Thanks. I only did the ones that were obvious/easy how to do
> manually. Could you link the software you used, please?
>

Sure, http://vote.sourceforge.net/

> > > > 95 votes were received.
> > >
> > > Finally, as I understand it, turn-out was 25% of voting members (not
> > > the 25% of SPI members that some press reported). Why was turn-out so
> > > low (at least, I think it's low for a vote-by-mail in this type of
> > > organisation) and what steps will SPI be taking to increase it?
> >
> > A review of who is a contributing member will be underway shortly.
>
> Does that mean you suspect the turn-out was low because there are too
> many voters and SPI will try to remove/reclassify voting members to
> improve the turn-out proportion?
>

Yes, as per the by-law's renewal requirements. This has been mentioned
this month in:
<20070713214308(dot)GB2793(at)techhouse(dot)org>
<Pine(dot)LNX(dot)4(dot)55(dot)0707161607100(dot)5347(at)baffin>
if you need further clarification.

Neil
--
Neil McGovern
Secretary, Software in the Public Interest, Inc.

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

MJ Ray wrote:
> Neil McGovern <neilm(at)spi-inc(dot)org> wrote:
>> I've done a few others (Borda, Borda Elimination, Minmax, Nanson, Ranked

> Does that mean you suspect the turn-out was low because there are too
> many voters and SPI will try to remove/reclassify voting members to
> improve the turn-out proportion?

I think it means that we have a lot of contributing members who don't
contribute and thus we will likely remove their contributing status.

Joshua D. Drake

>
> Thanks,

- --

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGuIw6ATb/zqfZUUQRAiDFAJwKf9ARI+U6wTsTSHuGUGK/WEz05QCffb9F
KhY41cBgN/B/2pTn6NDBZvo=
=L+GA
-----END PGP SIGNATURE-----

On 07-Aug-07, 10:14 (CDT), "Joshua D. Drake" <jd(at)commandprompt(dot)com> wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> MJ Ray wrote:
> > Neil McGovern <neilm(at)spi-inc(dot)org> wrote:
> >> I've done a few others (Borda, Borda Elimination, Minmax, Nanson, Ranked
>
> > Does that mean you suspect the turn-out was low because there are too
> > many voters and SPI will try to remove/reclassify voting members to
> > improve the turn-out proportion?
>
> I think it means that we have a lot of contributing members who don't
> contribute and thus we will likely remove their contributing status.

So the only way to contribute is to vote for SPI board members? :-)

Actually, I'll use this opportunity to ask a question: how does one
resign from SPI? There's no info on the website(s) that I could find.
And for the record, this has nothing to do with Joshua's comment; I
realized during the lengthy pre-election discussions that I had no
reason to remain a member.

Steve

--
Steve Greenland
The irony is that Bill Gates claims to be making a stable operating
system and Linus Torvalds claims to be trying to take over the
world. -- seen on the net

On Tue, Aug 07, 2007 at 10:19:18AM -0500, Steve Greenland wrote:
> So the only way to contribute is to vote for SPI board members? :-)

No. For example, current members of an SPI associated project such as
Debian or PostgreSQL who respond to an inquiry about wanting to remain
an SPI contributing member (similar to the one that Debian did recently)
will remain contributing members. I'd expect it will be similar for
other contributing members who meet the membership criteria in other
ways. Note that none of this is speaking for the current board or the
current membership committee, since I am on neither, but it is my
understanding of the plans as of the last time I checked on them with
the then-current membership committee while I was on board.

- Jimmy Kaplowitz
jimmy(at)spi-inc(dot)org

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Steve Greenland wrote:
> On 07-Aug-07, 10:14 (CDT), "Joshua D. Drake" <jd(at)commandprompt(dot)com> wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> MJ Ray wrote:
>>> Neil McGovern <neilm(at)spi-inc(dot)org> wrote:
>>>> I've done a few others (Borda, Borda Elimination, Minmax, Nanson, Ranked
>>> Does that mean you suspect the turn-out was low because there are too
>>> many voters and SPI will try to remove/reclassify voting members to
>>> improve the turn-out proportion?
>> I think it means that we have a lot of contributing members who don't
>> contribute and thus we will likely remove their contributing status.
>
> So the only way to contribute is to vote for SPI board members? :-)

No but I could easily argue that there is an implicit responsibility to
"maintain" your membership to SPI which would indeed include voting.

I will also concede that there is no explicit statement to this effect
which causes me to want to write up a motion to insure that we put it
into place.

Sincerely,

Joshua D. Drake

- --

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGuJGwATb/zqfZUUQRArHqAJ9FtoB99pxX2Rm7hGnJAPc6b2qEDwCfVrFO
g5allDxXkMyw1OhZF/600w8=
=g0HB
-----END PGP SIGNATURE-----

Neil McGovern <neilm(at)spi-inc(dot)org> wrote:
> Yes, as per the by-law's renewal requirements. This has been mentioned
> this month in:
> <20070713214308(dot)GB2793(at)techhouse(dot)org>
> <Pine(dot)LNX(dot)4(dot)55(dot)0707161607100(dot)5347(at)baffin>
> if you need further clarification.

Last month? I think the first one is in
http://lists.spi-inc.org/pipermail/spi-general/2007-July.txt.gz
but where's the second? (Do we have a nicer way to look up message-id?)

Thanks again,
--
MJ Ray - see/vidu http://mjr.towers.org.uk/email.html
Experienced webmaster-developers for hire http://www.ttllp.co.uk/
Also: statistician, sysadmin, online shop builder, workers co-op.
Writing on koha, debian, sat TV, Kewstoke http://mjr.towers.org.uk/

On Tue, Aug 07, 2007 at 10:19:18AM -0500, Steve Greenland wrote:
> On 07-Aug-07, 10:14 (CDT), "Joshua D. Drake" <jd(at)commandprompt(dot)com> wrote:
> > -----BEGIN PGP SIGNED MESSAGE-----
> > Hash: SHA1
> >
> > MJ Ray wrote:
> > > Neil McGovern <neilm(at)spi-inc(dot)org> wrote:
> > >> I've done a few others (Borda, Borda Elimination, Minmax, Nanson, Ranked
> >
> > > Does that mean you suspect the turn-out was low because there are too
> > > many voters and SPI will try to remove/reclassify voting members to
> > > improve the turn-out proportion?
> >
> > I think it means that we have a lot of contributing members who don't
> > contribute and thus we will likely remove their contributing status.
>
> So the only way to contribute is to vote for SPI board members? :-)
>
> Actually, I'll use this opportunity to ask a question: how does one
> resign from SPI? There's no info on the website(s) that I could find.
> And for the record, this has nothing to do with Joshua's comment; I
> realized during the lengthy pre-election discussions that I had no
> reason to remain a member.
>

"Members wishing to leave the organization may do so at any time by
notifying the membership committee in writing of their desire for the
membership to be terminated."

aka: mail the membership committee :)

Neil
--
Neil McGovern
Secretary, Software in the Public Interest, Inc.

On Tue, Aug 07, 2007 at 08:37:20AM -0700, Joshua D. Drake wrote:
> > So the only way to contribute is to vote for SPI board members? :-)
>
> No but I could easily argue that there is an implicit responsibility to
> "maintain" your membership to SPI which would indeed include voting.
>
> I will also concede that there is no explicit statement to this effect
> which causes me to want to write up a motion to insure that we put it
> into place.
>

I don't think there's a need for a motion here, see article 3 of the
by-laws.

Neil
--
Neil McGovern
Secretary, Software in the Public Interest, Inc.

On Tue, Aug 07, 2007 at 02:51:24PM +0100, MJ Ray wrote:
> Antti-Juhani Kaijanaho <antti-juhani(at)kaijanaho(dot)fi> wrote:
> > As I'm sure you know - as it was discussed during the soc ctte debates
> > at Debian - the election method used here fails proportionality.
>
> Why be sure? Assumption is the mother of all ... To be clear, I
> didn't know Condorcet-SPI was already discussed

I apologise for the assumption :)

As far as I can tell, the SPI method and the Debian method are similar
enough that the issue is common to both. Of course, I can't find
anywhere a precise definition of the SPI method for *multi-winner*
elections, but the result page makes me believe it's what I'd expect.

I'll revisit if you'll point me to the definition of the multi-winner
SPI method :)

> Where's that discussion, please?

The messages you mentioned were the ones I was referring to.

> Anyway, should/how could we try to fix this proportionality failure?

My current favourite is the Schulze STV method that generalizes the
classic Schulze ("cloneproof Schwarz sequential dropping") method to
multiple-winner elections. It is claimed to satisfy proportionality,
though I haven't bothered checking the argument myself.

--
Antti-Juhani Kaijanaho, Jyväskylä
http://antti-juhani.kaijanaho.fi/newblog/
http://www.flickr.com/photos/antti-juhani/

On Tue, Aug 07, 2007 at 05:05:01PM +0100, Neil McGovern wrote:
> "Members wishing to leave the organization may do so at any time by
> notifying the membership committee in writing of their desire for the
> membership to be terminated."
>
> aka: mail the membership committee :)

To clarify how to do this: membership(at)spi-inc(dot)org

- Jimmy Kaplowitz
jimmy(at)spi-inc(dot)org

On 07-Aug-07, 11:05 (CDT), Neil McGovern <neilm(at)spi-inc(dot)org> wrote:
> "Members wishing to leave the organization may do so at any time by
> notifying the membership committee in writing of their desire for the
> membership to be terminated."
>
> aka: mail the membership committee :)

Thanks.

You might want to add that to http://www.spi-inc.org/about-spi/membership,
which seems the obvious place to look.

Steve

--
Steve Greenland
The irony is that Bill Gates claims to be making a stable operating
system and Linus Torvalds claims to be trying to take over the
world. -- seen on the net

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Neil McGovern wrote:
> On Tue, Aug 07, 2007 at 08:37:20AM -0700, Joshua D. Drake wrote:
>>> So the only way to contribute is to vote for SPI board members? :-)
>> No but I could easily argue that there is an implicit responsibility to
>> "maintain" your membership to SPI which would indeed include voting.
>>
>> I will also concede that there is no explicit statement to this effect
>> which causes me to want to write up a motion to insure that we put it
>> into place.
>>
>
> I don't think there's a need for a motion here, see article 3 of the
> by-laws.

Agreed from article 3:

Membership in this organization shall be open to all who meet the
approval of the membership committee. Members have the right and
responsibility of overseeing the board members, officers, and committees
and ensuring that they operate in accordance with the goals and
principles of the organization. All board members, officers, and
committees are ultimately responsible to the membership, and should act
in accordance with its wishes.

Contributing and non-contributing memberships are available within this
organization. Members agree to support the principles and help further
the goals of this organization. All membership applications and renewals
will be reviewed by the Membership Committee. Applicants must meet the
criteria set forth in the charter of the Membership Committee.

Members may be individuals or organizations.

Members wishing to leave the organization may do so at any time by
notifying the membership committee in writing of their desire for the
membership to be terminated.

Thus it isn't even implicit. :) Excellent.

Sincerely,

Joshua D. Drake

>
> Neil
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Spi-general mailing list
> Spi-general(at)lists(dot)spi-inc(dot)org
> http://lists.spi-inc.org/listinfo/spi-general

- --

=== The PostgreSQL Company: Command Prompt, Inc. ===
Sales/Support: +1.503.667.4564 || 24x7/Emergency: +1.800.492.2240
Providing the most comprehensive PostgreSQL solutions since 1997
http://www.commandprompt.com/

Donate to the PostgreSQL Project: http://www.postgresql.org/about/donate
PostgreSQL Replication: http://www.commandprompt.com/products/

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFGuJ+xATb/zqfZUUQRArykAJ9HL0Hj+2qXdgGPiuRAvjVWR3uLmACggHa+
cEiOMxwjUoP1rAdP09VidP4=
=Ao4R
-----END PGP SIGNATURE-----

Antti-Juhani Kaijanaho <antti-juhani(at)kaijanaho(dot)fi> wrote:
> [...] Of course, I can't find
> anywhere a precise definition of the SPI method for *multi-winner*
> elections, but the result page makes me believe it's what I'd expect.
>
> I'll revisit if you'll point me to the definition of the multi-winner
> SPI method :)

The one I found at https://members.spi-inc.org/vote/election.php?id=6
says:

Votes will be counted using the "Condorcet" election method system
which will be used to select the most preferred candidate.
Conceptually, the election will be broken into a a series of
pairwise races between each possible paring of the candidates. If
one candidate beats each of the others in pairwise races, that
candidate wins. Otherwise, the "Cloneproof/Schwartz Sequential
Dropping" method is employed to choose the most prefered candidate
from those remaining. Some background reading on preferential
voting and Condorcet is available online.

As such, you are being asked to vote your preferences by ranking
the candidates. Your preference will best be shown when you
specify all canadiates, in order of preference (i.e. "XYZWTUV", X
being the most preferred). Voting "XY" states that you prefer
candidate X to Y. However, it does not express any preference
relationships for any of the other candidates.

Is that the same one discussed in
http://lists.debian.org/debian-project/2007/06/msg00261.html
? (= is Jacobo Tarrio describing C/SSD? I think)

So, this is non-proportional?

[...]
> My current favourite is the Schulze STV method that generalizes the
> classic Schulze ("cloneproof Schwarz sequential dropping") method to
> multiple-winner elections. It is claimed to satisfy proportionality,
> though I haven't bothered checking the argument myself.

Noted. Will any voting expert tell us how Schulze STV would result?

Thanks,
--
MJ Ray - see/vidu http://mjr.towers.org.uk/email.html
Experienced webmaster-developers for hire http://www.ttllp.co.uk/
Also: statistician, sysadmin, online shop builder, workers co-op.
Writing on koha, debian, sat TV, Kewstoke http://mjr.towers.org.uk/

On Tue, Aug 07, 2007 at 06:09:50PM +0100, MJ Ray wrote:
> The one I found at https://members.spi-inc.org/vote/election.php?id=6
> says:

Yes, I found that too. However, it describes the single-winner method
and leaves out the crucial point: how the single-winner method is
generalized to a multiple-winner method.

There are two obvious methods:

1. The single-winner method produces a beat graph
(http://www.spi-inc.org/secretary/votes/vote6/vote_006_results.png),
which is (after tiebreaking) a total ordering of the candidates. Take
as many winners from the top as you need.

2. The single-winner method is used to obtain a winner. The winner is
dropped from the list of candidates, and the procedure is rerun
until one has enough winners.

Both methods are, as far as I can see, nonproportional. (The methods
might actually create the same result, but I'm too lazy to check that.)
If the voters are split into two factions, one of which is ascendant
over the another, then the majority faction can force the new board
members to all be majority candidates, and the minority faction gets
nothing. The way to do it? Supporters of the majority faction just
need to rank all majority candidates ahead of all minority candidates.

> Is that the same one discussed in
> http://lists.debian.org/debian-project/2007/06/msg00261.html
> ? (= is Jacobo Tarrio describing C/SSD? I think)

The one I discussed there is Method 2 above.

--
Antti-Juhani Kaijanaho, Jyväskylä
http://antti-juhani.kaijanaho.fi/newblog/
http://www.flickr.com/photos/antti-juhani/

On Tue, Aug 07, 2007 at 02:51:24PM +0100, MJ Ray wrote:
> Anyway, should/how could we try to fix this proportionality failure?

I don't believe Condorcet of any stripe is intended to fix the
proportionality you seem to be interested in. My dim recollection of
Condorcet, from the days when I studied voting systems > 10 years ago,
was that it was intended to solve the problem where some (largish)
percentage of voters have _none_ of their preferences reflected. The
idea here is that voters are more willing to accept their second-best
(or third- or whatever-best) choice in preference to the thing they
regard as worst. This tends to mean that polarising choices are
always excluded, and that exclusion is regarded as a feature and not
a bug.

A

--
Andrew Sullivan | ajs(at)crankycanuck(dot)ca
Everything that happens in the world happens at some place.
--Jane Jacobs

Aha:

On Tue, Aug 07, 2007 at 06:09:50PM +0100, MJ Ray wrote:

> The one I found at https://members.spi-inc.org/vote/election.php?id=6
> says:
>
> Votes will be counted using the "Condorcet" election method system
> which will be used to select the most preferred candidate.

The bug is in the description, because it uses a question-begging
definition. Condorcet does not select the "most preferred"
candidate, unless you use a funny meaning of "most preferred". It
more accurately is described as a method to select the most agreed
upone candidate; that is, it tends to favour the choices that the
electors are willing to rank higher, but it does not guarantee
that the first choice of anyone wins.

A

--
Andrew Sullivan | ajs(at)crankycanuck(dot)ca
However important originality may be in some fields, restraint and
adherence to procedure emerge as the more significant virtues in a
great many others. --Alain de Botton

On Tue, Aug 07, 2007 at 01:57:50PM -0400, Andrew Sullivan wrote:
> I don't believe Condorcet of any stripe is intended to fix the
> proportionality you seem to be interested in.

Condorcet is a single-winner method. Proportionality is a nonsensical
requirement for single-winner elections.

--
Antti-Juhani Kaijanaho, Jyväskylä
http://antti-juhani.kaijanaho.fi/newblog/
http://www.flickr.com/photos/antti-juhani/

On Tue, Aug 07, 2007 at 09:37:05AM -0700, Joshua D. Drake wrote:
> >> No but I could easily argue that there is an implicit responsibility to
> >> "maintain" your membership to SPI which would indeed include voting.

> approval of the membership committee. Members have the right and
> responsibility of overseeing the board members, officers, and committees
> and ensuring that they operate in accordance with the goals and
> principles of the organization.

That doesn't say one must vote, though. Of course, I also happen to
think that not voting, in the absence of an "I abstain" or "I refuse
my ballot" option on the ballot, is a legitimate way of communicating
something. The problem, naturally, lies in understanding the message
so sent.

A

--
Andrew Sullivan | ajs(at)crankycanuck(dot)ca
The fact that technology doesn't work is no bar to success in the marketplace.
--Philip Greenspun

On Tue, Aug 07, 2007 at 03:12:04PM -0400, Andrew Sullivan wrote:
> That doesn't say one must vote, though. Of course, I also happen to
> think that not voting, in the absence of an "I abstain" or "I refuse
> my ballot" option on the ballot, is a legitimate way of communicating
> something. The problem, naturally, lies in understanding the message
> so sent.

Speaking only on my own behalf, I don't think that voting in SPI should
be compulsory for contributing members, and my reading of the bylaw does
not require members to vote. They do, however, have the responsibility
to oversee the board, but by parallel with other membership
organizations I would view that as a collective responsiblity instead of
an individual responsibility of every member. (Morally/ethically, every
member should individually oversee the board, but as far as a binding
responsibility imposed by the bylaws as opposed to a philosophically
grounded one, I view it as collective.)

- Jimmy Kaplowitz
jimmy(at)spi-inc(dot)org

On Tue, Aug 07, 2007 at 09:02:44PM +0300, Antti-Juhani Kaijanaho wrote:
>
> Condorcet is a single-winner method. Proportionality is a nonsensical
> requirement for single-winner elections.

Sure. But given that we have apparently adapted Condorcet rules for
multi-winners, my point was just that the whole point of Condorcet is
to optimise for some other set of conditions than proportionality of
preferences.

A

--
Andrew Sullivan | ajs(at)crankycanuck(dot)ca
The very definition of "news" is "something that hardly ever happens."
--Bruce Schneier

On Tue, Aug 07, 2007 at 04:01:25PM -0400, Andrew Sullivan wrote:
> Sure. But given that we have apparently adapted Condorcet rules for
> multi-winners, my point was just that the whole point of Condorcet is
> to optimise for some other set of conditions than proportionality of
> preferences.

Obviously. However, I believe that this is *not* because the need for
proportionality has been rejected, but because the need for
proportionality has not been considered before.

As recently as last year, I personally was an advocate of the
iterate-Condorcet technique, not having thought its implications
through.

--
Antti-Juhani Kaijanaho, Jyväskylä
http://antti-juhani.kaijanaho.fi/newblog/
http://www.flickr.com/photos/antti-juhani/

All,

Personally, given the tendency of some members of the SPI membership towards
extreme uncompromising views and interminable argument, that the Concordet
method of selecting for the "most acceptable" instead of the "most preferred"
is a feature and not a bug.

I don't buy the "Debian-dominated" argument because despite DDs being 75% of
our membership we currently have 3 non-DDs on the board. It would also be
trivial for any sufficiently motivated and organized project to get behind
one candidate from their project to ensure them a place on the board. And
the Liaisons have significant non-voting influence if they choose to exercise
it.

Overall, I don't see the problem here.

--
Josh Berkus
Treasurer
Software in the Public Interest, Inc.
www.spi-inc.org

Jimmy,

> Speaking only on my own behalf, I don't think that voting in SPI should
> be compulsory for contributing members, and my reading of the bylaw does
> not require members to vote.

Unfortunately, the quorum parts of our bylaws will paralyze the organization
if we do not purge "inactive" members. For example, changes to the bylaws
themselves require a majority of *all* members, not just voting ones. With
current levels of participation that will be impossible to achieve.

I'd bet you that at least 15% of our membership has gone no-mail on the lists
and is not even aware that they are still registered as members.

--
Josh Berkus
Board
Software in the Public Interest, Inc.
www.spi-inc.org

MJ Ray <mjr(at)phonecoop(dot)coop> wrote:
> Neil McGovern <neilm(at)spi-inc(dot)org> wrote: [...]
> > <Pine(dot)LNX(dot)4(dot)55(dot)0707161607100(dot)5347(at)baffin> [...]
> but where's the second? (Do we have a nicer way to look up message-id?)

(For the archives:) The second's on spi-private.
--
MJR/slef

Josh Berkus <josh(at)postgresql(dot)org> wrote:
> [...] For example, changes to the bylaws
> themselves require a majority of *all* members, not just voting ones. [...]

Ouch. It really does say just "members" not "contributing members".
That seems like a major bug to me, so I guess drastic action to fix it
is justified.

> I'd bet you that at least 15% of our membership has gone no-mail on the lists
> and is not even aware that they are still registered as members.

Hamish Moffatt wrote on my web site:
"Personally I did not vote as SPI does not seem particularly
relevant to Debian, other than as a money handler. In addition I admit
to being barely aware that the vote was on; despite being a
contributing member I was not subscribed to spi-announce until a few
moments ago. "
"Perhaps SPI could benefit from a members-announce mailing list
like debian-devel-announce."

Regards,
--
MJ Ray - see/vidu http://mjr.towers.org.uk/email.html
Experienced webmaster-developers for hire http://www.ttllp.co.uk/
Also: statistician, sysadmin, online shop builder, workers co-op.
Writing on koha, debian, sat TV, Kewstoke http://mjr.towers.org.uk/

Josh,

On Tue, Aug 07, 2007 at 11:36:42PM -0700, Josh Berkus wrote:

> Unfortunately, the quorum parts of our bylaws will paralyze the organization
> if we do not purge "inactive" members.

This argument begs the question. There is a difference between
inactive members who do not care about the organisation and don't
especially want to be members, and members who do care but choose not
to vote. Defining the latter as inactive deprives those refusing to
vote that option of communicating their indifference to the options
presented. It's a legitmate democratic tactic, particularly in the
absence of a "none of the above" option on the ballot.

> For example, changes to the bylaws themselves require a majority of
> *all* members, not just voting ones. With current levels of
> participation that will be impossible to achieve.

This is _also_ question-begging, because it assumes that the changes
should occur if they are desired. You simply do not know, in
advance, whether the non-voters are refusing to vote because they
don't care about the organisation, or because they're happy with the
bylaws as they are and don't think they need changing. If you know
that not voting is effectively the same as voting "no", why bother
going on the record?

Note that this isn't an argument against establishing criteria for
participation, but it is an argument for excluding a requirement to
vote in every topic put to vote by the organisation.

> I'd bet you that at least 15% of our membership has gone no-mail on
> the lists and is not even aware that they are still registered as
> members.

This oughta be easy enough to check, no? I'd prefer a real number to
speculation.

A

--
Andrew Sullivan | ajs(at)crankycanuck(dot)ca
In the future this spectacle of the middle classes shocking the avant-
garde will probably become the textbook definition of Postmodernism.
--Brad Holland

On Tue, Aug 07, 2007 at 11:28:34PM -0700, Josh Berkus wrote:
> Personally, given the tendency of some members of the SPI membership towards
> extreme uncompromising views and interminable argument, that the Concordet
> method of selecting for the "most acceptable" instead of the "most preferred"
> is a feature and not a bug.

I agree that the Condorcet method is ideal for single-winner elections,
for exactly this reason.

> It would also be trivial for any sufficiently motivated and organized
> project to get behind one candidate from their project to ensure them
> a place on the board.

That's only if no larger portion of the voting membership doesn't
coordinate against that project.

If the membership gets factionalized in a vote for any reason
whatsoever, the majority faction (which may not have >50 % support if
there are more than two factions) is currently able to dictate the
composition of the board, and can deny any other faction enfranchised
representation.

Note that the issue here is of principle. I don't expect this to be a
practical problem in the foreseeable future.

I'm still waiting for someone to point me toward the definition of SPI's
current multi-winner election method.

--
Antti-Juhani Kaijanaho, Jyväskylä
http://antti-juhani.kaijanaho.fi/newblog/
http://www.flickr.com/photos/antti-juhani/

On Wed, Aug 08, 2007 at 08:48:18AM +0100, MJ Ray wrote:
> Josh Berkus <josh(at)postgresql(dot)org> wrote:
> > [...] For example, changes to the bylaws
> > themselves require a majority of *all* members, not just voting ones. [...]
>
> Ouch. It really does say just "members" not "contributing members".
> That seems like a major bug to me, so I guess drastic action to fix it
> is justified.

It also says that only contributing members are eligible to vote. If we
read it literally, which we probably should where it's not
self-contradictory since it is legally binding, this would mean that we
need to have a number of contributing members at least equal to
two-thirds of the combined contributing and non-contributing membership
approve the changes in the bylaws. I have no idea if this is even
possible, since we may or may not have too many non-contributing members
in relation to contributing members for this to be a theoretical option
without a major drive to recruit new and active contributing members,
and the bylaws state that non-contributing membership does not expire.

Then again, if there is no feasible way to amend the bylaws reading
everything literally, I encourage the current board to seek the advice
of Greg (our lawyer) on how best to get out of this mess in a legal and
ethical way.

- Jimmy Kaplowitz
jimmy(at)spi-inc(dot)org

On Tue, Aug 07, 2007 at 11:36:42PM -0700, Josh Berkus wrote:
> > Speaking only on my own behalf, I don't think that voting in SPI should
> > be compulsory for contributing members, and my reading of the bylaw does
> > not require members to vote.
>
> Unfortunately, the quorum parts of our bylaws will paralyze the organization
> if we do not purge "inactive" members. For example, changes to the bylaws
> themselves require a majority of *all* members, not just voting ones. With
> current levels of participation that will be impossible to achieve.

I definitely support purging "inactive" members, by which I mean
downgrading them to non-contributing status (which is all we can do
under our current bylaws). However, this could make it much harder to
reach the amendment threshhold, as I explain in greater detail in my
reply to MJ Ray. I don't support requiring voting in SPI elections to
retain contributing status, because compulsory voting anywhere is
distasteful to me - it can cause people to cast uninformed votes, which
are usually worse than them not voting. I certainly don't think the
current rules we have include compulsory voting. Weeding out people who
don't vote and also don't respond to an inquiry email sent to the email
address in their SPI membership record, however, is reasonable, assuming
they can regain contributing membership by applying again in the normal
way subject to the normal criteria.

- Jimmy Kaplowitz
jimmy(at)spi-inc(dot)org

On Wed, Aug 08, 2007 at 09:50:01AM -0400, Andrew Sullivan wrote:
> Note that this isn't an argument against establishing criteria for
> participation, but it is an argument for excluding a requirement to
> vote in every topic put to vote by the organisation.
>

I don't think we've ever said that members will be required to vote :)

I think we'll be handling it something like:
http://lists.debian.org/debian-devel-announce/2007/07/msg00004.html
ie: voting will only be one way of 'showing activity'.

Additionally, this won't start until we've received legal advice from
Greg as to the status of quorum with regards to contrib/non-contrib
members.
--
Neil McGovern
Secretary, Software in the Public Interest, Inc.