Wikipedia talk:Governance reform/Archive 5

Archive 1

←

Archive 3

Archive 4

Archive 5

The Wiki way, much less Wikipedia way, is community self-governance

I completely agree with the sentiment of "natural consensus, polls, and no committees" expressed earlier.

Self-organizing and self-governing committees have always been the Wiki way. Governance by a select committee is the very antithesis of the Wiki way. For any Wiki the superordinate goal, that is the group goal, as opposed to an individual goal, is to continue the Wiki's existence. Whereas a committees' superordinate goals are often at odds with both the Wiki's and those of the community. Not to mention committee's adding another layer of bureaucracy and tending to attract those pathologically drawn to power.

I'll take a self-governed community over one controlled by committee every time, despite the inevitable, but healthy, difficulties. Barring any new ideas more compelling than what I've seen here already, I'll be opposing this effort if it comes down to a vote, er, comments. FeloniousMonk (talk) 03:04, 6 May 2008 (UTC)

I agree with FeloniousMonk. The relative inability of the community (and ArbCom) to solve the problems being addressed is due to an ossifying (or complication) of the consensus model, by-the-letter interpretations of the rules and excessive bureaucratic sprawl. More of the same is not a real solution and will only serve to exacerbate the problems and delays futher. Vassyana (talk) 01:47, 8 May 2008 (UTC)

Why not just do it?

A proposal has been made that would follow the wiki way and at the same time create a centralization of advice, if used in that way. This was WP:PRX, which was quickly marked Rejected, accompanied by serious attempts to delete, erase, and possibly salt it, based on a radical misunderstanding of the proposal, apparently. The proposal was not to establish voting, but to establish -- merely allow, not require -- users to name a trusted "proxy." The name may have been misleading because of various uses which proxies can be put to, including voting, but no actual use was part of the proposal, which was only about the technology.

If a segment of users names a proxy, and there is a central proxy table showing these assignments (or the data can be easily collected in some way), then it becomes possible to estimate consensus by assuming that a vote or position taken by a proxy, upon consideration, is, on average, more likely than not to represent the position of the "client." There are more possibilities as well. This was a proposal for "delegable proxy." Which can be used as an election method for a proportional representation assembly.

I'm opposed to the direct exercise of power by proxies on Wikipedia; however, there are at least two major problems with the community consensus model. The first is that "consensus" is not defined. In some organizations, it means total agreement, which is obviously desirable, and very small groups can often work out total consensus, and process for doing that is fairly well understood. But that process requires a lot of communication, which takes a lot of time, and medium-size organizations (it starts to happen with roughly twenty participants or less) can find themselves taking extraordinary amounts of time making decisions, and burnout starts to occur, especially if the situation goes on for years. With large organizations it becomes impossible for consensus to be negotiated directly. The classic solution is representative democracy through elections, but, as it is normally implemented, it, shall we say, violates the wiki way. This method can work, obviously, particularly if the election method is good. Among the best of those in common use is Single transferable vote; one might note that I'm a prominent critic of Instant-runoff voting which uses the same counting method as STV for single-winner, where the small problems of STV loom large, but STV, under good conditions, can work almost as well as what I'm going to suggest. A better method, simpler and more suitable for our purpose, was proposed in 1886 by Charles Dodgson, better known as Lewis Carroll. Dodgson was working on proportional representation, and he was familiar with STV. Apparently he realized that vote transfers could take place, rather than according to a preference list on a voter's ballot, but according to the discretion of a candidate who has received votes. And then a candidate who receives, directly or indirectly, a quota of votes, is elected.

What is remarkable about this method is that it is not oppositional. Voters, acting through representatives whom they unconditionally chose, cooperate to create seats in an Assembly. Delegable proxy can be used as the method of collecting votes. If we simply assume that editors name the person they most trust, among all the other registered editors, we can then analyze these assignments and create an assembly as a standing body. "Election" is continuous, because editors may change their proxy assignments at any time. There are quite a few obvious objections, but, let me assure my readers that there are solutions to the obvious problems.

How would the rules of the Assembly be determined? Traditionally, an Assembly makes its own rules. It starts with a General Assembly, where all eligible members (which would, here, presumably be all registered editors, perhaps registered for a certain minimum time, or with other requirements) meet and determine rules. Given that Wikipedia has already grown to a size where this meeting could be itself intractable )witness this page), I'd suggest that:

WP:PRX be re-examined, and the simplest and most efficient method of naming proxies and collecting the proxy information be chosen. There were two methods on the virtual table when the proposal was abruptly rejected, both involving transclusion of proxy files in user space: a fairly complex system proposed by Sarsaparilla (now blocked) with some help from Mangojuice, which has a central table and additional tables in other locations didn't work, and a more distributed technique that I set up, where the proxy files are simpler and a proxy table can exist anywhere, i.e., there can be many of them.

An Assembly page be created, with certain proposed rules. The first of these rules would be that the main Assembly page not be for debate, but for report. Rather, for any issue to be discussed or debated, a subpage is created, with a link from the main Assembly page.

These subpages are virtual committees. Committees would seek consensus among their participants, and proxy assignments would be considered in the estimation of consensus. Committees don't make decisions, they make recommendations, and each recomendation is accompanied by a report of the vote status of the recommendation or recommendations (there can be many alternate drafts, each with its own vote, but there would be, following standard deliberative procedure, always a central, main text, which is amended by, yes, majority vote. However, the complete report of the committee is not limited to that single draft.

Debate would be kept off the Assembly page *and* the Assembly talk page, except within certain rules that the Assembly would establish. It is not only unnecessary to set these rules in advance, it's a bad idea. The process I'm suggesting here is merely one of a practically infinite number of ways that the Assembly could operate, but key is that the Assembly has control over its own process, and it may take whatever steps are necessary such that its deliberations are thorough, complete, and represent genuine consensus as best can be estimated.

Voting, in this system, would never be undertaken until after consensus has appeared that discussion has been thorough. Under classic rules, it takes a two-thirds vote to close debate. In peer organizations where unity is important, and I'd suggest that Wikipedia qualifies, extraordinary efforts are made to ensure that consideration of any issue is complete, and a lot of effort is put into maximizing consensus in an attempt to satisfy everyone. But the decision to move on does not require consensus, and the majority is never dragged through tendentious debate without its consent. Consensus is desirable, but majority, properly, rules, and it will rule wisely if it seeks consensus.

So one of the defects of the current model is that voting and debate are mixed, and many votes are cast before all the relevant arguments have been made. Yes, sometimes people go back and change their votes, but ... since votes don't count anyway, many don't bother. I've argued that there should be no voting in AfDs, but only the presentation of evidence and argument, and it's enough, for this purpose, that an argument be made once. Making the same argument twenty or thirty times simply makes it harder to follow. *Then*, after the appropriate time has lapsed, and without objection, there is a vote. Which is only a vote, not another round of comment. It's a decision, not a debate, and the decision is made by each editor who votes. But does this control the outcome?

It shouldn't. It's *advice*, where the community advises the volunteers who will implement decisions. In other words, the basic system doesn't change, but the process by which consensus is discovered becomes clearer and more efficient.

The Assembly can create itself. Delegable proxy would facilitate this, as well as an understanding that the purpose of the Assembly is not to control, but to advise, on behalf of the community and for the benefit of the community. Actual control is exercised by those with the power to act as necessary. This is individual editors working on articles, administrators deleting and taking other actions requiring buttons, and officers of the Foundation.

As I mentioned, the Assembly would begin with an assumption that every editor is eligible to participate. As to voting in the Assembly, I'd suggest, that privilege would remain. The problems of scale don't come from voting, they come from deliberation and debate. The Assembly may decide to restrict who can post to the Assembly pages, and committees may decide likewise, whenever traffic becomes a problem. The decision is made, though, by vote, and everyone can vote. So, if it happens, people will be deciding to restrict their own right to argue before the Assembly. Delegable proxy helps, with special proxy assignments for committees, as needed (that's a feature of delegable proxy that wasn't described in WP:PRX -- a special proxy table can be set up that assigns proxies that override the general assignment. It's quite easy and simple to do, and nobody need name a special proxy: if they don't their general proxy stands, and, in any case, they may vote directly. Proxies are used to estimate the position of those who haven't had the time to participate directly.

(It could be thought that people won't vote to restrict their own right to debate, but, in fact, if we elect an assembly, that is exactly what we are doing, we are confining the right to deliberate (and to vote, probably, as well) to elected members. If such an assembly is created democratically, rather than being imposed, people will have indeed voted to remove certain rights from themselves.)

The classic, immediate response to proxies here and elsewhere for internet usage, has been "what about sock puppets?" It's actually not an issue. First of all, puppet masters don't like painting large signs identifying their puppets. Sarsaparilla named me as his proxy, and we were promptly checkusered. While that won't be routine, I'm sure, some level of suspicion is natural where an unknown user registers and names a proxy right away. But one of the reasons I wanted distributed proxy tables is that special tables can be developed that take into account such things as edit count or even choices made by the person seeking advice. The proxy assignments are used to set up a communications network that is used to develop advice, not control. And a good argument is a good argument whether presented by one or twenty people and the same is true of poor arguments. Socks, if there were *lots* of them, which seems rather unlikely to me, could distort the Assembly process decisions, which are necessarily made by majority vote unless we are to create officers (there goes the bureaucracy), but in a consensus system, mere majority vote can't do much, and suspicious votes, where the expanded vote radically differs from the vote of the participants, would surely create some investigation on the part of those who care, and suspicion would immediately center on a single person exercising many votes coming from suspicious users.

We have before us a means of creating an Assembly without elections, without bureaucracy, where small numbers may seek and express consensus and the degree of consensus is known far more accurately, and the discussion can be far more thorough, than at present. But, folks, it takes at least two people to start it, better if there are three. We had two, one was blocked. Any takers? Be aware, there are forces here that don't want true deliberative consensus to be found. Don't be surprised if they come out of the woodwork. I'd love to be wrong about this!

--Abd (talk) 14:37, 6 May 2008 (UTC)

• I hereby name Kim Bruning as my proxy in all debates unless I otherwise state otherwise. From this point forwards Kim Bruning's voice carries the weight of two Wikipedians, and I am now free to not get so frustrated with policy discussions. Hiding T 13:24, 7 May 2008 (UTC)

• ...you can't do that. Wasn't that idea absolute buried by the community? Lawrence Cohen § t/e 13:28, 7 May 2008 (UTC)

I think you'll find I just did. It is even part of policy, per WP:SILENCE, which explains an important facet of WP:CONSENSUS. I alos missed the policy which afforded you the right to tell me what I can or cannot do. Note I am not proposing anything preposterous, such as Kim now being subject to 6RR, with admin's having to block me or Kim at random if breaches occur. Please perhaps consider my response a rebuttal to that which precedes it. As well as positing the notion that in many areas me and Kim tend to agree and it makes more sense for Kim to express the view we tend to share as he has a better way with words. Where I disagree with Kim, you'll find I am soon quick to say so. It is either that, or a lot of "I think Kim has captured the flaws in this proposal well, to the extent that I agree with his conclusions" type postings. All the bst, Hiding T 15:43, 7 May 2008 (UTC)

This is too bizarre and I've brought it up on WP:AN. Lawrence Cohen § t/e 15:51, 7 May 2008 (UTC)

Quite obviously this is too bizarre. That's probably as far as you should have gone with it, but there you go. No real harm done. I would take care to read what I am actually saying, and see to what extent it goes. Hiding T 12:17, 8 May 2008 (UTC)

I thought about it and I disagree with the proxy idea, for two different reasons. First, it doesn't solve any problem. Either the decisions are ultimately made by the merit of the arguments (for better or worse, I don't like it, since it's too subjective, but I digress), and then it is irrelevant how many people are behind a given argument; so the proxy is worthless. Or the decisions are made on the basis how many people want the given decision, regardless of the argument (which is in fact democracy or voting); then the proxy idea has some merit (but see my second objection below), however, the people who propose it are against such system. So people who want proxying (Kim, Hiding, Abd) at the same time oppose the circumstances where it is actually useful (that's why I say it doesn't solve any problem).

The second objection is complexity. Say, Kim comes to a discussion and says, "I represent 2 people". How do I check that it is true? Kim can have it written on his page, but I would have to check if this was really edited by Hiding, and not by Kim. Now imagine, you have a discussion, and there is like 20 people, each representing 8 other users. Who will do all the checks? That's completely crazy. And don't let me even start when someone who is a proxy will also let another person to be their proxy, then you would have to check everybody all the way down. It's not humanly possible. Samohyl Jan (talk) 17:37, 7 May 2008 (UTC)

I want proxying? Oh ok. I'll take my cue from you then.

Nah, just kidding. In the mean time, I've stated on AN/I that I'll think about this interesting conundrum for at least 24 hours. :-) --Kim Bruning (talk) 19:04, 7 May 2008 (UTC)

Ok, maybe you don't support proxying, but that's irrelevant to my point. Samohyl Jan (talk) 06:23, 8 May 2008 (UTC)

• Sorry to have placed you in such an awkward position, Kim, but you are arguing the points well enough for me at this point. You certainly had my proxy when you marked the proxy proposal as rejected, per my not reverting it or commenting on it either way anywhere on Wikipedia until this moment. :-) Hiding T 12:17, 8 May 2008 (UTC)

The proposal, WP:PRX was marked Rejected by Kim, yes. He was expressing the status quo in doing that, though whether or not that was truly a consensus is another question. I still don't know if Kim understands the proposal, but I *can* say that, as it is commonly misunderstood, I'd be against it too.... WP:PRX did not set up any system whereby one editor could vote for another. It merely set up a system whereby editors could decide to express trust in another editor, and those expressions could be collected. If this expression of trust means "if I can't make a decision or take an action myself, I trust that this editor's judgment is likely to be better than my own uninformed choice, or at least not worse, or, at a minimum, better than average if I can't participate myself," then the information *could* be used to estimate consensus, but that wasn't actually the proposal, though it was mentioned as a possible use. It was just to set up a uniform means of expressing the trust. How that was to be used was to be up to the community, to experiment with. Given that it is a simple voluntary action, binding nobody, my sense was that this wasn't a proposal that *required* community consent, and so edit warring over what was moot seemed rather silly, even though it was clear to me that what was being Rejected wasn't the actual proposal, nor did it make sense to waste everyone's time with formal dispute resolution. And that's still true.... So it was quite amusing, I'll say, to see an ANI report over Hiding's comment. As if. Something about WP:PRX seems to bring out a level of strangeness in people that I find striking. The massive effort to actually delete a rejected proposal was certainly bizarre, as was the similar effort to delete the working files. There was also the quiet deletion of a related user talk page, likewise unusual. --Abd (talk) 15:39, 9 May 2008 (UTC)

The major disadvantage to my eyes is the potential abuse of a proxy system. If I had 40 proxied behind me, who basically trusted me, I might be able to effectively lie to my supporters a few times about what I was doing and why I was doing it to get some to believe me. I think people who do things like that tend to be referred to as "politicians". Creating such a system would almost certainly lead to, eventually, a few fringe and maybe even extreme POV editors getting proxies from individuals who support their positions. Effectively, I could have everyone in my church, The Church of Delusional Psychos Who Claim to Be from Mars, do the minimum to establish their accounts, designate me as their proxy, and then leave. I might even be able to create several socks which would have enough separate edits to give me their proxies and disappear. I can't see any real way of ensuring it doesn't develop, in fact. And then there's the question of whether and how to void an existing proxy. Could I block someone who wanted to take their proxy away from me? Probably wouldn't be that hard to set up. I think the idea is way too flawed to be workable at this point. If we are going to have expanded governance, and the Foundation seems to approve of that to at least a degree, something based on the existing ArbCom model to some degree is probably the least troublesome way to go. John Carter (talk) 00:53, 9 May 2008 (UTC)

Ah. But now we're getting somewhere, because, what am I doing if I am electing someone to arb-com or pol-com if not giving them my proxy? And now, all of a sudden, why don't your doubts apply? At the minute the system we have works fairly well. It only doesn't if you perceive the need to make a decision today, something which fundamentally contradicts "consensus can change". Most of the conflict with the consensus model comes when people attempt to force an issue, when there is no real need for it to be forced. The consensus model is actually very robust and tends to work even in the face of external forces, for example look at the progress made with the fair use image issue. That was something which has been an issue since before I became an admin, and look how well we're managing it. I also have to question the board's position on wanting to implement "steering committees" on local Wikipedias. It seems to contradict their stated principles, and it also makes no sense. Either they want this to happen to reduce their responsibility, since they want it implemented to relieve them of some decision making, or it has no effect on them, and therefore they really have no business steering us in any direction at all. This all tends to smack of politicking just a little bit too much for me.

Look, since this seems to be coming from arb-com, all I can really ask is whether arb-com have looked at how arbitration works off wiki and discussed whether that model can be brought to bear here. Arbitration is supposed to end disputes by providing an outcome that both parties have agreed to be bound by. If that really is the case on Wikipedia, I'm not seeing it reflected in arbitration case settlements. That a case has gone to arbitration twice indicates arbitration isn't working, and I personally think it isn't working not because we don't have a committee which will decide upon policy issues, but because the arbitration committee isn't currently settling issues relating to behaviour. However, I'll concede that's a thorny issue. My best guess is simply to keep blocking users who won't work within the behavioural code we've established. Let them respawn, and let us block them until either they get tired or they reform. I can't really think of a better model than that. I also think arbitration cases currently are handled very badly indeed. It's very rare to see arbitrators actively engaging in cases, requesting information and weighing the issues. That feels off to me. Well, that's my current thinking, at any rate. Hiding T 09:46, 9 May 2008 (UTC)

Particularly in the last third or so of the 20th century, techniques were developed, for use in peer organizations, for finding broad consensus. Wikipedia uses rough consensus, and sometimes it is *very* rough. Sometimes it is similar to mob rule, where passions are inflamed, most people opine without investigation and thought, and some truly bad decisions get made. Among those are some block decisions. Blocking used to be a measure of last resort. In the comment above, there seems to be an imagination that the behavior of difficult users is improved by blocking them. It actually has three major effects: it frequently enrages them, and they become vandals, or they go away made, creating one more agent in the world of Wikipedia hatred, and the effort of maintaining vigilance against new accounts from these users is highly inefficient. As some socks have recently shown, blocks really don't work, if the user is determined. A very few editors respond to polite, brief blocks, blocks which set clear boundaries, and which are clearly protective of the community rather than punitive. Punishment as a model *never* works for gaining voluntary cooperation.

In the outside world, there are real (and large) peer communities that have had to face this problem, and they found solutions that work. But Wikipedia is strangely uninterested in what they found. Apparently, not invented here. Imagine an Alcoholics Anonymous meeting. They have drunks wandering in, sometimes. They have some very cantankerous people. And they operate by consensus. How do they manage it? Most meetings, most of the time, are quite peaceable. They value unity. They prize total freedom of speech. I've never heard of a person being banned from AA. *Individual meetings* may ban where there is actual fear of physical harm, or disruption is blatant and continued, and they tend to forget about these bans very quickly. What does AA do. How do they deal with disruption? I can tell you, it's not with punishment. And, for many AA members, having functional meetings is a matter of life or death, so it's not that they simply shrug it all off.

Wikipedia is gradually building up a reservoir of toxic waste. Blocked editors is only the tip of the iceberg. What's it like as a naive editor who puts hours into an article that is gone when the editor logs in the next month, and, from what I've seen, often without any notice at all on the editor's Talk page? I've talked with librarians and academics who have an extraordinarily poor view about Wikipedia, and it's not article quality, per se. It's that perhaps they wrote about something they knew well, and they came back later and it was gone or it was radically distorted. I've seen certain articles devolving, under simultaneous pressure from POV-pushers and RS fanatics. Please, I'm not criticizing the requirement for RS, but Wikipedia was built with hordes of articles that were, quite simply, just written from personal knowledge. *Later*, the articles were improved with sources. We are being extraordinarily inefficient with editor labor, and, eventually, the endless pool of new editors is going to start to dry up. --Abd (talk) 18:55, 9 May 2008 (UTC)

For what it's worth, I would disagree that elections to a body which can and should be transparent to anyone would be quite the same as giving someone a right to vote for you "until further notice". Part of the problem is, admittedly, the lack of involvement by the community in the decision making process in the first place. I can't think that would decrease if we gave those with little interest in being involved the ability to hand the ability to make our decisions to someone else so we won't have to deal with it. Perhaps your own points could be addressed if we were to have a behavior committee, either ArbCom or something else. I would myself wonder at who would be willing to take on the role of judge, jury, and executioner as such a body would be asked to perform, and probably not trust most of the people who wanted that role very much, but that's a side issue. And part of the problem as I see it with the "real world" application of arbitration is that, from what I remember, it isn't necessarily the last step there. It would still be possible, in several instances, to file criminal charges in several cases if the arbitrataion were unsuccessful. We do need some sort of behavior monitors apparently, and it seems clear us admins aren't doing all that is needed, particularly when the cases involved are rather complicated and take a lot of work to go through. One option I could see, which would probably reduce the number of admin candidates in a big way, might be to institute a jury selection system of some kind for more complicated cases. It might be something like a standard RfC with a set number of competent outsiders guaranteed to be included in addition to the regular participants. I would assume in most cases the jurors would be admins, although former admins and other editors in long standing might qualify as well. I just don't know whether we could necessarily count on such a system as working, as people won't be "obligated" to take part. Al least with a committee we get people who have committed themselves to taking part. Any of the rest of us might find some reason to not participate, often with very good reasons. If there were a proposal for a competent non-"committee" jury system, I might favor that as well, although I think the proliferation of differing opinions which would likely arise in such a system might present serious problems. John Carter (talk) 14:14, 9 May 2008 (UTC)

Could a proxy system be abused?

It's a pretty common objection to proxy systems that they could be abused. What has been missed here is context. If we assume that power is transferred with a proxy, then, yes, the potential for abuse is greater, as it is with any power concentration system. However, what is proposed is not the transfer of power, but merely some kind of representation in deliberation and the generation of advice. While some have written about an Assembly that would "legislate," that is a totally separate thing, and I'm opposed at this time to such.

John Carter wrote: I might be able to effectively lie to my supporters a few times about what I was doing and why I was doing it to get some to believe me. I think people who do things like that tend to be referred to as "politicians".

This assumes, in fact, that the holder of "40 proxies" is doing something with the proxies. That's not the proposal, in fact. *Somebody else* does something with the proxies. For example, ArbComm, say, wants policy advice. Suppose a significant number of editors have set up proxy assignment files in their user spaces. ArbComm could create a proxy table, adding to it transclusions of those proxy files. It could do this for every registered editor who has created a proxy file, or it could select in some way, excluding, say, new accounts or accounts that have not edited within some time frame. It's up to ArbComm, or to whomever is being advised, not to the proxy. The proxy simply participates. There isn't any power to abuse. The proxy speaks for himself or herself, and so votes. Like now. But someone who wants to understand and discriminate between those who are merely loud and determined and those who actually do represent a broader consensus could now have a means of doing so that moves a bit beyond what arguments sound personally cogent to them, and therefore must represent consensus.

No proposal is made here that votes should control. Rather, indeed, part of what is special about Wikipedia is that votes often don't control. But sometimes they do. There are simple methods for composing an Assembly from a collection of proxy assignments, that produces maximized representation while restricting the right to "speak on the floor" to a specific number. One such method was first proposed in 1886 by Lewis Carroll. Using proxy assignments to do this, there could be a standing assembly that *continuously* represents the membership. Without elections, staggered terms, and all the other complications. Just the individual choices editors make to name one proxy.

As to the imagined abusive proxy, what he's doing, in presenting evidence and arguments on the "floor" of the Assembly, is all visible, and anyone could point this out with a couple of diffs. If he is "lying" to his constituents (we call them "clients"), they will find out pretty quickly, I'd suggest. And he'd lose those proxies in a flash. And it is his arguments that count, most of all, the numbers aren't all that import, they are a detail. If an Assembly can't find consensus, its advice is going to be discounted, mere majority vote has little power in an advisory system. An Assembly, though, with restricted direct participation (direct *voting* can and should still be allowed), would presumably prepare reports and condensed conclusions, and voting would be on these, in the and (as well as on intermediate steps; in deliberative process there is a *lot* of voting so that process decisions can be made efficiently, none of it is really binding but merely procedural, until the final vote on a report or "motion"), so advice would be much more cogent and clear than what happens in, say, an MfD with 600 votes, with arguments all over the map, raising details that are never addressed, and an administrator who writes, at the end, "Result was Delete." --Abd (talk) 16:02, 9 May 2008 (UTC)

I would respectfully disagree with at least a few of the points above. One, a realistic view of wikipedia would make it clear to us that we already do have several editors who are actively involved in pushing fringe POV. These individuals would, reasonably, select a fringe POV pusher to push their own position. That person would be allowed to speak, and generally ignored I would hope, as it would be recognized that he were pushing a fringe POV. However, those who have "supported" that proxy would then be able to raise complaints that their valid, reasonable positions (of course they're valid, reasonable positions to their supporters) are being completely ignored and that wikipedia is a corrupt system which ignores the real voice of consensus, blah blah blah. This would be particular an issue regarding cases where even many of the proxies are less than well informed on the matter, which might make them less likely to say anything. Also, frankly, as Hiding indicated above, we really wouldn't need to formally set it up. If another editor were to say, "I've spoken to Abd and want him to speak for me here" that would be almost certainly acknowledged by the listeners, be it ArbCom, MedCom, any other proposed Coms, or whatever. But we also strive to avoid instruction creep and creating a formal system to do something which can just as easily be done without a formal system reeks of instruction creep. If Kim were to say they speak for Hiding, based on what Hiding said above, fine, I'd accept that as is. And, of course, as someone who's been invovled in politics, once any governing entity gets its foot in the down, they tend to get more powerful rather quickly, and the expected limited system gets replaced with a much more developed one rather shortly. Personally, I'd probably prefer a system which acknowledges out front that it will limit some freedoms, because it's generally easier to ensure limitations be built in up front if people are honest about their objectives. In this case, however, I honestly don't see why there would need to be a formal system to do what can be done just as easily, and possibly with less problems down the road, informally. John Carter (talk) 16:17, 9 May 2008 (UTC)

I'm afraid I don't see the problem. We have some POV pushers, or any fringe group. If few people are using the proxy system, those proxy assignments are going to be ignored, just as 40 fringe editors are now ignored, if, indeed, we've ever seen 40 fringe editors in one place (I haven't). But if more editors *are* using the system, which is the only way I can imagine that anyone would pay serious attention to it (it has other purposes, by the way, use for measuring consensus is only one of them), then, quite obviously, there will be a majority non-fringe, maybe even an overwhelming majority. If enough editors use it, it is actually going to show consensus, so the problem above would reduce to ... they were heavily out-voted by a means that actually showed broad consensus contrary to their position. Remember, this system, as being conceived here, is only being used to measure consensus. It doesn't create arguments (though, in fact, mechanisms might arise where representation expands; in particular, *one* attentive editor who understands the issues could, in a proxy expansion, demonstrate broad consensus against forty POV pushers or one proxy of forty POV-pushers. Quite simply, Mr. Carter has not considered the implications. Which is common. Now, if we were considering assigning *power* to proxies, we'd have to be much more careful. But if it's just a way of measuring who trusts whom, I fail to see the problem. So forty POV pushers trust one editor, who pushes the same POV? Where's the beef? Right now, that POV pushing editor can contact those people who trust him, off-wiki, and they assemble in an AfD or RfC. And what happens? Do we create any new problem by having proxy designations? From my point of view, we drastically lessen the traffic. We'd see maybe *two* editors comment! That's an improvement, over two *positions* expressed by forty-one editors.

Specific assignments can certainly be made already. And so can proxy assignments. Until someone is blocked for naming a proxy, it can be done. Nobody has been blocked for it, though Sarsaparilla (who had become Absidy by that time) was blocked for suggesting it. Okay, wrong. He was only warned for that, he was indef blocked for getting upset at the response and being uncivil to the warning administrator, nobody else. Um, Physchim62 or Tango, anyone?

But until there is a *system* for trust expression, even if only put together by the few who think the idea interesting, there is no way to use it to become more efficient and more fully representative, no way to run an Assembly election without bureaucracy and fixed terms and all that, no way to experiment with it. The "formal system" Mr. Carter seems to dislike isn't what is being proposed. Just voluntary proxy assignments by those who care to do so, with, then, the opportunity for the community to figure out how to use the information. Or not, as it, in its sovereign wisdom, it chooses. Frankly, I'm not quite sure why I keep arguing against the ill-considered objections and misunderstandings of the proposal. I guess because they are in front of me. I'm not starting the Assembly myself because it takes two. Until someone else says, "Let's try it," I'd rather not waste more of my very precious time. I'll answer questions, that's about it, and bring it up occasionally when it is the obvious solution to a problem. Such as the election of an Assembly, as was being considered here. --Abd (talk) 18:38, 9 May 2008 (UTC)

1. We already have a form of delegable proxy. It's passive--people engage in only the discussions they want to, and leave the others to the people who are interested in them. The problem comes when the only people interested are from a particular POV, and this will be solved by encouraging participation from a broader range of people. Sure, there's too much to participate in everything, but all that it takes is for everyone to sit in on one random Afds a week--there are about 1000, & if there's even 10,000 active editors that would get good representation.
2. AA is different because people participate face to face. It's our manner of participation by rapid-fire exchange of postings that brings out the worst in interpersonal communication for most people.
3. and that has a solution--absolutely insist on politeness, and remove the editors who do not have the skills to work here. even the good ones. Even the ones who have been here since the previous century. Even the admins. the first step is inverting BRD and establishing that making a bold that removes content without prior posting on the talk page is a blockable offense. Along with warnings for the use of any of a few dozen key word families like obsess? and vand?

DGG (talk) 03:55, 10 May 2008 (UTC)

In partial response, I wouldn't mind seeing some sort of statement that admins and admin candidates are requested (we can't force people, of course) to take part in at least one XfD per whatever, probably week. I wouldn't mind seeing admins "drafted" into RfC on an occasional basis, either. So that, for instance, when I summoned for comment duty, I am requested to respond to at least one RfC, requested move, XfD, or whatever per day logged on during that week, or something like that. I know people would hate it, but it would help ensure that faulty consensuses don't have as much of a chance of being reached. John Carter (talk) 22:11, 9 May 2008 (UTC)

What we have is not at all "delegable proxy," it isn't even "proxy." If I only participate in certain decisions, it does not mean that I trust those handling all the others, it doesn't mean that I'm not interested in them, it only means I have so much time and not more. And, yes, it could also mean that I don't have any interest. You don't know which from the fact that I'm not there. And I find it remarkable that, given a possible avenue to a solution that is designed to foster the development of consensus (that's what delegable proxy was proposed for, outside) is considered, apparently, inferior to an iron fist. Who wields this fist? Who monitors all this behavior? And, how, pray tell, does this scale to a point where Wikipedia is ten or a hundred times the present traffic level? BRD is relatively *efficient*, if people are paying attention. The distributed decision-making of Wikipedia is relatively efficient in certain ways. In others, it never was good, but the problem was masked by the constant supply of new editors. In some of my research, I've had occasion to study the editing patterns of some of our most active administrators. I find it hard to imagine how someone can sit there and push a button to accept an AWB edit, several times a minute, for eight hours in the dead of night. I'll say this. He didn't keep it up!

Yes, random assignment of editors to "committees" -- which is what an AfD page is -- would help with participation bias, and such solutions have been suggested outside for similar problems. This is what is done, roughly, in the U.S. jury system. It takes a bureaucracy, and it takes a kind of coercion.

Rather, let me suggest that what we have is *almost* right. Let those who have interest participate. But also encourage *groups* of editors who perhaps share some perspective consult with each other and delegate their participation -- informally, but in a way so that they know the task is covered -- to one of their number or someone else they trust. If votes don't count, that's it. But sometimes there is an AfD or the like, and 90% of editors !vote one way, with a handful !voting the other way. If the closing admin accepts the arguments of that small minority, as I've seen, there can be a huge flap. (Flap is part of how Wikipedia works: when a small scale decision encounters some strong objection, the situation is escalated.) The flap can waste a huge amount of time. If, from the start, it were clear, for example, that even a single dissenter who presented a controlling argument against what the closing admin thought were the spurious comments and "me too!"s of the majority, happened to be a proxy, directly or indirectly, for hundreds of editors, it would be much more obvious what was going on. A few editors with a fringe point of view had piled on. And to the contrary, if, say, nine editors whose views had been overridden by the single admin and the single editor on the other side *also* represented more users, then we would understand that we have a situation where, quite clearly, the administrator was disregarding an apparent community consensus.

But *none* of this should automatically reverse the decision. I'm *strongly* in favor of the Wikipedia model, where votes, in fact, don't count. What *does* count, however, is consensus, which is a deeper phenomenon than a particular set of votes. What the counter-consensus decision means is that, probably, wider attention is needed. And that's what DRV is about, and the Village Pump or RfC or ANI, and beyond that ArbComm and possibly a policy committee, or indeed, Jimbo or the Foundation Board. Which, quite wisely, ordinarily avoid inserting themselves in these decisions. What proxy expansion of votes would mean is simply that a better measurement, less subject to participation bias, has been developed, and so when the extent of consensus is relevant, it is very easily available. And it can be checked, by the way, if that is considered important. That's one application, but another was hinted at here, which is the voluntary coordination of editors to ensure that various tasks are covered by someone trusted. I'm quite sure it is already being done by some groups. Openly, wikiprojects may do this, sometimes. Delegable proxy sets up a visible, documentable network, through a very simple action by participants, that's the basic thing it does. The word "proxy" is a bit misleading, for, as we imagine the system, it's really bidirectional. Proxies are noise filters, filtering noise in both directions. And what that means I'll leave for another day.

--Abd (talk) 23:56, 9 May 2008 (UTC)

part of what you are describing just above is Wikiprojects, where groups of editors do get together and decide based on common interests. I think forming formal groups based on views, though, would simply harden blind block voting--and that of course is the basic objection to proxy systems of any sort--it destroys individual responsibility. Totalitarian systems are the ones that love block voting. Or one party pre-filled out ballots, like the 19th century US party system. And even with multiple parties, If we had a formal Deletionistist Party and a Inclusionist Party, we'd decrease flexibility where people decided on individual things according to individual merits. similarly on more specific issues-- Look for example at the SPOV advocates--they openly want to have a party to promulgate the Expert Point of View. I myself share a SPOV personally, but institutionalizing it is the road to destruction of NPOV. Remember the classic liberal argument: if you try to legislate against what you don't like, people can legislate against what you do like--Mill's argument for total freedom of speech (that said, he did want to limit the franchise to the educated; I see that as the residual effect of cultural bias on even him.) DGG (talk) 04:05, 10 May 2008 (UTC)