It's patently ridiculous to ask people to register their lack of interest.

Well, obviously not to me, only, it's not "lack of interest". The people in my example are interested. They just don't have an opinion either way.

Am I harping on this one point too much? Maybe I'm liking the irony of the situation too much.

There are really only two choices, however, if we decide that we need a minimum level of voter turnout for a vote to be valid.

1. the level of voter turnout is counted from the number of people who voted yes or no
2. the level of voter turnout is counted from the number of people who voted, including abstentions/"informal" votes

What is it about counting the votes that gave you a particular perspective, Jen?

I've been saying that Australians are obliged by law to vote. And I did clarify that this does not oblige you to select a candidate.

I really wasn't having a go at you, billytea. I was thinking more of UnAustralians who were freaking out over the idea. Sorry if it seemed aimed at you.

Anathama - Quorum can be decided on a single ballot, or say two. Have a vote to decide whether it should be percentage or number (percentage not to exceed 50%+1). Optionally, if you vote for percentage, put in the percentage you favor. If you support a number, put in the number you favor (not less than 2).

OK so now we know percent or number, and we know the choice of percents if percent, and the choice of numbers if numbers.

So the second ballot lists all the possible numbers (if numbers are the choice, and I think they will be.) Or it lists all the suggested percents (if percents are the choice - though I think they won't be).

Regardless,they are put on a ballot in numerical order. Now here is the trick. People rank these choices.

Suppose the choices are

85,75,50,25,15,10,5,2

(That is this the complete list of numbers people submit for a quorum. OK, Say we rank on a scale of 1 to 8, 8 being best. (1 to 8 because there are 8 choices) OK, let us say you favor 25. (This is completely hypothetical. You will probably favor something different. And the choices will be different too.) So 25 would be your first choice (ranked 8). Maybe 15 would be your second choice (ranked 7). But this particular (hypotheical) you really hates the idea of anything less than 15, so 50, 75 and 85 are your next choices, and then 10, 5 and 2.

And others due the same type of ranking on completly different criteria. Remember, anyone can put a choice on - so we have any number anyone favors as a choice. Or we can do multiples of five between the imax and minimum suggest , plus the max and min if not a multiple of five, plus any other number anyone feels strongly about.

So in that case the choices would be

2,5,10,15,20,25,30,35,40,45,50,55,60,65,70,75,80,85

So now you rank on a scale of 1 to 18, 18 being best.

That being done, we count the ballots. There are various algorithms for counting ranked votes.

The easiest one - since we are doing manual counts is what is know as a Borda count. You simply add the ranks for each choice to get a numerical score, and the highest score is the winner. That is if 10 people rank 50 as their highest choice 18 and 10 rank it their lowest (1) then score for "50" as a choice 180 (18 time 10) + 10 (10 times 1) = 190.

It is also good in that it mesure intensity of support and tends to produce good compromise results. The disadvantage, In very rare cases in can produce "splits" the way simple plurality can - leading to the one with the most support losing. However in a case like this with a lot of options, it is less likely.

Another method - perhaps the best mathematically is the instant Cordecet round robin. It is a more complex method than that used by the Austrialians. (However a more complex counting option. The balloting is the same in all methods.)

You simply take the ranked options, and conduct a serious of simple simulated elections between them. That is you run "50" against "2" , "50" against "5" , "50" against "10" and so forth until "50" has been run in a paired election against every number but itself. Then you do the same with "45" (omitting a runoff against "50" because that is already done), and so forth. After the computer has run an imaginary election between every choice and every other choice, then it looks to see which choice, won the most of these hypothetical pairwise elections. And that is the winner.

Ok I know this sounds complicated. But the tough part is done by a computer, and is easy to program. The only thing you as a voter have to do is rank your choices.

I was going to wait until we decided whether to use a minyan based system. But since it looks likely, I thought I'd lay out some fair ways we can decide on a number without endless votes.

X-post with Rob, sort of. I like preferential ballots. I think some better ways of counting them have been developed since the Austrialian method was adapted. In a case like this board, it is not a major issue - the Australian method is certainly good enough. But since it has to be programmed anyway (counting preferential votes in a non-automated way is horribly tedious) we might at least consider other methods. Actually, I suspect only a small minority on this list are intereseted in technical details. If quorum is adapted, I suggest that those of us who are interested get together by email , forming an informal subcommitte.

I understand what you mean, Wolf. If 90% of the Buffistas don't care, then why shouldn't 12 people be allowed to institute massive changes? And if a majority of people agree with you, then 10 will end up the minimum number. Not trying to argue against your ten so much as I am trying to argue for taking our time with making that decision.

I think this paragraph will help me to clarify my position. If 90% of the Buffistas don't care, then 12 people should be able to institute minor changes, not massive changes. If it's a massive change, more than 90% of the Buffistas will definitely care. And I like the idea of having a voter range for quorum, though I haven't really thought through whether your way or Jon's way or a different way would be the best way to do this.

Edited to add, I think Gar's way may be the best way simply because he blinded me with mathematics.

I think this paragraph will help me to clarify my position. If 90% of the Buffistas don't care, then 12 people should be able to institute minor changes, not massive changes. If it's a massive change, more than 90% of the Buffistas will definitely care.

Wrdoy dwory word.

It's patently ridiculous to ask people to register their lack of interest.

No ... oh, John got there before me. If there's a minimum whatever in place, I'll vote abstain on measures I feel neither negative nor positive towards.

I think if 8 Buffistas want to have the site in Swahili, and 3 don't, then the board should be in Swahili. Obviously, 789 Buffistas don't care, so why not?

Really, I just came up with my ballot suggestion off the cuff. I merely wanted to make sure people were thinking about taking our time and offerring as many choices as possible for determining this minimum number that is being bandied about. And it seems people, or the few people who have posted this afternoon, do see that.

Cindy, none of the proposals is suggesting 51% of all registered users. I think most of us are thinking along the line of 51% of the active voting members (using this current vote-in-progress to determine how many voting members there are). So however many vote on this measure that's going on right now (150, 250, whatever), that's the number of voting members.

And I hate to say it, but no matter whether we use a numbre or a percentage, we will have to probably reevaluate that figure every year in order to be fair. The number of users here probably goes up and down regularly.

Time consuming and tricky stuff this democracy. That's why I still favor the blah-blah-Buffista method you have been using since the time immemorial.

Dammit, I thought I was avoiding mathematics. t grin

Robs proposal and both my proposals are the same from the voter standpoint. You rank candidates.

The Borda method is simple enough that it could be implemented in a spreadsheet, without programming. It has advantages and disadvantages in terms of fairness over other methods. It tends to strongly approximate consensus to result in a compromise choice. But if there are two extremely popular choices and a bunch of unpopular ones, the popular choices can split the vote and result in choice being picked less popular than either.

The Condercet method is by some arguments the fairest. But it is the most complex to implement - it must be programmed on a computer. There is too much drudgework for somene to simply receive e-mail and count it.

The Austrialian method is in between. It is extremely tedious to do manually, but it is possible. (I would say that in fairness to whoever has to count the ballots, we should automated the counting even if the Austrialian method is chosen.) Unlike Borda count, it (for practical purposes) eliminates any chance of the least popular option being chosen due to vote splitting. But unlike the Condercet method, it can still allow for a type of vote splitting on second choices. That is while the actual last choice is unlikely to be picked, it can result in a similar type of voting splitting between second and third choices. To take an American example - if we had the Austrialian system in place we would have ended with Gore not Bush as president - provide Nader voter made Gore their second choice. But if 2/3rds of the Nader votes had made McRenolds their second choice, and Gore their third, and one third had made Bush their second choice, then Bush would have still gotten in under the Australian system.

Or to put it another way, under the Austrialian system, you can vote for what you really want as your first choice. But you had better still think strategically, about what can win, rather than just what you want, when you make your second choice.

I think what Katie was saying is that the proposal we are voting for now only allows for simple majority votes. So if we then use the Borda method, we aren't doing what we said we would.

OK - both Austrialian (usually called Instant Runoff in the U.S.) balloting, and Cordecet are simple majority. They just do multiple counts to get a simple majority. And I think Borda is included in what is mean't by simple majority. Because I mean if 51 or more vote for something under Borda (or any of the systems) as their first choice that wins. All of these are counting methods to resolve choices between multiple options, which *simple* as opposed to complex majority requires. None of them require super-majorities - that is more than 50%+1 to pass somethings. Note that even the simplist system, plurarlity where the one with most votes wins is not a *simple* majority since if there are more than two options stuff can be passed under plurality by less than a 50% vote.

So that would mean that we were rushed into voting without thinking through what we are voting on. But I don't think that is th e case. I don't think anyone meant it to be interperted that narrowly and legalistically. Stuff like whether to use choice voting for mulitiple mutually exclusive choices is not on the ballot - and thus is not resolved by the question of whether the decision is simple majority or not. If we pick simple majority, that ruiles out super-majority and consenus, It does not rule out other counting methods, provided of course a simple majority support those counting methods.