Sunir is trying to formulate a MeatBall:BanningPolicy. Do we need something similar?

How about framing it as a constitution, describing the rights and duties members have. I would try to make it very general, open to interpretation, and not specific enough to try and game it, and without too much details. It should be very short. Detailed rules will always be tested and will provoke new conflict on their own.

My take:

  1. Everybody has the right to create, edit, and delete pages on this site and thereby join the community.
  2. The site exists to support the MissionStatement. The community has the duty to not stray far from the mission, or to change it in order to reflect new priorities.
  3. Participating should be a rewarding experience for the entire community. The community has the duty to maintain an environment where this is possible.
  4. Measures the community can take to carry out it’s duty include removal of content and banning.
  5. Grave decisions shall be made only after due consideration. There shall be no voting; instead members in good standing shall offer their opinion and reach a consensus.

I’d like to avoid the introduction of to much process at this part, because the number of “members in good standing” has always been very small – never more than a dozen.

I’d also make it clear, however, that many things are not ok that I don’t have to list explicitly. If a member in good standing does not like it, then that’s reason for concern enough. Something needs to be done (or learnt).

Basically this should be a document that we can use to help argue for some decisions such as reverting or banning.


5 second skimmed the doc here’s my gut reaction not really well thought out though:

  • voting: BayleShanks is working on voting stuff, and talked about the ability of a people to make it clear to observers “yes, we really think this way” – I thought it wasn’t a bad idea, but am undecided right now
  • our “mission” is kind of fluffy, hard to know how it would be violated, but, ok
  • I like the part about grave decisions and due considerations.

I’d like it if we said: We’re a wiki about communities, but that doesn’t mean we practice RadicalInclusiveness. We see no obligation to be infinitely open minded; We are SelectivelyOpenMinded.

I’m worried about the last numbered point (the only one, by the way, that addresses how things actually get done, rather than a philosophical statement of principals). My read of that one is that no one can be banned, and nothing else important can be done if there is not 100% consensus. This differs from the status quo, in which Alex has the right to be a GodKing as a last resort. 100% consensus may be fine for now, but it won’t scale if the community gets big. And if we get a constitution, amending it would itself be a “grave decision” so we could be paralyzed (of course we could always fork in that case).

In general, I think going halfway with formalization is pointless. I think we should either not have any constitution, or we should have one that defines a formal process (including defining “due consideration” and “member in good standing”). If we remain a small, cohesive group, then I think we will all agree with almost everything said above anyway should a newcomer every challenge it. But if we ever have a major argument, the above constitution probably isn as’t specific enough to help us decide what’s proper.

So, the only result of adopting this constitution, rather than sticking to the status quo, is to add verbiage, and to strip Alex of his emergency martial law power. I prefer the status quo1.

I think I agree with Bayle. Perhaps we just need to state that so that newcomers are not surprised when a crisis comes up.


This was written 1 year and some months ago.

I rediscovered it because I was surprised that, in the rendering of CommunityWikiGovernmentIrcConversation, this term linked– I had not expected that. (We were just brainstorming names of what an alternative to “CommunityWikiGovernment” would look like.)

“Perhaps we just need to state that so that newcomers are not surprised when a crisis comes up.”

Well, yes. We should write down something.

Here’s another initial suggestion (see “Proposal 2”, below).

Summary: The goals of this constitution are to allow us to have an official membership list and to create a set of official rules.

This proposed constitution makes it hard to legislate rules, very hard to amend the constitution or to change the membership list, but easy to repeal rules. It does not in itself specify anything “actual”; it deals only with the “metaissues” of how to change the membership list and how to make rules. It is intended for legislation to be possible even when most members are either absent or don’t care to spend time paying attention to such issues. Since rules are much easier to repeal than to pass, you don’t have to worry that if you don’t pay attention to a “bad proposal” that you will forever be stuck with it (this can happen with constitutional amendments, but presumably they will be few and far between since they are hard to pass). LiquidDemocracy is incorporated, so you have the option of giving your “proxy” to someone else if you care about some issue but don’t have time to intensely participate in the legislative process.

Detailed differences between Proposal 2 and Lion’s “Proposal 1” (which is on CommunityWikiGovernment) (if you are just skimming then skip this part; it’s probably longer than Proposal 2 itself):

  • Proposal 2 spells out the procedures for voting in more detail.
    • See comment below, about “formality”
  • Proposal 2 restricts itself to specifying the legislative process, leaving most aspects of the social structure uncodified (if we want to codify this, I suggest we do it as Rules, not as part of the Constitution). In Proposal 2, only two kinds of people are codified: Voting Members, and everyone else.
    • See the comment below, about “rules”.
  • Time limits for becoming a Voting Member are not codified in Proposal 2.
    • See the comment below, about “rules”.
  • Proposal 1 requires consensus of Voting Members to amend the constitution or to accept new members. Proposal 2 requires only 90% agreement.
    • If there are 17 Voting Members (as in the list Lion proposed), this means that at most 1 of them could disagree, and a measure could still pass. I am comfortable with this; if it was really 16 vs. 1, I think that 1 person would be obligated to drop the issue even in our informal notion of “consensus”. Allowing less than absolute agreement is necessary to make the system scalable, which will be important if we intend to continue to allow in new members.
  • Proposal 1 allows 10 members to declare a “rush vote”. Proposal 2 states that enough “Yes” votes are needed such that even if everyone else voted “No”, the proposal would still go through.
    • I am not particularly tied to this method of rush voting, but I don’t think we should have an absolute # of people like 10, because that’s not scalable.
  • Proposal 1 doesn’t make any provisions for making any rules outside of the constitution itself. Proposal 2 provides a rule-making process that requires less consent than is required to change the Constitution itself.
    • See comment below, about “rules”.
  • Since the only way to make rules in Proposal 1 is the same way that you amend the constitution, it takes a consensus among everyone present for 2 months (or 10 members to rush) to do that in Proposal 1. In Proposal 2, it is easier to change a rule; only 80% of everyone present for 2 weeks is needed. Proposal 2 also makes it even easier to repeal a rule; only 50% of everyone present for 2 weeks.
  • Proposal 2 incorporates LiquidDemocracy, that is, if you won’t be around for awhile or don’t care to follow certain issues, you may delegate your vote to other people.
  • Proposal 1 has a “quorum” of 2 affirmative votes. Proposal 2 has a “quorum” of ceil(sqrt(members)) affirmative votes. With 17 members, this is 5.
    • Again, I think this is more scalable.

Comment on formality: I’m sure those who don’t like formal voting will find this proposal odious in that it spends a lot of space detailing the formal minutia. But I think if we’re going to do formal voting at all, then we should (1) be precise, and (2) specify ways to change the system. And I don’t see how to have a formal list of “authoritative” members without some kind of formal voting (even consensus legislation is formal voting, it just has a 100% threshold). Note that Lion’s “proposal 1” also had a formal specification of how long voting was, how to do a rush vote, what kind of votes may be cast, how to amend, etc; I just don’t think you can get away from this sort of thing. Most of the verbiage in Proposal 2 seems to me to be necessary either for precision or for specifying a procedure to amend the constitution and the membership list.

Comment about rules: I think it’s important to keep a constitution as small as possible. Therefore, I have included a procedure for making rules which are separate from the constitution. This is important for a few reasons. First, it makes it easier to read. Second, it lets people distinguish between “issues” and “metaissues”, since some people think issues are worth more attention than metaissues, and other people think metaissues are worth more attention than issues. Third, you have to make it hard (in terms of time spent convincing everyone) to amend the constitution or to expand membership (which amounts to the same thing), but you may not want it to be so hard to change “non-meta” rules. Separating the two allows you to do that.

(a day later) I thought of another way to justify that. In general, there is a hierarchy of flexibility in types of rules.

Rules encoded into software (HardSecurity) are the most inflexible. They are hardest to change or to break. The constitution is the next step down. The next step down are ordinary laws (what I have here called “Rules”). The lowest rung is composed of social conventions.

As you go up the hierarchy, rules become more precisely worded, more time-consuming to make, more prone to discussions about nitty-gritty, and less flexible in application. A social more is flexible and can be easily bent when there are extenuating circumstances, but a constitution cannot be bent and must instead be formally amended, even when there is general agreement that extenuating circumstances exist. Hence, one can successfully exploit “loopholes” in a formal constitution (situations where it is clear to most observers that the letter of the law differs from its intent); this cannot happen with social mores because the shared understanding of the social more is paramount (even if someone wrote it down somewhere, that’s understood to just be a description of the more, not the thing itself). Hence, constitutions lend themselves to discussions about the nitty-gritty of their wording, which is a necessary but undesirable way to spend time. Hence, care must be taken when crafting a constitution to make sure it is worded correctly.

So, basically, as you go up the hierarchy, everything becomes a big pain. So I think you should keep as many rules as possible as far down as possible, i.e. make most things social mores, a few things laws, and very few things in the constitution.

So why have a constitution at all? Because without some sort of formal system, there is no agreed-upon way to interpret when (and if) a group of people has “spoken”. So some commandment is written on a wiki page. All that shows is that the writer wrote it, and that no one who has visited since has erased it. Does the writer have the authority to speak for the group? What if most of the group would have disagreed, but they are on vacation? Is the group that represents the wiki the set of people who have visited it in the past week, or some other set of people? What if most of the people visiting the page disagreed with that sentiment, but didn’t think it was worth arguing about? Does that mean it does express the will of the group (because a bunch of people who didn’t care about the issue consented to have it decided by the few people who did), or that it does not (because the majority disagrees)? What if most of the people who visited the wiki recently didn’t visit this particular page?

These sorts of questions lead to a lot of time being spent arguing over what the wiki’s “Community” has agreed even in an EngagedCommunity?. When there is an AbsentCommunity, I would imagine that the additional arguments over whether the absent members should even have any say, and if so, whether anyone present can be qualified to speak for them would make the continued maintanence of any rules at all difficult.

So, that’s why you have a constitution at all; to construct an unambiguous/formal, abstract entity called “the community” so that “the community” can say things.

But, given that, you still want to minimize use of the more inflexible tiers of rules. Which is why I resist encoding any social mores at all into the constitution except what is needed to define group speech (because I see the constiution’s proper role as merely a formal system for generating group speech), and why I think we should make as few Rules (laws) as possible, and still do as much as we can with informal social mores.

Proposal 2


This is [a proposal for] CommunityWiki’s Constitution. Only CommunityWikiMembers? may edit this page For discussion about the Constitution, please see the page ____.

Who is a Voting Member

The list of Voting Members of CommunityWiki may be found on the page ____. The list of Voting Members is by no means a comprehensive list of every valued or longtime contributor to CommunityWiki.

How Voting Members may be added or subtracted

The list of Voting Members can be changed by the same process as that specified for Amending the Consitution.

In addition, any Voting Member may unilaterally Resign, which means they are taken off the list of Voting Members.

How this Constitution may be amended

An 90% majority is needed to pass an Amendment. A proposed Amendment must normally stand for at least 2 months before being passed.

See Legislative Minutia, below.

How Rules may be passed

An 80% majority is needed to create a new Rule or to amend an existing Rule. However, any existing rule may be revoked by a majority of only 50%. A proposed Rule change must normally stand for at least 2 weeks before being passed.

See Legislative Minutia, below.

The mechanics of voting

To propose a constitutional Amendment or Rule, a Voting Member places their proposal on the OfficialLegislativeProposals page, being sure to note the date. Members are encouraged to air their proposals informally on the UnofficialLegislativeProposals page first.

To cast a vote, a Voting Member adds their name underneath the relevant proposal, in the relevant list (Yes, No, Delay).

To grant a proxy, the giver notes who receives the Proxy, and under what conditions, next to the giver’s name on the page ____ [the same page with the membership list].

To vote a proxy, the Member who holds the Proxy casts the vote by writing “for (original person’s name), cast by (receiver’s name), proxy chain: (chain by which the receiver came into possesion of the proxy vote)”. The receiver is responsible for keeping track of which proxies they have accumulated.

Legislative minutia

Any Voting Member may at any time propose any constitutional Amendment or Rule. The time elapsed since a proposal was added to that page is said to be the amount of time that the proposal has been “standing”.

Underneath each proposal will be a list of “Yesses”, “Nos”, and “Delays”. These lists start out empty.

For each particular proposal, define the number of Participating Voters to be the sum of Yes, No, and Delay votes, including votes cast by proxy.

A proposal passes when both of the following conditions obtain:

A proposal fails and is removed from the list when both of the following conditions obtain:

Note that the effect of Delay votes is that, in order to pass, a proposal must be strong enough so that even if all “Delay” votes were turned into “No” votes, the proposal would pass. In order to fail, a proposal must be weak enough so that even if all “Delay” votes were turned into “Yes” votes, the proposal would fail. In other words, to cast a Delay vote indicates that the voter reserves the right to make up their mind later; the proposal is not decided until either they make up their mind, or their decision is rendered irrelevant to the passage of the proposal.

Voting Members may give other Members their Proxy for any subset of issues. This allows the receiver to cast the giver’s vote for them on those issues. Proxies are transitive; that is, a Voting Member may forward a Proxy that they have received along to another member.

Any Member may change their vote or take back their Proxy at any time until the conclusion of a vote (which is defined as when the proposal passes or fails).

Thresholds for votes to pass and for affirmative quorums which are fractional are rounded up; for example, if 90% of 9 votes are required, then 9 votes are required.

These rules are for proposing and conducting votes for changes to the constitution and rules, and are not suited to other sorts of elections, where there are multiple choices to be voted upon. If we ever want to vote on something else, such as officers or budgets, the constitution should be amended to detail the vote couting system to be used.

Just a technical note - I think the timing here is not precise enough. I particular - it states that a rule passes after it is ‘standing’ two weeks, but when you start the counter?

I think these proposals above great and well thought out procedural mechanisms. Yet, I see only 2 different SocialNorms? explicitly stated in both of these proposals, and they are LiquidDemocracy and SelectivelyOpenMinded. I think we really need to state explicitly, and collectively, the SocialNorms? that we intend to live by.

We should recognize and understand why we want, or don’t want a given set of SocialNorms?. For instance, SoftSecurity social norms are apparently desireable because they potentially create an environment where people AssumeGoodFaith, AvoidConflict. These SocialNorms? also DevolvePower, encourage PeerReview, trust, etc.

I think I was attracted to SoftSecurity oriented SocialNorms? because, when they work, they avoid the need for lots of formal procedure. However, I’ve discovered in different ways that they don’t work without an EngagedCommunity?. SoftSecurity breaks down with AbsentCommunity.

So, when Lion first started talking about some of the issues that led to discussion of CommunityWikiGovernment and revived CommunityWikiConstitution, I think that a central issue is and was:

What do we do, or how do we handle AbsentCommunity?

Answers to LionKimbro’s concerns in the middle of CommunityWikiGovernment:

  • I’m not clear to me: Is “Delay” the same as “Abstain?” Is it the same thing as “I haven’t made up my mind yet?” When a vote starts, is everybody automatically in delay?
    • No. I rewrote some of the Legislative Minutia section to clarify – does it make more sense now?
  • Also: Proposal 2 has a “quorum” of ceil(sqrt(members)) affirmative votes. With 17 members, this is 3. Nr.r…. Wouldn’t that be 5? Because, sqrt(17) = 4.12, and the ceiling is 5..? You need 5 yes votes to pass..?
    • Oops, thanks, you’re right. In an earlier draft I was thinking of putting sqrt(sqrt()) as the quorum-ish thing for Rules, and sqrt() as the quorum-ish thing for Amendments. But then I thought that was too complicated, so I took it out. But I forgot to change that 3 back.
    • On the one hand, 5 sounds like alot. On the other hand, if all 17 of the proposed Voting Members really desired to be active in the legislative process whenever they were around, then it’s not a lot. Also, if even a few people use the LiquidDemocracy system to give their vote as proxy when they are away or don’t care, then 5 sounds easy enough. Since it is so much easier to repeal than to pass Rules, I expect that people won’t be too afraid to do this.

ZbigniewLukasiak: thanks for catching that, I added a sentence defining “standing”.

SamRose: Yes, I think you have hit on a good, concise way of saying what the central issue is here that we are trying to solve. To paraphrase you, “SoftSecurity doesn’t work without an EngagedCommunity?. What do we do, or how do we handle AbsentCommunity?”.

The method that the above constitutional proposal tries to use is to make an explicit list of Voting Members, and to give those members the power to make Rules which stand even when those members are on vacation. This is definitely a choice, and I think it is even a choice different from the “default”; I think the SocialAffordances of wikis promote a social structure where power is vested solely in those present in last few weeks.

I completely agree that we should try to clearly and completely detail the SocialNorms? that we are following. First, this makes things clearer for newcomers. Second, it is an interesting research project. Third, it allows us to propagate useful social norms in a modular, “prepackaged” form to other communities (WikiProcess). Forth, it exposes hidden disagreements between us.

However, I think that whenever possible, we should do this informally, on regular wiki pages, rather than legislating Rules. I feel that it is even more important to keep substantive (rather than procedural meta) rules out of our Constitution as much as possible. See the section above under “Comment about rules” for my reasoning. I just added a large section to that part that explains it more.

So, I don’t think substantive issues should go in the constitution at all. I think we should have a rule-making mechanism, but use it sparingly. I think most stuff should remain informal.

I see what you mean, Bayle. For instance, they did not put these substantive statements into the US constitution. Instead, they put them into a seperate “Declaration of Independence”. DoI?, states things like “all men are created equal”, etc, and these appear to guide the formal laws in the constitution. But the constitution itself does not make these declarations. So, we may also benefit by updating our MissionStatement, or some other delcarative document to state the substantive views that we try to follow. This can help us both in our informal interaction, and in our formal decision making, should it become necassary.

I wrote some more about your proposal, and Alex’s latest comment on CommunityWikiGovernment, at CommunityWikiGovernmentChat2.

I have contradictory feelings about writing a constitution for CommunityWiki.

  • Pro: When newbies first show up, it is a good thing to tell them how things work around here, rather than let them wonder and make up conspiracy theories. ImportanceOfCommunityBuilding
  • Anti: On the other hand, haggling and debating the by-rules is boring.
  • Pro: If we write a really good, easy-to-read constitution, we could point at it as a shining example of the success of PlainTalk.
  • Anti: We do want people to contribute good things here <3 <3 <3 , even before they get Official Approval from N people, and read 20 pages of rules and regulations, and wait out a N week waiting period.
  • Pro: The bigger a community gets, the more “precedent” accumulates. Better to summarize the common understanding in one place. If the only way to understand things is to read the entire history of the community, that’s annoying. That summary makes a good place to try to persuade people to act in certain ways in the future, even though we were doing things differently in the past.

“it won’t scale if the community gets big.” I agree that to scale this wiki up to 1000 people posting a week, there are some things we have to do differently, no matter how boring and annoying they may be.

There are 3 ways we can design this constitution:

  • Assume that more and more people is a good thing. Try to design things that scale up indefinitely. (Most people building discussion forums seem to attempt this. However, AlexSchroeder points out “A wiki for a group of friends is good enough … Therefore I don’t plan to scale.”).
  • Throw together something much simpler and easier to use, that works fine with the number of people we have now. If we grow so big this doesn’t work any more, patch it up.
  • Somehow LimitGrowth to (say) a dozen or two active posters. Design things to work with that number of people. (I think it would be interesting to talk about good ways for a wiki founder to do this, even if we decide this isn’t appropriate for CommunityWiki in particular).

Alas, it seems like coming to a consensus on “What’s the official CommunityWiki process?” is taking a long time, because I’m not the only one to get side-tracked into talking about “Wouldn’t it be interesting if at least one wiki used this other process?” and “Here’s what I think about constitutions in general, on every wiki”.

I personally don’t care about growth. But I do care about the constitution. So if Bayle has provisions for future growth, I basically just ignore them: I just want to see a thing ratified. I don’t care if it’s perfect, or even really really good.

Something that says:

  • Here are our members.
  • Here’s roughly what we think.

Perhaps optionally:

  • Here’s how we make rules.

I think that it adds to the authority of the constitution, if it says how it is amended and so on.

But I do not care much about the details here. Our constitution is mainly about insiders and outsiders and authority on the wiki. It is not so much about how we treat other insiders. We already do that fine, for the most part.

I don’t forsee a lot of talk about rules and regulations. If they do happen, though, and you are uninterested, just don’t read them! It’d just be a rehashing of what we already do.

I do think we should perhaps get rid of the concept of “Rules” in the constitution, and include the concept of “Endorsements.”

For example, TurnBasedVsInterruptedThreadMode is how we do things here. It’s not necessarily a rule. But if a foreigner comes by, and we say, “Use TurnBasedVsInterruptedThreadMode” here, I would like that page to say something to the effect of: “Yes, this is in fact a CommunityWiki Endorsement. Treat it as fact, if you are a newcomer, if you wish to respect our policies.”

Yes, it doesn’t need to be perfect, as long as we can improve it later.

So, do you picture one big page with a list of all the “Endorsed” procedures? Or perhaps a tag at the bottom of each one, indicating it has been endorsed? Or perhaps a short list (in some standard format) of the people who endorse each policy, and when enough people add their own names to the list on one page, that page becomes an Official Endorsement?

I suspect I’m going to regret ever mentioning this. The most readable “Official Policy” (perhaps because it is the funniest Official Policy) that I have ever read on a wiki is the "Encyclopedia of Stupid style guide". (It has a “Consequences of Not Heeding This Document” section – do we need one also?)

Just to add a ew of my own cents…I think that our activity here in CW in general tends to limit growth, because it requires a level of engagement. I think that Lion has a good idea with just having us all read Bayle’s CommunityWikiConstitution contribution, and either sign it, or propose changes, etc, and basically work together as a group until we collectively come up with something we are all willing to sign. Because, I agree that we need a social contract, and we need authority of the group in order to work in a decentralized way in an open environment.

The Summary Text


Do we need one? Probably not. Alex has emergency admin powers. Perhaps we just need to state that so that newcomers are not surprised when a crisis comes up.


constitution proposal 2; timing not precise, What do we do, or how do we handle AbsentCommunity? Fixed timing inprecision. Clarify that proxies count towards affirmative “quorum”, that votes may be changed. Clarify delay votes, rewrite Minutia a bit. Bayle answers Lion, Sam. Sam likes Bayle’s take on things. Endorsements rather than rules?


1. actually, as I noted on WikiVoting, eventually I’d love to be in a wiki with a more formal process, because I’d like to experiment with scalable group decision-making. But I realize that others here don’t share my excitement. And it’s not necessary to do that as long we remain small and forkable. So let’s do without a constitution

Define external redirect: CommunityWikiMembers SocialNorms EngagedCommunity DoI

EditNearLinks: DevolvePower AvoidConflict PeerReview GodKing AbsentCommunity HardSecurity