User talk:Robert McClenon/Crisis

Untitled

Robert, I only read part of what you wrote on the main page here, but I agree that things could be improved and that ideally mediation would come before user conduct RFCs.

Maybe we should start Wikipedia:Dispute resolution reform? Maurreen (talk) 00:50, 28 August 2005 (UTC)

Caseworkers, etc.

Here are a couple ideas, food for thought:

1. Enlarge or otherwise change ArbComm so it can hear more cases more quickly.
2. Invent a new designation -- maybe "caseworker". These people would be empowered and possibly have the duty to resolve any disagreements as they see fit. Think of them as combining the functions of a mediator and enforcer. They would need to be approved by at least 90 percent of those voting. Maurreen (talk) 01:11, 28 August 2005 (UTC)

I agree that the ArbCom needs to be enlarged and have smaller panels.

Before I can support the idea of another designation of functionaries, I would like an explanation of exactly what they would and would not do. How would they resolve disagreements beyond mediation? If they would resolve disagreements by locking the article, that solution already exists, and is undesirable. If they would resolve disputes by blocking any revert that was inconsistent with what had been mediated, then I am concerned about the possibility of abuse. As I mentioned later in the paper, there need to be better limits on the use by administrators of special powers. The vast majority of admins use their powers wisely and cautiously, but a small number of disputes, such as an admin blocking an admin, result in considerable tension. If we were to create a tier of "caseworkers", we would need to decide how to remove the occasional rogue caseworker. Also, would caseworkers resolve content disputes by finding consensus and presenting alternative views as POV, or by making unilateral decisions? I assume it would be the former, but what would be the recourse if an occasional caseworker declared that there was a consensus when 45% of the editors said that there was not? Robert McClenon 13:56, 28 August 2005 (UTC)

Good points, Robert. I realize the "caseworker" idea is fairly radical, and I wouldn't be surprised if it gains little support.

Maybe it would be more useful for me to back up and explain what it's intended to address.

Most WPians agree that WP is intended to be governed by consensus. But sometimes that consensus is difficult to achieve or determine or interpret. And of course, sometimes people ignore consensus or essentially dismiss disagreement.

Our dispute resolution process in intended to help with the above. But for the most part, no parties uninvolved with a disagreement have any duty or individual responsibility to help with any disagreement. So sometimes things more or less fester.

People might list things on RFC and hope for a definitive answer to emerge, but not help on any other RFC. For instance, I've listed things on RFC and got little or no response. (But whenever I do list something, I try to help at another disagreement.)

If there is a consensus, and one or more parties are acting against it, the options are fairly limited, unless someone is breaking 3RR.

If there is a consensus, and one or two parties are acting against it, then Maurreen is right that the options are limited, but they are sufficient. If there is a consensus, then the community can revert to the consensus version. Either the one or two non-consensus editors will violate 3RR, or their edits will be reverted. Also, if the consensus and the minority are both reasonable, then perhaps the minority POV can be presented as a minority POV.

Admin powers only come into play when people are making significat policy violations. "The community" is THE higher authority, but it's often too slow and open to varying interpretations.

Ideally caseworkers would operate like this, in a nutshell: If there is no consensus, they would mediate or otherwise work to achieve one. If there is a consensus, they would enforce it. And I agree that there should be mechanism for removal. Hope that helps. Maurreen (talk) 14:22, 28 August 2005 (UTC)

What does Maurreen mean by enforcing a consensus? Robert McClenon 18:33, 1 September 2005 (UTC)

I think that the "caseworker" concept is worth further consideration, but I also have serious doubts about whether there might be abuses, so that the procedures would need to be very well-defined. Robert McClenon 18:33, 1 September 2005 (UTC)

1. Part of what I mean about enforcing consensus, and part of the problem in general, is that sometimes disagreements boil down to a question of who is going to be more stubborn, which party is willing to revert more.
2. Maybe it would be better to suspend the caseworker idea and consider possible improvements to arbitration. (To be fair, arbitration is probably a largely thankless task.) Maurreen (talk) 22:10, 3 September 2005 (UTC)

random thoughts

The problem is that the the dispute resolution system appears to be designed by people who have no training in actually resolving disputes, mediation, and what not.

Simple case in point, the "user RFC" is not actually a request for comments, but rather a form that any editor can fill out to 'make charges', build a case, and get a bunch of friends to sign on to support it. alternatively, a bad editor with an RFC can get a bunch of editors to come to his defense.

A user RFC should be renamed something like "incident report" and it should have some objective measures that limit waht can be put into the same report. it should be limited to an "incident", not some vague, ill defined problem that effectively allows a snowball effect, with many editors with separate grudges to snowball the incident report.

On the one hand, I think that FuelWagon has an interesting idea to introduce a new type of report that is an "incident report", which could concern 3RR violations or a series of closely related personal attacks. On the other hand, I disagree with renaming the user conduct RfC to be an incident report. The user conduct RfC is a penultimate step toward arbitration, as a way of punishing or getting rid of problematical users. There does need to be a way of demonstrating a long pattern of abuse, and a long pattern of abuse should be the basis for punishing a user. Robert McClenon 18:33, 1 September 2005 (UTC)

I find it odd that a document that you consider to be the penultimate step to arbitration and possible banishment be called something that is considered completely harmless in any other field. Are you familiar with Request for Comments used on the internet? Right now, article RFC's are like the RFC's that the rest of the world uses. The user RFC you describe with its penultimate step to arbitration is more like a grand jury, which decides if a case should go to trial. If you put these ideas in the "dispute resolution" spectrum, you get an "article RFC" at teh first step, mediation at level 2, a user RFC->arbitration at 3. If that's what you want, then I really think article RFC's and user RFC's should have different names. A user RFC should be named something that indicates this grand-jury nature that you're describing. FuelWagon 23:34, 1 September 2005 (UTC)

some admins may need to specialize. mediators shouldn't be able to block. "caseworkers" or whatever they're called, may not be able to block eitehr, but would have final control over an RFC.

A violation should be treated as a violation, regardless of who committed it. One of the things that I am sick of hearing is so-and-so is an editor with XX,XXX edits under their belt and so punishing them won't help wikipedia. Actually, if an editor violates wikipedia integrity and integrity is not restored in some way, then that makes wikipedia worse. If any editor causes a dispute with another editor and violates policy, letting that editor off the hook because of their edit count only furthers the accusations of a cabal at wikipedia, and tells new editors that they're not as important as long-time contributers.

admin priveledges should be no big deal, but when an admin breaks policy, it apparently becomes a big deal that they're an admin. Part of the root of this problem is that the 'arbitration committee' seems to be generally feared because they essentially have final authority and can hand out any punishment they wish.

I either do not understand, or do not agree. Abuses of administrative power should be a big deal, because they are a misuse of trust. Robert McClenon 18:33, 1 September 2005 (UTC)

Wikipedia policy needs to be designed so that Jimbo Wales can hand control over and have a sense that it will be able to run on its own. That doesn't mean he steps down, but it means that policy is designed so that it needs less and less subjective interpretation and is much more objective in what is and is not acceptable.

wikipedia needs finer-grained control. An RFC is sort of like "over the last few weeks, this user has been a real pain in the ass, and here's why". It's a subjective interpretation of policy violations like NPOV that require human interpretation.

I agree that there should be finer-grained standards in some areas. However, it appears that FW is suggesting more bureaucracy than is appropriate. Unfortunately, sometimes a user really is a pain for several months and really does need to be told to chill for a month. Robert McClenon 18:33, 1 September 2005 (UTC)

one thing that would help would be to have a link beside any "edit history" that says "report this edit". If an editor violates NPA in a single edit, that single edit should be dealt with. It should require an RFC and evidence and everything else to deal with it. It should be reported, and an uninvolved admin should make the call.

Another change that would help would be to put objective ways to measure whether an admin is "invovled" or not, so as to determine whether they can enforce policy violations with any individual editor. FuelWagon 16:18, 29 August 2005 (UTC)

I agree about uninvolved administrators. There do need to be finer-grained controls there. Robert McClenon 18:33, 1 September 2005 (UTC)

Oh, and having admins who can (1) admit when they make a mistake, (2) apologize, (3) clean up any issues they create would also go a long way. FuelWagon 17:43, 29 August 2005 (UTC)

Terminology

FuelWagon is completely correct on the terminological disconnect. I am familiar with Internet Requests for Comments. They are also, oddly, misnamed in a different way, which is that an RfC on the Internet may be either a real Request for Comments, or a standard. In Wikipedia, an article Request for Comments really is a Request for Comments. On the other hand, a user conduct Request for Comments is something else. It is not an Incident Report. However, it probably should be called something like a Request for Corrective Action (RfCA). The corrective action can be for the subject user to change his behavior, or for the subject user to stop editing, or for mediation (that should have happened earlier), or for arbitration.

Sometimes misunderstandings about terminology can result in deep divisions. This appears to have happened between FuelWagon and SlimVirgin. FW, relying on common Internet terminology, filed an RfC. SV, relying on well-known Wikipedia context, treated it as a personal attack. They were (in words once used by Winston Churchill) divided by a common language. FW did not intend to be attacking SV. SV thought that FW was attacking her. That is to be understood because they were divided by a common language.

The terminology should be changed. Then can we move on beyond these two valuable bold editors insulting each other? Robert McClenon 01:04, 2 September 2005 (UTC)

uninvolved editors

[1] The policy proposal for allowing admins to enforce NPA with a 24 hour block reflects the root of teh problem. Some want to reserve teh right for subjective interpretation to each individual admin as to whether or not the admin is "involved" with the editor in question. I proposed an objective definition saying an admin can't have edited the page that contains the attack one week prior, and that after the admin imposes a warning or block, that same admin cannot warn or block the same editor for one week after. Howls of protest have erupted saying that bad editors will game the system. Any possibility of admin abuse is dismissed. The fact that an objectively defined system would make it harder for accusations of admin abuse to be harder to make falls on deaf ears. Poeple want vigilante justice. No sense of rule of law is honored. Admins are "trusted" and should therefore be granted whatever power is needed to enforce policy, included deciding whether policy was broken or not, no matter how subjective the definition may be. FuelWagon 22:27, 29 August 2005 (UTC)

I don't think that I understand, but I think that FW is just venting his frustration in that paragraph, and will ignore it. Robert McClenon 18:33, 1 September 2005 (UTC)

suggested solution: random

OK, so, if you're a registered editor, when you view an edit history, you'll see (curr) (last) and (report). If you click on the report button, it will report that diff as a policy violation. This will get put into a queue of all reported policy violations. Admins who wish to handle a reported violation will go to a page, click on a button and the wikipedia website will pick which reported violation they are to investigate. Ideally, this decision making would filter out reported violations that occured on pages that the admin regularly edits, or users that the admin is often involved with. Beyond that, the software could randomly pick a reported violation (or pick from the ones that have the largest number of editors reporting it) and assign it to that admin. The admin would then investigate and either administer a warning or a block or close the report as insufficient to warrant punishment. Gaming the system by admins would be impossible. Which means cries of "cabal" and "old boys network" would be silly. Basically, an admin is nothing more than an editor and maintenance person and can only act as a policy enforcer on issues assigned by this randomized tool. This would require some changes to the wiki software, but it could be done. FuelWagon 22:38, 29 August 2005 (UTC)

See [2] and [3]. FuelWagon 22:41, 29 August 2005 (UTC)

Neutral Policy Enforcement

I think it would be useful to attempt to define when an admin may and may not be considered "neutral" in enforcing a possible policy violation. The current definition is that the admin is not "editorially" involved with an editor. I think it would be useful if this was nailed down to more objective measures. And it could apply to all policy enforcement, not just personal attacks.

any admin may enforce any arbcom decision

an admin my block someone for a violation on a page if the admin hasn't made an edit to that page within the last week.

after an admin enforces policy with an editor, they may not enforce policy violations with that same editor for one week afterwards.

It's a thought. It basically requires an admin to cool off after blocking someone and to get some distance. It also means that multiple back-to-back violations will be enforced by multiple admins, which makes almost automatically gets the opinion of several different admins regarding the problem. FuelWagon 08:19, 31 August 2005 (UTC)

#2: Editing a page within a week of someone else doesn't imply in any way a dispute with them. Editing a page the same day as someone else doesn't. I don't see what's wrong (unseemly, maybe, but not actually wrong) with blocking someone you've been involved with on a clearcut case like 3RR.

#3: I can see the point of this, but look at, for instance, SlimVirgin's dealings with Gabrielsimon/Gavin the Chosen (here). ~~ N (t/c) 14:18, 31 August 2005 (UTC)

(1) if it was always "clear cut", it wouldn't be a problem. Determining if a policy violation occurred often requires some subjectivity. Determining if an admin is "editorially" involved with an editor requires some subjective interpretation. Two things that require subjective interpretation (policy violation, involvement) means two sources of possible problems. The idea is that you could at least come up with an objective definition for "involvement" that would remove possible problems.

(2) I skimmed through the RFA, Gavin violates 3RR, gets a 24 hour block, the block is dropped, and 24 hours later, he gets another 3RR block. You would need only 3 admins to keep this up indefinitely. Is there only one admin willing to block this editor?

So, here's the deal. Say you take the definition of "involved" and make it objective, with a one week before and one week after restriction. Say you write policies in objective terms, the way 3RR is pretty objective to determine if a violation occurred. If you do this, then you have reduced the risk for abuse by admins within the system, and you could actually give them more power for enforcement. The example that comes immediately to mind would be you could increase the duration of a block as the violations accumulate because those violations would require multiple admins to enforce. This would mean that the Gavin thing would require 3 or 4 admins for the first blocks, and then after that, policy enforcement could say that 3RR violations can be blocked for 48 hours if there are 3 violations in the last week, or 6 violations in the previous month. By designing the system with higher confidence that it will operate properly, you can give admins greater power in enforcing violations because you'll know it's harder for them to misuse. You could design an objective policy that would allow a few admins to deal with Gavin indefinitely without it ever going to arbcom. That seems like an advantage to me. FuelWagon 17:54, 31 August 2005 (UTC)

I agree with FW's concept of neutral policy enforcement. I don't think that an admin who has been actively editing an article should be the one who blocks an editor for 3RR violations. Most admins are scrupulous about not misusing their privileges, but allowing the same person to be a content editor and a policy enforcer could result in an admin taking a proprietary interest in an article. Robert McClenon 18:33, 1 September 2005 (UTC)

Also, I do not think that administrators should be the continuing line of defense against persistently problematical editors. Persistently problematical editors should be taken to the ArbCom. In order to permit this, and relieve the need for admins to shoot on sight, arbitration needs to be streamlined. (Perhaps the demands for vigilante justice arise because due process justice by the ArbCom is too slow and should be speeded up.) Robert McClenon 18:33, 1 September 2005 (UTC)

objective policies, enforced neutrally

One of the problems is "scale". The number of editors has exploded and the number of admins is around 500. wikipedia needs to be able to have possibly thousands of admins to deal with trouble. The current system of subjective policies and subjective definitions of neutrality would work when you have a small number of admins and Jimbo Wales being able to keep an eye on all of them directly. If you want to scale wikipedia, you need a system that can allow for two thousand admins to work reliably with little oversight. To achieve that, wikipedia policies should be defined as objectively as possible and they should be neutrally enforced. This would allow a large number of admins to be appointed and they could could be set loose on wikipedia. if policies are defined objectively, then admins will know what they can and cannot do, and it will be easy to determine if an editor broke policy and if an admin overstepped their authority. If neutrality is defined objectively, then it will be easy to determine if an admin can enforce policy on an editor based on past edits. The idea of "open policies, enforced neutrally" would be a sort of "meta-policy" for all otehr policies. And it should make it easier to scale the number of admins, quickly enforce policy, easily find rogue admins, and manage the whole process neutrally. And it has a cool acronym too (OPEN). FuelWagon 20:25, 1 September 2005 (UTC)

This would have the decided advantage that admins could handle individual violations quickly and easily. it would also mean that several admins could easily enforce multiple violations in a row from the same editor, which could then also be used as a way of flagging that editor as a repeat offender needing to go to arbitration, and possibly defining objective ways of increasing block lengths for repeated violations, because it would mean multiple admins have looked and found several different violations. The "this editor has been a pain over the last few months" should be easily documented by multiple policy violations that have been logged by multiple admins of the last few months. Arbcom could simply look at a block log history and issue a ruling. FuelWagon 20:30, 1 September 2005 (UTC)

Arbitration

Maybe it'd be good to do a little research about how long arbitration cases take. I'd be willing to do some if anyone else will also. Maurreen (talk) 22:15, 3 September 2005 (UTC)

Public Hearing

Robert, the name "Request for Corrective Action" could be interpreted to mean the RFCA is a request for arbcom to come in and "correct" something. I wouldn't want a name that some would interpret to mean this is an application/request to get to arbitration. I would prefer a name that clearly reflects (1) the idea that this step in the process is intended as a way to resolve something in and of itself and (2) is as serious in that it is one step from going to arbitration, the way a grand jury is one step to going to trial. I thought the name "public hearing" might work, since that's really what a user rfc is. Both sides make their cases public, the public (all editors) hears them, and the public (all editors) can give its opinion. It is still non-binding and not punitive and all that, but it is a step unto itself, it may resolve a problem without needed to go on to arbitration, or if the dispute is not resolved, then RFA would be the next step.

So I suggest the term "Public Hearing" as the new name for the step in the dispute resolution process that we currently call a "user RFC". FuelWagon 15:48, 5 September 2005 (UTC)

proposal

I have created a proposal here Wikipedia:RFCReform