Talk:Basic principles

From Metagovernment - Government of, by, and for all the people

Rating based weighting

As far as I am aware, there is no consensus about weighing user scores.

--AurSaraf 00:42, 14 May 2008 (UTC)

Weighting contributions and comments based on the user's rating is not the same thing as weighting the user's 'vote' (yes, i know this is a contentious term) on a particular item of proposed legislation. I don't think there is anything currently about voting or scoring in the Basic Principles ... and, I'm not sure that the voting/rating/etc mechanism by which a proposal becomes legislation should be a Basic Principle .. it seems almost more like a procedural detail. thoughts?

--Jacki Buros 05:10, 16 May 2008 (UTC)

I am still vigorously in favor of weighted scoring. However, if I cannot convince the rest of the group soon, then I agree that this at least should be pulled from Basic Principles and put as a suggestion for how to run Metascore. Note that Basic principles#Geographic distinctness within a global community (currently the last principle) would also have to change.

The reason I think weighting is important is because it enables communities to protect themselves from being overrun by newcomers who do not have a vested interest in the community.

What is anyone's interest in trying to form a small community of people with a related interest, if at any time a group of vandals can join in and degrade their community with irrelevancies?

— Ed Pastore 21:58, 16 May 2008 (UTC)

Below is a rough idea responding to your point. I haven't audited it, so there's probably reasoning errors. But I'd like to present it, anyway.

My intuition for the right way to solve this is that a person gets a vote in a particular bill if and only if the bill affects that person; it's the community created by the bill, not the community created by the forum of discussion, that matters. If I have a rule about a lake that's private or joint property, which really has no effect on the outside world, the owners of the lake are the ones who would vote on the bill. If that same bill was about a lake that is public property, we need to define which shard of the "public" in question will be affected, and put them as the stakeholders. Roving trolls can't sabotage discussion unless they have a legitimate stake in its outcome - in which case a single person's veto ought to have all the effects that any other person's veto would have.

Humphrey 22:30, 20 June 2008 (UTC)

But how do we define who is affected by a bill? What software mechanism can we use to determine someone's degree of impact from a bill? Especially, when it varies... a person who lives on a river is highly influenced by what happens right there on that river. Someone who lives a thousand miles downstream is somewhat affected by those upstream decisions... but not as much. The idea of user-weighting solves this. If a user has a score within each community which is defined by their participation in that community... then we have an organic, working solution to deciding how much say someone should have within any one community. — Ed Pastore 19:07, 22 June 2008 (UTC)

Hrm. I have some ideas as to that, but they would add a lot of complexity to what is yet a very conceptually simple system.

Lemme get back to you on it. - Humphrey 01:01, 23 June 2008 (UTC)

Something else I'd like to say on this matter:

Currently, every ranking system proposal we have includes a "veto vote" which requires very large opposition to remove (or in some cases, offers no method of removal at all if the vetoing party stays involved in the discussion). Would the *10 multiplier in your proposal simply be applied atop the community score multiplier? I'd also like to brainstorm how community scores might affect vetos in Aur's proposal - while the threshold of +s necessary to flag "consensus" could certainly be weighted by score, vetos are an all-or-nothing effect in that system.

Aur, any thoughts on this? - Humphrey 01:01, 23 June 2008 (UTC)

I personally like and agree with the idea that votes should be wighted according to various elements. I think it is open to speculation what those elements are. But generally we might make a general system where there are various parameters. And then each community decides how open or close they are by setting the parameters to different values. WHat cannot be changed that easyly is what gets multiplied and what summed. In short the shape of the formula. ANd I think we need to work on that first. But we could start bymaking a list of all the elements that might affect the weight. And if we disagree that something shoud or shouldn't be there, we just add it, with a parameter before, like:

(k1*w1+k2*w2+...)*V

where the ki are the parameters. And then if some community prefer not to use them, they can just set them to 0. I think the priority should be to make the software general, not to fit only one of us, when we disagree.--Pietro 14:33, 24 June 2008 (UTC)

It seems no matter how hard we try we can't increase our ambition above the notion that we will be unable to achieve sythesis and consensus and therefore have to resort to imposition of disagreeable solutions achieved by majority vote. We haven't yet got a model or process to achieve sythesis and consensus on important and emotive issues and it is precisely this that we should focus our efforts on developing. This is not a software problem, but software may make it a whole lot easier to implement on a mass scale. So let's use a rating system where the contributions of people with more connection, history or reputation in a particular area are awarded greater visibility and recognition in the debate. Votes will be used where clear and uncontested information about the points of the debate cannot be found, as a way of substituting the missing information with the intuition of the crowd. But only on the basis that it is followed up with actions and experiments to validate the crowd's intuition against it's alternatives. --CauliflowerEars 11:50, 26 June 2008 (UTC)

Remember that votes here are meant to let the software track whether or not there is consensus within the relevant community, NOT for their usual purpose of breaking deadlocks - in every proposal thus far one or two people have the capability to hold up a proposal if they think it necessary to do so, so only a proposal that is at least consensus will be allowed to pass. However, without some filtering mechanism (for instance the proposed connection/history/reputation rating) affecting that process, anyone in the world can conceivably enter the debate and stop it in its tracks, just because they feel like ruining someone's day. People like that exist. Do we want the votes to be used to determine "majority rule" when no consensus can be reached, as opposed to detecting consensus itself? And if not, how do we deal with this issue? --Humphrey 15:29, 26 June 2008 (UTC)

Does not the idea of synthesis score and consensus score (as currently defined on the Main Page) provide a mechanism for driving people toward a consensus? — Ed Pastore 00:39, 27 June 2008 (UTC)

It does provide a mechanism. I agree with that. I was just responding to CauliflowerEars's comment that seems to think votes would be used in a simple "majority rules" fashion rather than in the way we had described before. -Humphrey 04:37, 29 June 2008 (UTC)

Yes I think it goes a long way to doing that. Thanks for pointing back to it Ed. I'm slightly struggling to put the meat on the bones of the idea tho and I'd love to see a "for instance..." example of how these might play out in practice. It might be useful to highlight the differences between how it will work and the traditional compromise and majority rules scenarios.

I think it also needs some notion of bias towards content in synthesis resultions which uncovers and addresses the root causes of issues rather than just their downstream effects. If that's not contentious, I'll have a go at coming up with some appropriate wording for that for us to play around with. --CauliflowerEars 20.40, 30 June 2008 (UTC)