Misplaced Pages

Revision as of 23:41, 5 February 2011

Abuse

What happens if this is abused, as admin ship is on occasions?Slatersteven (talk) 14:14, 29 January 2011 (UTC)

First off, note that abuse is much easier to identify that it is with normal sysops, because of the restricted nature of what a sheriff can do. Sheriffs have a clear and limited mandate, and are held to a higher standard: If they can justify their acts as unbiased and necessary for the preservation of peace on the page they should be more or less immune, but if there's even a reasonable suspicion that they've gone awry then the community should take action. There are different gradations: a sheriff who makes an error in judgement should probably get a reminder about the restricted mandate, then be allowed to do a mea culpa and get back to work. A sheriff who seems not to understand the limited mandate and consistently behaves in un-sheriff-like ways should be removed and replaced, possibly with a bar on sheriffing again until they understand the rules better. A sheriff who grossly and intentionally violates the limited mandate (e.g. where someone volunteers to sheriff on a page specifically to use those powers to push through a POV), should be removed, barred from sheriffing, and probably desysoped for the violation of community trust. It's pretty much like cops in the real world: they should be given broad benefit of the doubt because they are doing an unpleasant job for the community's benefit, but woe unto them if they violate that community trust.

It's actually a fairly self-balancing system. Sheriffs can pretty much guarantee to ruffle feathers on a page, and editors with ruffled feathers will be watching and documenting every darned thing the sheriff does, looking for a way to get him/her in trouble. perceived abuse will be noticed and reported rapidly, and evaluation of such claims is straight-forward (does the reported incident lie within the sheriff's limited mandate or not?). Sheriffs will get quick feedback on their actions, will have an incentive to stick to the letter of their mandate religiously, and the clear specification of that mandate means that whole thing will be handled with far less wiki-drama than your normal ANI "this sysop is being mean to me" thread. --Ludwigs2 16:02, 29 January 2011 (UTC)

Im seem to recall that that has all be sadi about admin ship, yet admins regulay abuse their power with apparetn impunity. Such as the recent incident over Carol where admins clealry acted in a way that was (at best) partial. Unless it is will and not should then we will get the saem cabalisation gangs of Sherifs protecting their pets. Its not only the rules, but the enforcement that would have to be tougher.Slatersteven (talk) 14:02, 30 January 2011 (UTC)

Slater, pardon me for leading off with this, but you really need to (1) pay more attention to your typing, or (2) invest in a background spell checker. posts like above are hard to parse, and make it hard to take you seriously. I mean, nine spelling/grammar errors in one short paragraph??? My 12 year old niece wouldn't do that badly, ever...

The key difference here is that the rules for a sheriff are specific and restrictive. The problem with controlling normal sysops is that sysops have broad powers, more or less complete autonomy, and an ambiguous mandate. Anyone who wants to say that a sysop is doing something s/he shouldn't be doing has to get over the hurdle that no one has a clear idea of what a sysop should be doing, so such discussions boil down to wandering debates with huge loopholes and lots of wiggle room. That's not the case here, where the intent is to restrict the job to a limited area of action and take a jaundiced view of behavior that seems to step outside that area. It will be easy to tell when a sheriff is doing something s/he oughtn't, and quickly resolvable by removing the sheriff (who is only there because the community asked him/her to be there int he first place).

In short, I understand your concern, but the system is designed specifically to minimize that concern Rogue sheriffs will get the axe without fuss, muss or bother. --Ludwigs2 15:29, 30 January 2011 (UTC)

I would suggest then the addition of a new restriction. "that no Sheriff may be appointed to pages that they have been involved in acting in the role of an editor or as a party to dispute.” Lets make sure that the old gangs do not show up by having truly uninvolved sheriffs.Slatersteven (talk) 16:21, 30 January 2011 (UTC)

Well, that would probably be a natural part of the selection process, but it's easy enough to work in. I'll add it. --Ludwigs2 17:47, 30 January 2011 (UTC)

HAs it been added yet>Slatersteven (talk) 13:19, 4 February 2011 (UTC)

I see two things here: one, that a sheriff would need to apply the rules equally to people on both sides of a debate. I've seen it happen otherwise. Second, a sheriff would need to be someone who is inexpert not only in the subject area, but also in the kinds of debates taking place. There are too many people on WP who simply have an ax to grind concerning particular types of POV pushing, whether it be about race, politics, or other kinds of things. These admins have seen so much POV pushing that they are angry, and they will be unable to see that a POV pusher may have a point. They will subtly favor one group over another. I really don't like it that the admin corps is going to enforce this. If this is what we're going to do, then create a new usergroup in addition to admins. Here's what I propose:

Create a new usergroup called "sheriff" like this

$wgGroupPermissions = true;

That added to the WP Localsettings file would add a new usergroup and people could be given that right (by a bureaucrat), as "temporary admins" and they would be allowed only to block users who are editing on the page where they've volunteered to be a sheriff. (And yeah, it's that simple to implement in mediawiki). BE——Critical_{__Talk} 20:01, 3 February 2011 (UTC)

Another suggestion: what if the editors on a particular page had to agree to accept the sheriff? They could be told they have to have a sheriff, perhaps, but they could choose which one they wanted by full consensus. Full consensus probably wouldn't be too hard to achieve on such a matter, if the pool of sheriffs were large enough. BE——Critical_{__Talk} 01:36, 4 February 2011 (UTC)

I've added in most of your stuff, plus some other material. check it over. with respect to your last point, sure, if a group of editors decides they want a sheriff, they could ask (either over at AN, or maybe on a special Sheriff's noticeboard for things like that). That's kind of covered under the 'Various procedures for doing so may occur...' bit.--Ludwigs2 09:49, 4 February 2011 (UTC)

Two problems jump out at me. First, the name is too reminiscent of Citizendium's constables.

Secondly, a bigger concern is that admins do this kind of thing already. Why would a special group need to be formed for it? SlimVirgin 10:08, 4 February 2011 (UTC)

Also, the reverting back to a version of the sheriff's choice is highly problematic for all the obvious reasons. Admins have leeway to do this now, but it has to be used sparingly—there has to be something like a BLP violation, or some other clearly inappropriate thing. We can also choose to revert to a version before a 3RR violation. But making a purely editorial decision (this is the best version) is not on. And I think I saw somewhere on the page (though now I can't find it) that sheriffs should apply their restrictions to all parties equally? SlimVirgin 10:17, 4 February 2011 (UTC)

Why so many 'mays' A sheriff who breaks rule A may lose his powers. Sorry but if they have extraordinary powers then the sanctions need to be extraordinary too. I would say that any sheriff, who is found to have broken the rules, should not may, lose their status as a sheriff.Slatersteven (talk) 13:19, 4 February 2011 (UTC)

@Slim: I've never edited (or even looked) at Citizendium, so I don't know what constables do there, or why the comparison is problematic. That being said, the idea here is to create a class of 'Officers of the Peace' (aka cops), and so any appropriate name is going to be reminiscent of every other appropriate name. unless, that is, you want to reach for obscure referents. The thing I like about the TS name is that it retains that sense of lawlessness that Misplaced Pages already prizes - 'sheriff' feels more freewheeling and liberal while 'constable' sounds stuffy and establishment oriented (to my ear anyway). do you have a better suggestion?

with respect to reversion - this kind of thing isn't strictly necessary, but it's a useful tool for the toolbox. As noted, the goal here is to take away the advantages of bad behavior. One of the more frequent activities I see in page fights is editors struggling over which version gets displayed to the public during discussions, and this often has ramifications for the discussion - editors who manage to get their preferred version displayed will often drag their heels or try to squelch discussion, because they already have what they want. and that's not even considering tendentious edits and reverts in mainspace designed merely to infuriate other editors. Giving the sheriff power to dictate a particular version of the page for the purpose of the discussion pulls the rug right out from under that. If it helps, think of it as an ad hoc implementation of flagged revisions for troubled pages. Remember, the operant thought here is that sheriffs are going to be heavily monitored by page participants and sysops at AN, and will lose the sheriff status if they do anything controversial or one -ided, or at least anything that they cannot explain adequately on demand.

@Slatersteven: I'm still rewriting. however, I do want to leave the opportunity for sheriffs to explain an action and have them discussed rather than just get a bureaucratic axe on appearances. I expect some acts might look odd at first blush, but actually have decent reasons behind them which place them well within the Sheriff's scope. We wouldn't want good sheriffs de-sheriffed for doing their job well. --Ludwigs2 17:04, 4 February 2011 (UTC)

I am not sugesting that what I am saying is that if a 'cop' breaks the rules (and is found to have broken the ruels) there should be no may about them being sanctioned (and indead givent the extent of their powers I see no reason why they should not lose those for even minor infringements). I see that this will be the same as aminship with 'Cops' arguing the toss about the fact they did not breach the rules for days on end at RFC's. There is also the issue of consesnsus, If some one does indead become an advocacy 'cop' then he will have allies who will step into any complaint and assist.Slatersteven (talk) 17:18, 4 February 2011 (UTC)

Hi Ludwigs, constables on Citizendium are or were the admins, and the name attracted a fair bit of ridicule because of the cop aspect. We had a similar thing on Misplaced Pages years ago, which didn't last, and I'm writing now entirely from memory, but it was an investigations/mediation group that offered to troubleshoot at any article it was invited to look at, and it also used cop imagery, and was sunk almost entirely because of the name. So I think "sheriff" wouldn't work. SlimVirgin 17:19, 4 February 2011 (UTC)

So why not call them a Protector? Or (as a joke) a page lord.Slatersteven (talk) 17:22, 4 February 2011 (UTC)

lol - thank you for not suggesting "unterführer". The two keys to making this not a joke are establishing authority and legitimacy. achieve that and it will work, fail to, and it won't. To my mind the name is largely irrelevant, because what's ultimately going to make it fly is the broad understanding that this person (whatever s/he's called) can make it impossible to achieve editing goals on an article unless the goals are sought in civil, deliberative discussion. To use a not-great analogy, this is the same reason they have teachers in study hall: the teachers don't do anything, they just inhibit the kids from acting out on the normal hijinks that kids will get into unsupervised. Without the hijinks, the kids either sit in boredom or study; what else is there to do? The teacher needs to have the authority to send students to the principle, obviously, and needs to pay attention to stop sneaky hijinks, but once that authority is recognized as legitimate it almost never needs to be used because the rules get respected.

and you're right about all the 'mays'. I'll fix that now. --Ludwigs2 21:05, 4 February 2011 (UTC)

I changed the reverting back part per SV. These Sheriffs have greater powers than current admins at the pages they monitor. Their mandate is for much deeper control than I've ever seen an admin take under the current system. In accordance with this, they have to have stricter rules so they don't abuse their powers. But of course there should be consensus among neutral editors that they did abuse them. You can't have a Sheriff feeling too unsafe. It would be embarrassing to actually lose your status.

I have another suggestion, which I don't really like, but which might make this fly much more easily: give Sheriffs the ability to give page bans, and subject area bans, but not blocks. However, if an editor breaches a ban, a sysop should automatically block. That way, there won't be so much blather about it.

The Sheriff should have the power to make a call on consensus and then enforce it for a period of time- say a month or a week: once consensus is reached you don't want a couple of editors to come in and say "oops, no consensus anymore."

I'd be fine changing the name... how about "Crisis Mediator?" Sound psychological enough? BE——Critical_{__Talk} 03:22, 5 February 2011 (UTC)

And another suggestion: a Sheriff should try to get a bunch of other editors involved; that should be easier for a page under a Sheriff, because sometimes all the non-POV accounts have been driven off and would come back under a sheriff's protection. This would be necessary to getting real NPOV. Sometimes most of the current page editors are of one opinion, and they would call in a Sheriff thinking they will get their way. BE——Critical_{__Talk} 03:30, 5 February 2011 (UTC)

BC, you're starting to lose the edge on this. it's very important from a sociological perspective that sheriffs are self-contained authorities. If we don't grant them the power to impose short blocks, then they have to rely on sysops for their power, and they will (almost immediately) find themselves having to request help and having to explain before they act, and that will mean that the editors they are trying to moderate will be able to start doing run-arounds (wheedling sysops, organizing disputes and debates, playing politics, attacking the sheriff's character and legitimacy). If that happens you might as well not even have a sheriff. they will not be able to do the job effectively, and will just be another bureaucratic layer for people to squabble in.

I understand that this is very uncomfortable (and I know it will be uncomfortable for many people on project), but the only way this will work is if you give the sheriff complete self-contained authority over the page, and rely on the strict restrictions, close monitoring, and the ability to easily unseat him/her to keep his/her behavior in line. Imagine what the world would be like if police officers had to call their superiors each and every time they wanted to discharge their weapon: the ones who didn't get shot waiting for permission would not have one ounce of respect from anyone.

I'm less concerned about the reversion thing, but I'd still like to see it available as an option. again, the point here is to give the sheriff the tools s/he needs to quickly, quietly, and efficiently put a stop to nonsense. In disputes, people use mainspace to do all sorts of nonsense. Half the time I see edit wars start, they start because a group of editors is trying to bait another editor into getting blocked so that they don't have to discuss changes in talk. Now a sheriff could put a stop to that quickly enough with BoP blocks, but that will leave legitimacy issues of its own (much like the 'always locked on the wrong version' thing with edit protection). Maybe we could compromise - rather than give sheriffs revert power, give the sheriff a special template that says something like "This article/section is currently disputed; discussion is ongoing"; they can slap that onto the article and insist that it stay there until editors work things out. The point is not to let the end product of a nascent edit war be seen as a victory for either side.

In other news... If you really don't like the name, I suggest we go with the simple but factual "Moderator". That's really what they are, anyway, but I like the folksy feel of the word sheriff. Reminds me of The Andy Griffith Show.

with the respect to your last point - I don't think sheriffs should canvass, and I think they should steer entirely clear of content issues. The most a sheriff would do would be to recommend an RfC or noticeboard posting to participants without insisting on it. if you get a page that's dominated by editors of a single mind, then what they decide is what they decide; The sheriff's only job is to make sure that if editors of a different opinion show up, the conversation remains civil and productive, and doesn't devolve to dirty tricks.--Ludwigs2 04:06, 5 February 2011 (UTC)

Okay, so let's see if we can get them blocking power. I don't think Mediawiki gives the option for single page blocking, which would be ideal. I think moderation would be part of the job. Also, I think a moderator should have the ability to enforce consensus. IOW, make a judgment about what the consensus is, and not let one or two objecting editors mess it up for at least a while. Install a consensus version, and then continue discussion, then a week later maybe there's a change if someone still objects. It should be official also that a Sheriff can block/ban for continual IDIDNTHEARTHAT which has been explained in a thoroughly convincing way over and over. Which is about the same thing as saying enforce consensus for a period of time. As to reverting to an NPOV version of the page, that requires content decisions on the part of the moderator. But reverting to a page before the current dispute should be acceptable. BE——Critical_{__Talk} 04:47, 5 February 2011 (UTC)

If you want to change it to "Moderator," there's a draft in my sandbox you can use. BE——Critical_{__Talk} 05:18, 5 February 2011 (UTC)

Sorry if you’ve already addressed this point somewhere else Ludwig, but if you have I haven’t seen it. Your proposal for this policy says that editors can make a complaint if they think the sheriff is using his power irresponsibly, and that if this is indeed the case it’ll lead to the sheriff being replaced. What I’d like to know is, who will be making the ultimate decision about whether the sheriff has misused his power? Will it just be determined by community consensus?

If that’s the answer, I’m concerned that situation will lend itself to the same cycle of bias that already exists. If an article is being dominated by a large enough group of editors with the same viewpoint, they would be able to select a sheriff who shares this viewpoint, who could then help drive away everyone who opposes this group. Even though this would obviously be an abuse of the sheriff’s power, if the dominant group of editors support him and all of the editors who oppose him have been driven away, it would still be impossible to reach a consensus to replace him. If this proposal gets implemented, how would we prevent this sort of situation from happening? --Captain Occam (talk) 16:32, 5 February 2011 (UTC)

I am also concearned about this. I have raised doubts about this as well. The idea that Sheriffs will not be able to enforce on pagesd that are already active on is one answer to this. But enforcement by consensus would still be a potential issue.Slatersteven (talk) 16:38, 5 February 2011 (UTC)

Sheriffs definitely should need to meet the criteria to be considered “uninvolved” in the pages they patrol, but I don’t think that’ll be enough on its own. There have been plenty of examples of nominally uninvolved administrators who’ve still enforced policy non-neutrally. From what I’ve seen, this problem seems to have arisen especially often in the climate change topic area. --Captain Occam (talk) 16:49, 5 February 2011 (UTC)

I agree its not enough. I shall start a new sectio to discuse this point.Slatersteven (talk) 17:21, 5 February 2011 (UTC)

The point of this (and this is an idea that might take some getting used to) is to make the sheriff's responsibilities as non-interpretive as possible, and to impress upon sheriffs that they should interpret their abilities narrowly. that makes the system self-balancing. so, as an example: the sheriff has a right to redact personal attacks, but shouldn't disrupt the flow of conversation; yet some editor posts "If that idea weren't so stupid it would be laughable, but assuming you idiots really want it, we could try this..." Possible sheriff actions and results

• Sheriff does nothing - s/he's not fulfilling the mandate for the job. should be reminded of the job or even replaced if the pattern continues.
• Sheriff redacts the first two clauses, leaving the "we could try this..." part - fulfills the job mandate, can in no way be seen as exceeding it.
• Sheriff redacts the whole line - fulfills the job mandate, but exceeds it - should be reminded of the restrictions, asked to restore the non-attack part, and if the pattern continues should be removed or replaced.

What makes this work is the fact that some editor on the page is inevitably going to be (at minimum) miffed by any action that the sheriff takes which the editor thinks is unfair, and there will be an open thread somewhere (at AN, perhaps), where perceived unfairness can be reported. This creates a nice tension. On the one hand, miffed editors have an incentive to keep a watchful eye and complain about perceived unfairness, but must make clear and compelling arguments because the sheriff will be assumed to be acting in good faith by default. On the other hand, the sheriff (if he's not an idiot) will recognize that people he restricts are going to be pissed at him and looking for ways to get him in trouble, and that's a wonderful incentive for the sheriff to be a pedantic stickler about remaining within the office's limits and doing things correctly by the book. Add that there's nothing a sheriff can do that's seriously irremediable (redactions should all be inline, and the worst a sheriff can deal out is a 24hr block). Let's take the absolute worst case scenario I can think of: a sheriff goes totally off the deep end, starts redacting every post an editor made and starts blocking him every 24 hours. well, sheriffs don't have the power to lock pages, so the editor will be complaining in his talk (and probably emailing sysops), no one can really argue that the sheriff's actions are within the job's restrictions (even if they want to), so within 2-3 days at the outside all that sheriff's actions are undone, the sheriff is off the page, desheriffed and probably having a very unpleasant time at ANI; problem solved.

The trick is keeping the complaint threads tightly focussed on the "Is this allowable?" question, not "should this have been done" or "was this right". any questions of the latter type should be directed to a thread about modifying the rules that sheriffs work by; so long as a particular sheriff stays within the rules and applies them fairly he should be more or less immune. if he steps outside them, he should be toast.

(Just as an aside, this is a standard move in governance everywhere: create a set of rules, make individals accountable for performing those rules correctly, but make any legitimacy questions focus on the rules themselves. cops do the job the legislature tells them they should do; people can't really argue with cops for doing their job, but should get the legislature to change laws they don't like.) --Ludwigs2 18:12, 5 February 2011 (UTC)

Okay. I do think this could work, but you haven’t really answered my question. In the hypothetical situation you’ve described, where a sheriff goes off the deep end, who will actually be making the decision about whether or not they’ve abused their power? Even if what is and isn’t an abuse of their power is defined in completely objective terms, it’s still going to have to be someone’s job to examine complaints about this, and determine whether or not the complaints have any merit.

From what you’ve said here, it sounds like this would be determined by discussion at AN/I, and the decision to replace the sheriff would be made by ordinary sysops in response to complaints there. Well, if the sheriff is a popular editor whose actions are supported by many other editors, even if those actions are clear policy violations, I’m a lot less confident than you that an AN/I complaint would be able to resolve the issue. You probably remember the way this went when we were dealing with Mathsci’s personal attacks last spring. Some of the comments in question were completely obvious and clear-cut violations of WP:NPA, and ArbCom eventually agreed with us about that, but every time this issue was brought up at AN/I most of the people there ended up supporting Mathsci regardless, and it was impossible to find a sysop who was willing to do anything about how he was acting. How can we be confident that this same situation won’t arise when someone is trying to report a popular sheriff who’s abused his power? --Captain Occam (talk) 23:41, 5 February 2011 (UTC)

Broader questions

Isn't there a fundamental contradiction in empowering a "commander-in-chief" for the purposes of policing the "consensus process"? Consensus decision making is the means elected to writing the encycopedia, and not the overall objective of the project itself. And in my mind, the dispute resolution process in place now does a better job of incorporating "consensus decision making" in dealing with disruptive behaviors than this proposal does. While I agree DR is unwieldly, frustrating and uneven--but isn't this true of consensus building in the writing of articles as well? Why is consensus inadequate in this one context-editor behavior? Should we appoint sheriffs for NPOV, original research, fringe, and sourcing disputes too? The most contentious disputes I've witnessed in WP took place in policy, RFC/U and AN pages. Do we need sheriffs for those too? And user talk pages? Why not cut to the chase and simply appoint "Town Sheriffs" to write the articles themselves?

All admins currently have most of these "powers" already, one significant exception being granted any sort of "dominion" for their decisions in particular articles. We currently have DR up the ying-yang. WP:3, Misplaced Pages:NPOVN, WP:WQA, WP:RFCC, WP:AN, WP:RFAR--why aren't they enough? Assuming for the sake of argument that the "town sheriff" will be a real human being, that is, a person who is no more and no less gifted with neutrality, intelligence, and good intentions in DR, what do we expect the office Town Sheriff will accomplish that ordinary adminship through DR doesn't? Professor marginalia (talk) 17:50, 4 February 2011 (UTC)

Prof Marginalia: First let me point out that I'm trying to strike a personal balance here. I'm relying on a whole lot of research and academic training in the construction of this, but I don't really want to come off as an authority, so I am intentionally underplaying things that I could (literally) write volumes about. I'm only saying that because I'm having a hard time striking that balance correctly, and so I apologize if I'm coming across as confusing in the following response.

The thing I recognize that generally is not immediately evident to most people is that deliberations (in the formal sense of discussions aimed at achieving collective goals) always operate on two distinct levels:

• the ostensive level: things that are actually being discussed and considered by the group.
• the structural level: actions and things involved with determining intra-group dynamics and interpersonal relations.

When wikipedia people talk about consensus they are almost always thinking solely about the first point - consensus as the process of discussing the material - and generally don't recognize that consensus is first and foremost a structural ideal. Equality among editors, civil interaction, use of reason, reference to real-world sources: these are conceptions about the structure of the discussion that apply independently of any particular content that might be discussed on a particular page. If you've ever participated in real-world mediation you can see this in action: professional mediators use a variety of tools designed to construct an artificial discussion structure so that people can escape whatever entrenched social dynamics are preventing proper communication. For example, any time you hear a mediator say something innocuous like "So, John, you've heard what Matilda has to say: what's your response to that?", the mediator is implicitly (for the short list) (i) putting John and Matilda on the same social level, (ii) enforcing turn taking, so that everyone gets heard, and (iii) legitimizing both John and Matilda's opinions, and their right to express an opinion in that context. Freeing up the conversation structurally like that is a necessary first step in real-world mediation; it's what allows further substantive discussion to happen.

Misplaced Pages has almost no structural controls. In fact, wikipedia follows an 'ignore until punishment-time' strategy. We basically allow editors to do and say anything they want until they "go overboard", and then start applying cumulative warnings and sanctions; worse, the "gone overboard" criteria varies dramatically according to subjective sociological factors (time-on-project, popularity, impressions of factual correctness, etc.). This creates a conversational structure that (despite its superficial appearance of broad freedom) is actually anti-liberal and anti-egalitarian: it encourages demagoguery, cliquishness, Machiavellian intrigues, and other rhetorical gambits as editors unconsciously try to create a conversational structure that gives them advantages in the ostensive deliberation. You see this in real-world mediation as well: If it's corporate mediation, for instance, managers will often expect to be seated at the head of the table (an empowered position), and will get miffed if they are asked to sit at the side of the table opposite workers; even then, managers will often try to adjust the agenda for the mediation to focus on their concerns and avoid worker's concerns.

This whole thing is designed to create a particular and favorable structure for conversations without actually affecting the ostensive deliberation itself. Real-world deliberative bodies almost invariably use some variation of Robert's Rules of Order to accomplish the same task, but RRO isn't suited for the kind of free-ranging discussion that happen on wikipedia - that would require an authoritatively-structured system that would not work (or be appropriate) on project. instead, wikipedia needs a retroactive remedy - let people say what they want, but quickly rein them in when they start to go off in bad structural directions, and do it before it has the chance to amplify.

There's no more contradiction in 'enforcing consensus processes' than there is in a liberal society choosing to establish laws for itself. Anarchic democracy is a beautiful system if you have a community of people with the wisdom and detachment of buddhist monks. However, I don't think anyone would accuse wikipedia of that, and for a less enlightened community, structure is a decided blessing.

The reason why a separate 'sheriff' class is useful is four-fold:

• legitimacy. Sysops have all the power they need, yes, but it's too easy to delegitimize their actions. The few times I've seen sysops try to intervene on pages in ways consistent with what I've been talking about here, they have been roundly assaulted and abused for overstepping their bounds (or worse crimes against wikipedia). By setting up a separate class, we can imbue the class with the necessary powers, and place necessary restrictions on the use of those powers, and remove all legitimacy objections at the start.
• limitations. sysops have a lot of concerns aside from the flow of discussion, including more serious sanctions and policy considerations. when sysops enter page conflicts they are usually inclined to start evaluating things more broadly than a sheriff would be allowed, and that (again) starts to impact on their legitimacy. Sheriff's have one general task: they do it well or they get removed. no conflicts.
• accountability. again, sheriffs have one clear, general task. if they perform it badly they lose their sheriff status and no one has to worry about them again. Contrast that with sysops, and the painful process of evaluating and desysopping even the worst admin, and you'll see the accountability advantage.
• division of labor. last, and possibly most important, shuffling off these kinds of non-project-threatening behavioral issues onto a class of sheriffs frees up sysops tremendously. They can apply their efforts to other, more important tasks; they can enter into discussions about more serious sanctions against editors without being burdened by a history of minor conflicts. In fact, I fully expect that many sysops will become sheriffs as well, because then they can choose between acting as a sheriff (which wold give them community legitimacy to protect them from certain kinds of recriminations) or acting as a sysop (which gives them a greater range of powers), so long as they were careful to keep the two separate.

Long post, sorry, but I wanted to get things across as clearly as possible.

P.s.per NPOV, fringe, rs, etc.: sheriffs would not engage in content discussions - they are structure-only. A sheriff might be appointed to a dispute over (say) some issue of NPOV, but he would have nothing to say or do about the content issue, he would just keep he conversation within the bounds of behavioral guidelines. --Ludwigs2 20:39, 4 February 2011 (UTC)

This in-depth explanation is useful, thanks. I'm pessimistic that this sheriff would be spared the "painful recriminations" like those admins are subjected to, but leaving that aside...

Some of the most brutal "consensus building" I've seen on the project has taken place outside the confines of the articles themselves--in RFC/U and AN pages, the date linking dispute, FA reviews, policy and guidelines proposals, AFDs, and spread over an array of articles within a given topic area. Given the fact that in highly controversial topics, coming to a consensus over what's encyclopedically valid content within the article relies so heavily on evaluating claims against sources (and far less so, in my mind, inviting one-and-all to spout their own opinions, which are already overabundant as a rule)--how can the sheriff facilitate this necessary task without judging content? I ask this because in my experience editing controversial topics, it almost always comes down to this. Wouldn't the content neutral Town Sheriff idea be more realistic in smoothing consensus building when the legitimacy of sourced content isn't at issue?

Models in place now that are often imposed once the "community's patience has been exhausted" include arbitration imposed article probation and discretionary sanctions, which tend to eliminate most "legitimacy" conflicts over admin blocks, etc. And true, or not true? This is often, but not always, successful in eliminating disruptions in the article(s)--even while certain editors may continue to push the boundaries, escalate through other means, and subject themselves to tighter restrictions. Editors who receive these "discretionary sanctions" frequently moan and groan more than ever, and campaign about how corrupt, arbitrary or unfair the process is, attempting to de-legitimize the authority of the arbitration committee and administrators. How would this be any different under the authority of a Town Sheriff? Wouldn't the discipline they try to impose be even easier to undermine? Professor marginalia (talk) 18:51, 5 February 2011 (UTC)

First, no, they shouldn't have any trouble maintaining their legitimacy if we make the rules tight enough. Then it would be a cut-and-dried case of whether or not they've broken the rules. A normal administrator has a lot of leeway, and thus more ways they can be attacked. As to whether the sheriff would need to rule on content: there is, contrary to Ludwigs' desire, a very tight body of rules regarding content. If strictly followed, those rules practically dictate content. The Sheriff can comment on procedure and policy. What you're asking here is what happens when a couple of editors are going against consensus, based on their arguments and policy. What does the Sheriff do if they're right and yet going against consensus? That's a good question, and I don't think we've answered it yet. The following is a possible way, but is it too much like determining content? I would love for you to review the argument on the Jerusalem talk page, because I think it's a perfect example: fully RS sources are ambiguous on the matter of whether Jerusalem is the capital of Israel. POV pushers are trying to keep the statement that "Jerusalem is the capital of Israel" in the article lead. It's a totally obvious case of POV pushing and breaking the policy. What would I do if I were a Sheriff there? First, I think I'd be able to say that there are RS sources on both sides. Everyone there agrees on that. I could do a vote/hand count to determine if that's consensus. Given that consensus, I could assume that if WP takes a stand on the matter, it's against the part of NPOV that says we describe the controversy. I could then request a refutation of the proposition that stating in the lead that "Jerusalem is the capital of Israel" is not taking sides on the matter, that it is, in fact, merely "describing the controversy." Having asked that leading question, I could determine whether the editors who are practicing IDIDNTHEARTHAT on the matter are being disruptive. I could then warn/sanction them, and make a determination that in fact we need to change the article to the new proposed lead or something like it. But still... what if you have 10 POV-push editors who want it one way, and 1 who is right but going against consensus? Do we just enforce CIV and let the editor who is actually upholding the rules be driven from the article? We have, perhaps, the same problem as usual, in that we can't enforce NPOV and consensus doesn't always support NPOV. Ludwigs? BE——Critical_{__Talk} 19:30, 5 February 2011 (UTC)

Also, Arbitration enforcement has the same problems as usual for administrators, in that it's easy to attack someone who is relying on their own judgment too much. BE——Critical_{__Talk} 19:32, 5 February 2011 (UTC)

I largely agree with you that policies practically "dictate content" to a large extent, at least according to how I interpret those policies. But in my experience, disputes in articles are almost always entangled with disputes over whether or not rules have been broken. The broadly agreed "cut-and-dried" situations rarely lead to extended disruptions in articles because the existing systems work well managing them. But "cut-and-dried" situations often go unrecognized because (too many times) editors chime in who haven't read the sources or acquainted themselves with the topic enough to know which sources are good and which ones are crap. (That's a problem with RFCs in my opinion-it tends to stimulate "just some guy's opinion" from flyby's giving no attention to sourced opinion on the issue.)

To be effective, admins, arbs and the Town Sheriff all have to constrain themselves to rules but there is no alternative, they also must be empowered to "use their own judgment". Professor marginalia (talk) 20:20, 5 February 2011 (UTC)

(e/c) The fundamental difference between sheriffing and arbitration (and really the key to this concept) is the way that power and authority is localized in a single individual whose behavior is limited by established rules and fully accountable. Editors dissatisfied with the sheriff's behavior should be directed to start a discussion about changing the rules; the sheriff's behavior should only come into question where the sheriff breaks or stretches the rules. that's what saves the sheriff from undue recriminations (he's just doing the job, as written and authorized).

Now compare the two: in the arbitration model, there have been weeks of ugliness at the article page, and further weeks of ugliness at noticeboards, ANI, or other administrative venues before it even gets to arbitration. and then there are further weeks of ugliness as various people lay out their cases about how evil and horrible other editors have been. The arbitration result is usually restricted to sanctions on people: topic bans, admonishments, instructions to sysops that they can block certain editors for certain behaviors, etc. By contrast, a sheriff, steps into the talk page dispute directly and puts a stop to the nastier elements of the wrangling right there: he doesn't really have the power to sanction people themselves except in minor ways, but stops them from getting away with the kind of ugliness that would eventually lead to an arbitration case. In both approaches we get the same basic result - editors being more cautious about how they present themselves on the page - except the sheriff gives the results months earlier, in a way that doesn't actually excluding anyone from the conversation, and without all of the intermediary screaming, moaning, and ugliness.

With respect to the 'necessary task' of evaluating content... this system relies on the dual-level observation I made above. The sheriff's job is essentially to enforce the liberal, egalitarian consensus structure that most Wikipedians would expect already exists (fair, reasonable discussion about content without nastiness, chicanery, or melodrama). Once the overt nastiness, chicanery, and melodrama are removed by the sheriff, the discussion in talk will be relatively calm and reasonable and it will become much easier for participants on the page to evaluate content discussion points on their merits. This may let the problem resolve itself, but more likely the varying sides will express their viewpoints in the most reasonable terms they can, and the conversation will get stuck there. This is actually a good state to be in, because it will clearly delineate the content differences without all of the superfluous interpersonal problems. At that point, the sheriff might do a number of things: make a suggestion for a compromise, suggest wording for a 3O or an RfC, suggest that mediation might be helpful, suggest that a particular policy or guideline is the solution to the problem and point the participants to a noticeboard... The sheriff doesn't have to judge content himself; he just has to remove all of the obstacles to reason and common sense that will naturally and inevitably crop up.

(Just for an example, I've recently been involved in a discussion where one editor mostly relies on the the argument that a particular sourced viewpoint is being suppressed by advocate editors - he uses that line every time someone questions the reliability of his sources. A sheriff would simple redact that claim as uncivil wherever it occurred (without further judgment), implicitly forcing that editor to make an argument based on the merits of his sources rather than on the behavior of other editors. it would have saved the talk page a couple of thousand lines of cross-chatter and focused the issue right where it really needs to be.) --Ludwigs2 20:29, 5 February 2011 (UTC)

(after edit conflict comment. to prof M) yes, about using their own judgment. One of the most important qualities of authority is that the authority is trusted to act knowledgeably, responsibly, and independently within the boundaries of its mandate. The sheriff should be allowed to interpret the rules in terms of their spirit and intent and not be bound by their literal phrasing. There will obviously be an ongoing process of sheriffs taking action and the community retrospectively considering those actions and tweaking the rules to limit or expand the sheriff's power, but w should trust both that they will do their job well and that it will become evident and clear when they don't. --Ludwigs2 20:42, 5 February 2011 (UTC)

Instatement process

I think it should be the same as for full Sysops, though it should be easier to get in considering that you get desysopped if you abuse your powers. Regular sysops can't be desysopped (really) so it's a big deal, but this is different. But if you just let people volunteer with only the vetting of "good standing," well, I can think of a number of editors who would get in and have no business being a sheriff or moderating skills. BE——Critical_{__Talk} 04:51, 5 February 2011 (UTC)

I was actually thinking something milder - rather than requiring full community approval, the editor has to pass certain basic requirements, after which there's a waiting period of a week or so for others to register objections. basically it would be a "approved unless good cause to reject is shown" rather than "support needed to approve". I'd rather it didn't turn into a popularity contest, which would tend to stack the sheriff's box with editors who can gather lots of supporters. but more tomorrow - it's late here. --Ludwigs2 06:57, 5 February 2011 (UTC)

Hmmm, well from what I gather, you've been in a lot of conflict areas. If you think that the editors who participate in what you call "gangs" would make good Sheriffs or would not be approved, then I would say you probably know what you are talking about. So like you say it's in the details. BE——Critical_{__Talk} 07:01, 5 February 2011 (UTC)

SJS (Special Judicial Squad)

Who will judge the Judges? It has been pointed out that ther are already probloms wiht Sysop with cabals operating to protect (lets call them) tame admins. How are we going to determine if a judge has not upheld the law?Slatersteven (talk) 17:23, 5 February 2011 (UTC)

"There have been plenty of examples of nominally uninvolved administrators who’ve still enforced policy non-neutrally" I'm also concerned about this, and I've seen it too. We need to take as many precautions as possible without vitiating the role. One thing we need is to explain exactly when a ban or block is necessary, seeing that the Sheriff would be redacting other's posts. So when do they actually block/ban? And, one thing we could do is not have a judicial squad, but rather have judgments from editors who are not active, now or in the past, in the subject area broadly interpreted. How about that? But the main thing is as Ludwigs said: make sure that the rules are strict and clearly stated. If done properly we should be able to make it hard to subvert. BE——Critical_{__Talk} 18:47, 5 February 2011 (UTC)

Banning and blocking

Will there be a right of appeal, whi will hear this appeal and will a sheriff be able to block user talk pages as well (thereby stopiig any potential appeals)? I( read it that they can only apply sanctions on the actual articels pages is that in fact the case?Slatersteven (talk) 21:07, 5 February 2011 (UTC)

blocks (and I assume bans) will be a maximum of 24 hours, and non-escalating - I see no reason for an appeal process. I someone gets blocked for grossly unfair reasons, the sheriff will be in trouble for it and the block will get undone if it the problem is realized before the day expires (obviously). but if it's a reasonable block within the sheriff's purview - well, 24 hours is not long to wait in the grand scheme of things, and it will be a very good incentive not to repeat whatever it was earned the block in the first place (those of you who have young kids will recognize the 'time-out' strategy of behavior modification).

It will also do wonders to help editors realize that the "hurryhurryrushrushneeditrightNOW!" mindset is not the best way to approach Misplaced Pages. That alone will make the editing process so much more pleasant. --Ludwigs2 22:49, 5 February 2011 (UTC)

Blocks and bans

When does a Sheriff block or ban instead of redacting? BE——Critical_{__Talk} 23:15, 5 February 2011 (UTC)

Name

Any other potential names instead of "Sheriff?" SV for one had a problem with that name. BE——Critical_{__Talk} 23:15, 5 February 2011 (UTC)