Talk:InstaGov/introduction

Leftover Bits which may be just rambling

I wanted to say something about the importance of establishing beliefs.

Advocacy groups take the beliefs for granted, and organize their actions around those beliefs. For example: can we get together and agree, once and for all, that there's nothing wrong or immoral about homosexuality? Or, at least, that the only arguments against it are based in irrationality? It seems to me that a group of people coming together without a specific agenda and deciding on a few basic truths they can all agree on would have far more sway than the same statements being issued from advocacy groups dedicated to those specific causes (you know -- those "single-issue groups on the left") -- even if the advocacy groups have more members than we do.

The obvious retort to a "single issue" advocacy group: "Of course you believe that; that's where your support comes from."

Furthermore, there has been a history of such groups becoming corrupt and extremist; I assert that this is in large part because their value documentation (typically a "mission statement") has not been maintained -- updated and clarified as inaccuracies and vagaries become apparent, and as people's ideas about what is appropriate evolve over time.

If a group's mission is, say, to prevent cruelty to animals, this can morph through "preventing needlessly cruel scientific experimentation" into "preventing all animal experimentation of any kind", "releasing all test animals from captivity", and even "vandalizing animal research facilities and sending threatening messages to the researchers". It seems to me that using InstaGov (or something like it) to manage such a group would have allowed moderation (the original goal of preventing cruelty) to keep the fanaticism in check (e.g. by pointing out that freeing animals raised in captivity is no kindness, or that the increased security necessitated by "animal rights" incursions has led to less pleasant living conditions for many animals not to mention wastage of animal lives as butt-covering regulations prevent experimenters from re-using animals in multiple experiments.)

--Woozle 15:05, 11 June 2008 (EDT)

Interested

Hi,

I am planning to devote my energies along similar lines. In general, I want to provide tools which greatly lower the barriers of entry into political knowledge/discussion.

I'd love to hear your methods of collective decision making. I'm concerned that trolls or interest groups can control issues.

Methods of Collective Decision-making

Yeah, I need to spell that out more. I'll sketch it out here and later write it out more formally... Note that my basic plan is to start small, build gradually, and see where things need tweaking as we go along. I strongly suspect some of the problems I've worked out elaborate counters for will turn out not to be problems, and I'm sure that problems will emerge which I haven't anticipated. The key thing is to have everything well-documented, so we can reconstruct what went wrong when something does.

First: Of all the voting methods known to science, winner-take-all (WTA) – the one used in the overwhelming number of elections everywhere, to the point where it's what you think of when you hear the word "voting" – seems almost purpose-made to be abused by powermongers. It favors two strong groups over all others, forces people to choose "the lesser of two evils", results in really weird choices for winner if there's a lot of disagreement... etc. I can imagine that WTA had the sterling advantage of being possible with the information tech of the 1700s (no computers, no calculators, no long-distance realtime communication of any kind), but I think we can do better now.

As I understand it, range voting – basically rating each proposed solution to a given problem on a scale of approval – seems to actually represent the wishes of the voters the best, resulting in the smallest amount of error between what people say they would like and what choices actually win.

So the first part of my solution would be to use range-voting instead of WTA. It's far less prone to manipulation overall, rendering ineffective many of the manipulation techniques commonly used in WTA systems (e.g. gerrymandering).

Second: Secret ballots were originally instituted for two reasons that I'm aware of:

• To prevent voter intimidation ("You vote for X or I'll kill you!")
• To prevent vote-buying ("Vote for X, and I'll give you $50") or worse ("Vote for X, and I'll see about getting those charges against you dropped").

While these are both good reasons, I suspect that the consideration of accountability may outweigh them – as demonstrated by the rash of GOP-driven voting machine fraud in the last two presidential elections (2004 and 2000) – if we can come up with some other ways of dealing with that sort of manipulation. (I also mean accountability in another sense: I think we'd get a lot more meaningful dialogue between people if everyone's vote were on public record. I can see a lot of immediate objections to this, but none of them really hold up under scrutiny. "My opinion is my own, and my vote is private!" Well, why? How is this good for anyone? I suppose certain in-groups might start putting pressure on people to vote a certain way -- but that needs to be talked about, not hidden under a rock.)

My immediate proposal there is this: rather than discouraging the trading of votes for money/influence, make it much easier to do it openly than covertly. This way we can monitor who is buying the influence, and see if it is causing problems. The system I'm designing has a specific function whereby any voter can designate another person to vote for them. There are legitimate reasons to do this (some people don't want to deal with the issues, but know who they trust to make decisions in certain areas) as well as illegitimate.

The existence of records showing who-voted-how directly addresses your trolling concern (if by "trolling" you specifically mean sock-puppetry; if not, please elaborate?), in that someone would have to have multiple accounts in order to vote multiple times. For trivial security, perhaps we'll make IP connection data for each user available; a troll using a proxy would keep logging in from different addresses, and a troll not using a proxy would be using the same IP (or group of IPs) for multiple accounts. Either one could be a flag to trigger further investigation. Another way to handle it would be to require voter verification by mailing address -- more expensive, but at some point this thing will have to have a budget (and indeed should ultimately be involved in deciding the fate of quite considerable sums of money). A troll might plausibly register more than one account at the same address, but again there could be further verification in this case and some kind of penalty to discourage gaming the system.

It's an experimental idea, of course, and I'm open to suggestions for the details as well as suggestions for other solutions.

Third: The next problem you run into, once you've got a basic system for posting issues to vote on, collecting poll data (votes) and tallying up the result, is managing issues -- if anyone can suggest an issue, or a candidate solution for a problem (whether the problem is a practical one or "who should be president?"), then we're likely to end up with something like a 1/x distribution for issue significance: a few very important things, and vast mountains of utter trivia (or stuff that is only important to a small number of people).

The "traditional" solution would probably be to create some sort of group to review issues and decide which ones were worthy of being voted on, but I really don't like that; it brings with it a lot of the traditional baggage of the way we do things now. Instead...

My proposal here is to create a taxonomy, and let people "mod" issues into slots in the taxonomy, i.e. vote on whether they think given issues are important to a given taxonomy slot or don't belong there, also using range voting. There's also a feedback system to discourage people from mis-rating stuff and reward people who accurately gauge an item's significance to a given slot.

Individual voters would then "subscribe" to only the topics (short for "taxonomy slots") which interest them or which they care about, and leave the other issues for other people to decide.

Possible pitfalls with this:

• Do we need to have a way to force-override the individual choice not to pay attention to a certain issue? Like, to grab people by the throat and say "Look, I know you're not interested in this topic normally, but THIS IS REALLY IMPORTANT!!!" My tentative solution is that topic-subscription will be a threshhold kind of thing – you'll only see issues which meet your threshhold for their assigned topic (which makes sense anyway; people have different levels of interest in different topics) – but there will be a maximum cutoff you can set, so that if an issue is modded as REALLY IMPORTANT!!! by enough people, then everyone will see it. I dislike having to set an arbitrary level like this, but then again anything arbitrary like this can be put to a vote if people think it needs adjusting; the system can help maintain itself.
• Should we allow issues to be modded into more than one topic? My inclination is to say absolutely yes, but there may be reasons not to do this.

Fourth: We will probably need some additional auxiliary tools. Issuepedia essentially as it is now is one such tool; another would be a "debate engine" -- there are several web sites providing features like this, but they need work and they need to be part of a larger context such as Issuepedia where decisions reached in a debate can be used as evidence in other debates, and perhaps inserted as sort of "live truths" within other writings. You can see some static, incomplete examples of this idea in the category:debates category.

I have to keep stopping myself from making it too elaborate, because ultimately any elaborations should be voted on; my "job" is to get the basic system up and running so we can start doing stuff with it. I want to anticipate any problems which are likely to occur early on, but beyond that it should be up to the users (and the experience of actual usage).

Hope that makes sense... I've begun the technical documentation here, fwiw.

--Woozle 18:58, 29 July 2008 (EDT)