W3C home > Mailing lists > Public > public-w3process@w3.org > August 2014

Re: voting simple illustration

From: David Singer <singer@apple.com>
Date: Mon, 25 Aug 2014 15:44:26 -0700
Cc: timeless@gmail.com, public-w3process@w3.org
Message-id: <2621D6AC-791B-4EEB-BC6D-AA7E23926158@apple.com>
To: chaals@yandex-team.ru
OK, we are beginning to converge. I think.

On Aug 25, 2014, at 14:56 , chaals@yandex-team.ru wrote:

> Note that this situation is like the one where the French managed to choose between Jacques Chirac (who was about evenly preferred and loathed) and Jean-Marie LePen (who was preferred by a tiny minority and loathed by an outright majority, including but by no means limited to those who also loathed Chirac). Chirac won by a margin that usually only occurs in places where elections are run along the lines described in Asterix in Corsica (TL;DR: By cheating). Under our rules it would be a close run thing, and there wouldn't be the run-off where you could select the person you hated less.
> 
> Note also that the proposal is for an experiment, the results of which would be available to the W3C Team for analysis, with an indication that they would release data they felt was both interesting and not prejudicial to the perception of legitimacy of elected candidates - elections will actually continue to be run under the system we have for the foreseeable future.
> 
> And to the points raised…
> 
>> ‎> In no elections we have do we give people anti-votes (votes against), or blocking votes (“I get my way or the whole election gets blown up”).
> 
> (Which "we" do you mean? In W3C the only votes with a mandated process have rules that have landed you in the pickle you're now in).

I mean W3C elections.  We don’t have either concept at the moment, is what I am saying.

> 
>> * Motion of no confidence
>> * AGM (annual general meeting) votes where a certain portion of the vote must be in favor
> 
> Monetary supply (i.e. refusing to allow money for a government to function) is a similar thing in practice. You live in the US, so you're probably more familiar with votes where it is seriously threatened. The last serious one in Australia was 1975 (and yes, the government was kicked out as a result).
> 
> Votes with a quorum requirement are from time to time affected by walkouts, which is a "yes/no" version of "none of the rest" - more usually in smaller societies than the State of California, or even the Regional government of León.
> 
> Presidential college votes in states where the candidate who gets 51% of the state gets all the votes for that state are not the same, but demonstrate that there are all sorts of strange voting systems possible, and that some seem a lot fairer than others. (See also "Gerrymander”).

Sure, I have been involved in elections with quorum requirements.  Do we (w3c) have any now?  Even if we did, then some of the cases are covered by a simple failure to list a preference for some candidates (r.g. the quorum requirement that, to be elected, one has to receive yes votes from >Y% of the voting pool).

> 
>>> ‎ Do we have any elections where I can state on my ballot “or I want the election to be invalid”?
>> 
>> I recently almost lived through one under Roberts Rules of Order [example]
>> 
>> You could claim that it wasn't really a case of declaring for election invalid, but it was close.
>> 
>> I also have been called to vote for a mutual fund because their (AGM) quorum requirement failed for their slate vote and so they were having a second vote and hoping to reach quorum.
>> 
>>> I don’t think so; the closest I know in any election is a quorum requirement;
>>> “for the election to be valid, X% of the voting pool must file a vote” or “for an election to be valid, the winning candidates must receive support from at least Y% of the [votes cast | voting pool]”.
> 
> Actually when the Soviet Union started taking elections more seriously under Gorbachev, a number of candidates were undone by the suggestion you make, that if they didn't get 50% of the votes cast, they were out. In several cases there was only one candidate, who failing to get the 50% was not declared elected.
> 
>> There's also the recall election, which is becoming more common in the states. It's done roughly by initiative -- you're in California, so you're probably quite familiar with it ;-)
>> 
>>> The first is clearly irrelevant to a discussion of the roll-over method, and the second is covered by a simple “don’t rank those candidates”
>>> If I fail to vote for someone, then my vote does not roll to them; quorum requirements may now fail. I still don’t see a benefit to “and no other candidate”.
>> 
>> I'm pretty sure that it didn't occur to me that quorum requirements could fail this way -- perhaps it's worth outlining and explaining in the ballot instructions.
> 
> This is a technical detail about the counting methods used. The default (rightly IMHO) in most STV systems I know of in politics is to let the votes get exhausted, and count preferences among the remaining candidates, based on the assumption that "I don't mind - I prefer my candidate but these are better than endless elections" is a more common default position than "I'd rather keep trying to force elections than have these candidates win".
> 
>> For the AGM I described initially, our Constitutional scholar explained which votes had which values for numerator / denominators (it varied by election class).
> 

Right, I thought we were conducting a voting experiment;  I didn’t conceptually include adding quorum requirements of either kind into that. Even with them, I think most cases are covered.

As I see it, the one case that isn’t is where people want to express “don’t elect anyone else”, which is kinda like an anti-vote for all other candidates;  I think you’re saying that if a candidate gets more than a certain threshold of such anti-votes, they are not elected.  That threshold might be
a) > the number of positive votes cast for them
b) > X% of the number of people voting
c) > Y% of the entire electorate size

and so on

I guess I still doubt we need that rule.  All rules that say “but so and so is not elected unless this other condition is also met” suffer from the problem that one might not fill all available seats, whereas in today’s W3C elections, if there are enough candidates, we will fill the seats.

Even if we were to go there, I would suggest quorum rules before anti-votes (for the election to be valid >X% of the electorate must vote, or the candidate must receive >Y% affirmative botes — possibly transferred — of the ballots cast).

The “and no other candidate” adds complexity and I want to be sure we both need it and can explain it.


David Singer
Manager, Software Standards, Apple Inc.
Received on Monday, 25 August 2014 22:44:56 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:35:11 UTC