Re: What does FD Mean

[Russ Allbery]

"Barak A. Pearlmutter"  writes:

> Let's say a cohort of voters prefers option APRICOT to option BANANA,
> but would like neither (FD) even better. However they are well aware
> that there's no way FD will win.

> It is possible that if they vote their true preference,

> FD > APRICOT > BANANA

> then BANANA will win, while if they vote

> APRICOT > FD > BANANA

> then APRICOT will win, due to majority/quorum issues. In other words,
> they are penalized for voting honestly.

Isn't this true regardless of majority/quorum issues?  That looks like an
example of compromising, and the Wikipedia page on Condorcet says that it
(without our majority addition) is susceptible to compromising if there is
a majority rule cycle.

It's not clear to me that the majority requirement makes it substantially
worse.  I believe the majority requirement can only affect the outcome in
the case of majority rule cycles, and (please correct me if I'm wrong)
those seem to be rare in our votes.

(By definition an option without a majority was voted below FD by a
majority of voters, and therefore I believe it's impossible in Condorcet
for that option to defeat FD if there are no cycles, and thus it would be
impossible for it to win if there are no cycles since FD would always win
instead.)

I'm not as sure about the quorum rule, but in project-wide (as opposed to
technical committee) votes, I don't remember an option failing quorum that
would have had any realistic chance of winning.  Generally that means the
option is losing by a huge margin.

Obviously that could be different in a very low turnout vote.

-- 
Russ Allbery (r...@debian.org)  

Re: opinion on Choice 1

[Salvo Tomaselli]

People are free to express their opinion.  That does not mean the Debian
Project is obligated to provide a platform for those opinions on the
debian-vote mailing list, which exists to facilitate discussions among
voting members of the Debian Project regarding matters that will be voted
on.

Non-voting posters to debian-vote are almost exclusively outside agitators
and there's no reason subscribing to debian-vote should mean receiving their
bullshit in our mailboxes.

I only posted because a debian mailing list was used to reach me about 
the issue and it contained incorrect statements that I corrected.


If no GR had been made, I wouldn't have received anything via debian 
mailing lists and I would have not replied via debian mailing lists.


If you don't want people to reply, just don't email them.


Best

Re: opinion on Choice 1

[Steve Langasek]

On Mon, Apr 05, 2021 at 03:03:48PM +0200, Pierre-Elliott Bécue wrote:
> Le dimanche 04 avril 2021 à 16:37:15-0700, Steve Langasek a écrit :
> > On Sun, Apr 04, 2021 at 03:09:10PM +0200, Bernd Zeimetz wrote:
> > > On Tue, 2021-03-30 at 12:18 +0200, Ulrike Uhlig wrote:

> > > > People without voting rights repeatedly tried to lobby or push for a 
> > > > certain agenda on this list. 

> > > Welcome to Debian.
> > > People are free to express their opinion, even if they are not owning
> > > an @debian.org email address. And that is actually a very good thing.
> > > The interested reader is able to filter messages and maybe maintain a
> > > list of people to ignore if needed. It might be annoying for you, but
> > > free speech is not always fun.

> > People are free to express their opinion.  That does not mean the Debian
> > Project is obligated to provide a platform for those opinions on the
> > debian-vote mailing list, which exists to facilitate discussions among
> > voting members of the Debian Project regarding matters that will be voted
> > on.

> > Non-voting posters to debian-vote are almost exclusively outside agitators
> > and there's no reason subscribing to debian-vote should mean receiving their
> > bullshit in our mailboxes.

> Even though it's hard and can be tiresome to many of us (and maybe
> drives some away), as long as possible, I'd like the majority of our
> lists to stay open to all people willing to express something.

> Blocking potentially relevant comments from non contributors because
> some trolls are trying to wreck havoc is giving them too much importance
> and therefore giving them an easy victory.

Can you point to an example of a post you consider actually (not "possibly")
relevant from a non Debian voter to debian-vote in the past 2 years?

Why should we allow third parties to lobby Debian electors using our mailing
list infrastructure?

> And, despite what I personally think, a non-contributor calling the RMS
> vote a "witch hunt" is not necessarily a troll.

I never used the word "troll", which for me has a very specific meaning
grounded in its historical usage in online communities.  I referred to them
as "outside agitators", which I believe they are - whether or not a
particular individual's intention is to derail the discussion, it is
certainly their intention to influence the outcome of Debian's decision
process according to their own interests, whether or not those align with
the interests of the Debian voters as a democratic body.

-- 
Steve Langasek   Give me a lever long enough and a Free OS
Debian Developer   to set it on, and I can move the world.
Ubuntu Developer   https://www.debian.org/
slanga...@ubuntu.com vor...@debian.org


signature.asc
Description: PGP signature

Re: re. RMS

[Thomas Goirand]

Someone privately made me realize that what I wrote is probably not nice
to read for autistic people. It was not my intention to hurt anyone, and
I have nothing against autistic people, who in some cases, are capable
of being leaders.

On 4/5/21 4:12 PM, Pierre-Elliott Bécue wrote:
> Whether RMS has any neuroatypical condition and whether or not it should
> give him lenience and/or preventing him from leading the FSF is paths I
> won't go along, as I am no doctor and have no expertise on these
> matters.

Also, I agree with what PEB wrote above.

Though I wasn't making the case that RMS has any "neuroatypical
condition" (Miles did). I'm saying that Miles's explanation/excuse is
probably counter productive, and is instead a case *against* RMS being
in charge. If it was the case, I would not agree with Miles that says we
should "give him leeway for that".

I hope it's more clear like this, and it doesn't hurt anyone this time.

Cheers,

Thomas Goirand (zigo)

Re: Re: What does FD Mean

[Adrian Bunk]

On Mon, Apr 05, 2021 at 06:35:52AM -0700, Felix Lechner wrote:
>...
> would it be better for a voting system to
> quadruple-count, or otherwise strengthen, options voters rank in the
> middle—thereby recognizing that a compromise between two or more sides
> is always a prerequisite for peace?

This is the first time Debian holds a GR for a position statement
about issues of the day outside of Debian.

There is a valid question whether or not a 2:1 or 3:1 majority 
requirement would be more appropriate for such GRs.

> Kind regards
> Felix Lechner
>...

cu
Adrian

Re: re. RMS

[Pierre-Elliott Bécue]

Le lundi 05 avril 2021 à 16:04:58+0200, Thomas Goirand a écrit :
> On 4/5/21 3:29 PM, Miles Fidelman wrote:
> > Given that Friday was Autism Awareness Day, it might be worth noting
> > that RMS is clearly "on the spectrum" - and well known since the days he
> > slept in his office at MIT (my student days).
> > 
> > Why is it that nobody ever gives him any leeway for that?
> 
> Without expressing myself about one side or another, or even not
> expressing myself about RMS case (ie: if he's an autistic person or not)
> autism is a mental disease. Putting someone with a mental disease in
> charge is not exactly a good idea.

I am not sure to agree with your last sentence. Of course, some
pathologies may prevent someone to be able to lead a group/activities,
but it's not something I'd state as generally as you tend to do.

Whether RMS has any neuroatypical condition and whether or not it should
give him lenience and/or preventing him from leading the FSF is paths I
won't go along, as I am no doctor and have no expertise on these
matters.

Regards,

-- 
Pierre-Elliott Bécue
GPG: 9AE0 4D98 6400 E3B6 7528  F493 0D44 2664 1949 74E2
It's far easier to fight for one's principles than to live up to them.


signature.asc
Description: PGP signature

Re: What does FD Mean

[Marc Haber]

On Mon, Apr 05, 2021 at 04:26:25PM +0300, Adrian Bunk wrote:
> For people living in a country like Germany where the shares of 
> representation in parliament are based on the nationwide vote,

The Bundestagswahl is a weird combination of direct vote and
proportional vote with a minority blocking clause, independently
conducted for each of the 16 states and then combined. You need to be a
mathematician AND an expert in public law to understand.

> Debian is usually the first and only contact with anything
> like Condorcet.

Confirmed.

Greetings
Marc

-- 
-
Marc Haber | "I don't trust Computers. They | Mailadresse im Header
Leimen, Germany|  lose things."Winona Ryder | Fon: *49 6224 1600402
Nordisch by Nature |  How to make an American Quilt | Fax: *49 6224 1600421

Re: re. RMS

[Thomas Goirand]

On 4/5/21 3:29 PM, Miles Fidelman wrote:
> Given that Friday was Autism Awareness Day, it might be worth noting
> that RMS is clearly "on the spectrum" - and well known since the days he
> slept in his office at MIT (my student days).
> 
> Why is it that nobody ever gives him any leeway for that?

Without expressing myself about one side or another, or even not
expressing myself about RMS case (ie: if he's an autistic person or not)
autism is a mental disease. Putting someone with a mental disease in
charge is not exactly a good idea.

Are you trying to make a case that we should be indulgent with RMS
because he's autistic? If that is your intention, you're achieving the
exact opposite thing with your message...

Cheers,

Thomas Goirand (zigo)

Re: What does FD Mean

[Adrian Bunk]

On Mon, Apr 05, 2021 at 02:34:28PM +0200, Pierre-Elliott Bécue wrote:
> Le lundi 05 avril 2021 à 14:07:13+0200, Marc Haber a écrit :
> > On Mon, Apr 05, 2021 at 12:15:25PM +0100, Barak A. Pearlmutter wrote:
> > > Making a system more complicated to try and address a specific
> > > deficiency rarely reduces its attack surface. In this case, our voting
> > > system involves multiple levels (quorum, majority, ranking resolution)
> > > each with its own criteria and threshold and (due to Arrow's Theorem)
> > > unavoidable flaws, and every feature of this sort increases the
> > > system's attack surface to both strategic voting and to just plain
> > > doing the wrong thing given honest votes. Moving FD around in the
> > > ordering is an example of this, as is a quorum boycott.
> > 
> > I have been a DD for nearly 20 years and I have not yet understood how
> > we vote. Before I joined Debian, I thought that the way Germany votes
> > for the Bundestag is a complex method.
> > 
> > Greetings
> 
> It's probably because I'm a mathematician, but I really enjoy our voting
> system, despite it also having flaws.

For me it is also mostly mathematical curiosity, and there is no 
situation in real-life elections where it would be relevant.

Voting methods like Condorcet try to solve problems in single-round 
first-past-the-post systems with more than 2 candidates that are common 
in the UK and some former British colonies.

For people living in a country like Germany where the shares of 
representation in parliament are based on the nationwide vote,
Debian is usually the first and only contact with anything
like Condorcet.

cu
Adrian

Re: Re: What does FD Mean

[Felix Lechner]

Hi,

On Mon, Apr 5, 2021 at 4:33 AM Barak A. Pearlmutter
 wrote:
>
> Moving FD around in the
> ordering is an example of this, as is a quorum boycott.

When a center option is likely to fail our majority requirement [1]
should I rank preferable extreme choices above FD even if I am
strictly moderately inclined?

Along the same lines, would it be better for a voting system to
quadruple-count, or otherwise strengthen, options voters rank in the
middle—thereby recognizing that a compromise between two or more sides
is always a prerequisite for peace?

Kind regards
Felix Lechner

[1] Section A.6 (3) of the constitution:
https://www.debian.org/devel/constitution

re. RMS

[Miles Fidelman]

Given that Friday was Autism Awareness Day, it might be worth noting 
that RMS is clearly "on the spectrum" - and well known since the days he 
slept in his office at MIT (my student days).


Why is it that nobody ever gives him any leeway for that?

Miles Fidelman

--
In theory, there is no difference between theory and practice.
In practice, there is.   Yogi Berra

Theory is when you know everything but nothing works.
Practice is when everything works but no one knows why.
In our lab, theory and practice are combined:
nothing works and no one knows why.  ... unknown

Re: opinion on Choice 1

[Pierre-Elliott Bécue]

Le dimanche 04 avril 2021 à 16:37:15-0700, Steve Langasek a écrit :
> On Sun, Apr 04, 2021 at 03:09:10PM +0200, Bernd Zeimetz wrote:
> > On Tue, 2021-03-30 at 12:18 +0200, Ulrike Uhlig wrote:
> 
> > > People without voting rights repeatedly tried to lobby or push for a 
> > > certain agenda on this list. 
> 
> > Welcome to Debian.
> > People are free to express their opinion, even if they are not owning
> > an @debian.org email address. And that is actually a very good thing.
> > The interested reader is able to filter messages and maybe maintain a
> > list of people to ignore if needed. It might be annoying for you, but
> > free speech is not always fun.
> 
> People are free to express their opinion.  That does not mean the Debian
> Project is obligated to provide a platform for those opinions on the
> debian-vote mailing list, which exists to facilitate discussions among
> voting members of the Debian Project regarding matters that will be voted
> on.
> 
> Non-voting posters to debian-vote are almost exclusively outside agitators
> and there's no reason subscribing to debian-vote should mean receiving their
> bullshit in our mailboxes.

Even though it's hard and can be tiresome to many of us (and maybe
drives some away), as long as possible, I'd like the majority of our
lists to stay open to all people willing to express something.

Blocking potentially relevant comments from non contributors because
some trolls are trying to wreck havoc is giving them too much importance
and therefore giving them an easy victory.

And, despite what I personally think, a non-contributor calling the RMS
vote a "witch hunt" is not necessarily a troll.

An angry person, surely. But being angry and being a troll are two
orthogonal things.

That's my two cents, and I'm no one to decide, of course.

-- 
Pierre-Elliott Bécue
GPG: 9AE0 4D98 6400 E3B6 7528  F493 0D44 2664 1949 74E2
It's far easier to fight for one's principles than to live up to them.


signature.asc
Description: PGP signature

Re: opinion on Choice 1

[charlie derr]

On 4/4/21 7:37 PM, Steve Langasek wrote:
> Non-voting posters to debian-vote are almost exclusively outside agitators
> and there's no reason subscribing to debian-vote should mean receiving their
> bullshit in our mailboxes.


i don't disagree with any of the above, but in order to try to lend a
bit of balance (because i do think it's a shame), i just wanted to offer
my own point of view, as an avid user of debian/GNU linux for something
like 2 1/2 decades who is *not* a debian developer.

Obviously there's a breadth of opinion among debian developers on this
particular issue, and as often seems to be the case *whenever* GRs get
proposed and discussed, tempers and emotions can run high.

But i'm quite impressed with the overall tenor of the conversation which
took place in this discussion period. Especially the way that (contrary
to what i've witnessed in so many (almost all that i've ever paid
attention to?) other internet forums) many sub-threads which became
quite inflammatory (with respect to the rhetoric used) at certain
points, so often were brought back to salient issues and *away* from
outright hostility by other participants working hard to redirect the
conversations back "on track". Not always (as i'm sure most who've been
following the list postings well know), but often enough for me to take
heart in both the structure (debian constitution) and the character of
almost every single developer i've witnessed engage in this incredibly
divisive debate/dialogue.

It seems a pretty great example of tolerating dissenting opinions (that
i strive to emulate), which is doubly interesting as in this case, the
specific topic is also about the same thing (tolerating dissenting
opinions (or not) and to what degree, it is appropriate to do so).

   thanks,
   ~c




-- 
Charlie Derr   Director, Instructional Technology 413-528-7344
https://www.simons-rock.edu Bard College at Simon's Rock
Encryption key: http://hope.simons-rock.edu/~cderr/
Personal writing: https://medium.com/@cderr   Pronouns: he or they

Re: What does FD Mean

[Pierre-Elliott Bécue]

Le lundi 05 avril 2021 à 14:07:13+0200, Marc Haber a écrit :
> On Mon, Apr 05, 2021 at 12:15:25PM +0100, Barak A. Pearlmutter wrote:
> > Making a system more complicated to try and address a specific
> > deficiency rarely reduces its attack surface. In this case, our voting
> > system involves multiple levels (quorum, majority, ranking resolution)
> > each with its own criteria and threshold and (due to Arrow's Theorem)
> > unavoidable flaws, and every feature of this sort increases the
> > system's attack surface to both strategic voting and to just plain
> > doing the wrong thing given honest votes. Moving FD around in the
> > ordering is an example of this, as is a quorum boycott.
> 
> I have been a DD for nearly 20 years and I have not yet understood how
> we vote. Before I joined Debian, I thought that the way Germany votes
> for the Bundestag is a complex method.
> 
> Greetings

It's probably because I'm a mathematician, but I really enjoy our voting
system, despite it also having flaws.

-- 
Pierre-Elliott Bécue
GPG: 9AE0 4D98 6400 E3B6 7528  F493 0D44 2664 1949 74E2
It's far easier to fight for one's principles than to live up to them.


signature.asc
Description: PGP signature

Re: What does FD Mean

[Marc Haber]

On Mon, Apr 05, 2021 at 12:15:25PM +0100, Barak A. Pearlmutter wrote:
> Making a system more complicated to try and address a specific
> deficiency rarely reduces its attack surface. In this case, our voting
> system involves multiple levels (quorum, majority, ranking resolution)
> each with its own criteria and threshold and (due to Arrow's Theorem)
> unavoidable flaws, and every feature of this sort increases the
> system's attack surface to both strategic voting and to just plain
> doing the wrong thing given honest votes. Moving FD around in the
> ordering is an example of this, as is a quorum boycott.

I have been a DD for nearly 20 years and I have not yet understood how
we vote. Before I joined Debian, I thought that the way Germany votes
for the Bundestag is a complex method.

Greetings
Marc

-- 
-
Marc Haber | "I don't trust Computers. They | Mailadresse im Header
Leimen, Germany|  lose things."Winona Ryder | Fon: *49 6224 1600402
Nordisch by Nature |  How to make an American Quilt | Fax: *49 6224 1600421

Re: Re: What does FD Mean

[Barak A. Pearlmutter]

On Mon, 5 Apr 2021 at 11:57, Adrian Bunk  wrote:
> On Mon, Apr 05, 2021 at 11:46:23AM +0200, Kurt Roeckx wrote:
> > A possible solution is to drop the majority requirement
> > and have a quorum on the number of people that vote ...
>
> A quorum on the number of people who vote means that a vote against the
> proposal counts for the quorum.
>
> Assuming a quorum high enough, this gives a coordinated boycott of the
> vote a higher chance of defeating a proposal than voting against it.

Making a system more complicated to try and address a specific
deficiency rarely reduces its attack surface. In this case, our voting
system involves multiple levels (quorum, majority, ranking resolution)
each with its own criteria and threshold and (due to Arrow's Theorem)
unavoidable flaws, and every feature of this sort increases the
system's attack surface to both strategic voting and to just plain
doing the wrong thing given honest votes. Moving FD around in the
ordering is an example of this, as is a quorum boycott.

Since voting systems are necessarily vulnerable (Arrow's Theorem!) our
objective cannot be perfection, but rather good performance under
realistic conditions.

But I'm not sure this is the right place to discuss these issues.
Maybe debian-meta-v...@lists.debian.org?

Cheers,

--Barak.

Re: Re: What does FD Mean

[Barak A. Pearlmutter]

What you say is all correct, although I suppose people might be able
to get at least a rough poll of voter preferences if they actually
care. Assuming people don't know enough details about others'
preferences to vote strategically is basically security by obscurity
so I wouldn't want to rely upon it.

As a matter of methodology, I think we should separate the process of
identifying problems in our voting system and gauging their severity
from the process of looking for ways to improve it. This makes it much
easier to stay objective.

Warren D. Smith wrote some code to measure the performance of various
voting systems (including various Condorcet systems) under various
simulated conditions.

Cheers,

--Barak.

Re: Re: What does FD Mean

[Adrian Bunk]

On Mon, Apr 05, 2021 at 11:46:23AM +0200, Kurt Roeckx wrote:
>...
> A possible solution is to drop the majority requirement
> and have a quorum on the number of people that vote
>...

A quorum on the number of people who vote means that a vote against the 
proposal counts for the quorum.

Assuming a quorum high enough, this gives a coordinated boycott of the 
vote a higher chance of defeating a proposal than voting against it.

> Kurt

cu
Adrian

Re: What does FD Mean

[Adrian Bunk]

On Mon, Apr 05, 2021 at 12:47:58AM +0200, Kurt Roeckx wrote:
>...
> But the reason for yes/no is the majority requirement. In this GR
> all options have a majority ratio of 1. This means more people
> need to put the option above of FD than people who put the option
> below FD, or the option gets dropped.
>...

On a side note, there is a bug in the vote counting software or its 
configuration where it would generate a nonsensical result that also 
violates explicit wording of the constitution in a corner case.

The constitution says:

  An option A defeats the default option D by a majority ratio N, if 
  V(A,D) is greater or equal to N * V(D,A) and V(A,D) is strictly great

>From the latest systemd GR[1]:

  * Option5 passes Majority. 2.185 (271/124) >= 1
  * Dropping Option6 because of Majority. 0.808 (173/214) <= 1

>= 1 is not strictly great, and an option both passing the majority 
requirement and being dropped for failing the majority at the same
time in the = 1 case would not make sense.

> Kurt

cu
Adrian

[1] https://www.debian.org/vote/2019/vote_002

Re: Re: What does FD Mean

[Kurt Roeckx]

On Mon, Apr 05, 2021 at 09:45:15AM +0100, Barak A. Pearlmutter wrote:
> 
> Let's say a cohort of voters prefers option APRICOT to option BANANA,
> but would like neither (FD) even better. However they are well aware
> that there's no way FD will win.
> 
> It is possible that if they vote their true preference,
> 
> FD > APRICOT > BANANA
> 
> then BANANA will win, while if they vote
> 
> APRICOT > FD > BANANA
> 
> then APRICOT will win, due to majority/quorum issues. In other words,
> they are penalized for voting honestly.

It's true that our system isn't perfect, and that if you know how
others will vote you can manipulate the outcome. You can try to
influence which options get dropped or not due to quorum or more
likely majority requirements.

I think in your scenario, I assume APRICOT was dropped due to the
quorum or majority requirements and changing their vote to mark
APRICOT as acceptable means it was considered and can now have
more people that prefer it over BANANA.

It can also be abused the other way around. People who
prefer:
BANANA > APRICOT > FD
might vote:
BANANA > FD > APRICOT
in an attempt to drop APRICOT as an option, for instance
because they know there are a lot of people who're going
to vote FD > APRICOT > BANANA.

I don't believe this is actually a problem in Debian for
general resolutions and elections. I don't think we know
enough details about what others will vote or have group
coordination to abuse it.

A possible solution is to drop the majority requirement
and have a quorum on the number of people that vote, and this
would work for most votes, except those where we now have
a supermajority requirement.


Kurt

Re: Re: What does FD Mean

[Barak A. Pearlmutter]

Kurt Roeckx  writes:

> There are 2 ways the FD option has an effect on the result.

> The first option is the quorum requirement. For a GR the quorum is
> 3*Q, which is around 47 for this vote. 3*Q people need to put the
> option above FD to meet the quorum, or the option is dropped.

> But the reason for yes/no is the majority requirement. In this GR
> all options have a majority ratio of 1. This means more people
> need to put the option above of FD than people who put the option
> below FD, or the option gets dropped.

> Note that you can rank the option the same as FD, which is neither
> yes nor no. So it's more than 50% of those votes that voted yes or
> no for that option that need to vote yes for the option to be
> considered.

> There are also 2:1 and 3:1 majority requirements, which you could
> translate as 66.6% and 75% need to say yes.

That is all true: we have a yes/no majority/quorum mechanisms combined
with a Condorcet system.
This Frankenscheme actually creates some interesting extra pathologies
for the Debian voting system, beyond those you get from plain
Condorcet.

Let's say a cohort of voters prefers option APRICOT to option BANANA,
but would like neither (FD) even better. However they are well aware
that there's no way FD will win.

It is possible that if they vote their true preference,

FD > APRICOT > BANANA

then BANANA will win, while if they vote

APRICOT > FD > BANANA

then APRICOT will win, due to majority/quorum issues. In other words,
they are penalized for voting honestly.

This is unrelated to the current vote (maybe?) but if a DPL candidate
would put "get a committee together to investigate whether the Debian
voting system and related processes could be significantly improved"
I'd consider that a big plus.

Cheers,

--Barak.