User talk:Lawrence Cohen/Arbitration RFC draft

From Wikipedia, the free encyclopedia

Contents

[edit] Thoughts

Please feel free to leave any thoughts you have on this. Obviously, still a draft. Lawrence § t/e 17:03, 13 March 2008 (UTC)

Could you add a section on scale problems? I believe arbcom have had internal discussions on this in the sense of what happens as arbcom gets bigger, and as wikipedia gets bigger. In fact, it might be an idea to ask if arbitrators would be willing to tell us what they have discussed themselves in terms of arbcom reform, so that we avoid re-inventing the wheel. Also, the role of Jimmy Wales might be relevant. Finally, that new committee that arbcom set up to report back to them on nationalist editing (I think). That should be mentioned. Plus any other new stuff that I've forgotten. Hey, why not discuss the recent change for format on the arbitration pages (asking people to submit full proposed principles, findings of fact and remedies, rather than separate PP/FF/R). Um. The clerking system as well, though there doesn't seem much wrong with that. Maybe the request system as well. Oh, and the voting rules, some of which may seem arcane or need better explanation to some. But maybe this is going to far. Start at the top and work down, picking out the major issues. Carcharoth (talk) 17:29, 13 March 2008 (UTC)
Asking the AC directly is a good idea to weigh in on the draft, I'll post it on that page. I was tempted to get down and dirty into the minute details, but then the format alone would be massive. My idea for wider, sweeping topic sections would be to allow people to fit in those bits. I might think something in the AC is wonky, but another 100 users may not... Lawrence § t/e 17:36, 13 March 2008 (UTC)
This is the kind of thing people do when all else fails. I've been disappointed with several recent decisions also, yet the immediate cause for today's misery is something that's much easier to fix than getting an undertaking like this one off the ground. Namely, community sanctions need to be better defined and standardized.
A community ban is a block that no admin is willing to undo is a concept that went into policy in summer 2005--when this site had a tenth of its present registered accounts and just a few hundred admins. It isn't scalable. Particularly, that weakness in the banning policy dovetails with vagenuess in the definition of wheel war in a way that made today's debacle inevitable. This is a dynamic that crops up whenever we have a high profile controversial community banning proposal. The difference this time is that we're used to kicking the situation over to ArbCom, and now we're stuck because it's already been there.
As a community, we've been stuck in the dark ages about community sanctions. Wouldn't it be better in general to have some standardized way to intervene earlier with milder sanctions? Suppose we did 1-3 month topic bans before things reached this point? Although I haven't followed all the details on the early history of the Mantanmoreland case, imagine how things might have gone if a brief topic parole had been applied in first half of 2006?
Suggest you put this on hold for now and see what we can do about those areas. It's easier to revise a policy than to haul all of ArbCom into RFC. DurovaCharge! 17:36, 13 March 2008 (UTC)
This isn't really a specific reaction to the Mantanmoreland situation. You got railroaded quickly through your RFAR. The IRC RFAR was basically the RMS Titanic. The Waterboarding RFAR and the Mantanmoreland one both basically went out of their way to not address major concerns of the community. Out of the recent ones I've stuck my eyes on, only Bluemarine has worked out completely well. 1/5 is not so good, and that's just me. My other reasoning for this was that an open-ending RFC will show what the community expects of the AC, and the consensus(s) if any will be the mandated changes the AC must take on in it's service to us. It seemed like a simple elegant solution for any possible problems like that. If 1-2 months after the RFC opens someone goes to them, and says: "200-300 users here have instructed that the AC is doing this, this, and this this way from now on," thats it, and some problems fixed. That is the hope. Lawrence § t/e 17:42, 13 March 2008 (UTC)
Well, then put the 2-3 months and numbers of votes you hope to get, in the document. And make it a specific timescale and forbid people to moan about currently open cases. If this is going to be open for 2-3 months, it certainly does need to be rock-solid. Carcharoth (talk) 17:48, 13 March 2008 (UTC)
I would say 3 months, actually. Though I wouldn't be unhappy about 2 months. Carcharoth (talk) 17:49, 13 March 2008 (UTC)
I tagged it for 3. Lawrence § t/e 20:59, 13 March 2008 (UTC)
I see your point. If you haven't looked at it already, try Wikipedia:Requests for arbitration/Matthew Hoffman. DurovaCharge! 02:18, 14 March 2008 (UTC)

[edit] Who to ask for feedback

I see that the current list of what links here shows who has been asked about this. Any more? Noticebaords? Or do you just want individual feedback at this stage? (I came here from FT2's page). Carcharoth (talk) 17:31, 13 March 2008 (UTC)

Since it's still a draft, I was figuring on starting small, then working out by "word of mouth". This is the sort of RFC that needs to leave the gate rock-solid when I hit the move button and kick it over to Wikipedia space. Lawrence § t/e 17:34, 13 March 2008 (UTC)

[edit] Omnibus

I've seen a bunch of ways that ArbCom has failed. The most recent IRC case was a disaster... a spectacular case of private interests prevailing over the public good. However, when a person does an RFC like this, the effect will be to catch all sorts of voices. The unfrozen cavemen of the ethnic struggles will talk about how they hate it, the long time content people will talk about how they hate it, the people who think the only problem is people using bad words will talk about how they hate it, etc., and yet all of these people will be hating different, and sometimes diametrically opposed, things. Therefore, we can all agree that the mess is a mess, it's in the clean up that we fail.

I would point, however, to one thing above all others for its reason: it is not elected. Because it is selected, not elected, and selected by someone without a day-to-day experience of the project and the mechanics of the place, the selection basis gets highly suspect, even if it isn't necessarily corrupt. I.e. the persons.

A second thing is the lack of any appeal structure and the lack of accountability for bad decisions. These are both horrible given the size of the project and the necessary narrowness of any one user's point of view. Utgard Loki (talk) 17:55, 13 March 2008 (UTC)

In regards to selected vs. elected, that as anything on any local Wikimedia project can be changed by the local community so long as whatever is changed does not interfere with the core Foundational principles. The community can easily convert it to a straight election if enough support the change. Lawrence § t/e 17:59, 13 March 2008 (UTC)
I think the selected/elected issue is largely a semantic one; in the past couple elections the top vote-getters (by percentage) have gotten seats. If there's a problem with the committee being out of touch, then the community really only has itself to blame. Mackensen (talk) 20:45, 13 March 2008 (UTC)

[edit] What's this all about, then?

This seems really ambitious. It needs a lot of preface to make it very clear what is going on, and since it's a deviation from normal RFC process, it needs explanatory info on why following the deviation is appropriate and what the deviation is achieving. It's going to be hard to maintain an orderly process, and moreover, hard to justify. None of that is to say it's a bad idea, just that I'm dubious without a lot of preliminary work. I do wish it well. ++Lar: t/c 18:06, 13 March 2008 (UTC)

[edit] my statement

The statement I just made could conceivably be put under "Statements about what does not work well in the current Arbitration Committee process". I'll leave it up to Lawrence. daveh4h 18:18, 13 March 2008 (UTC)

I rv'd your statement and put a "RFC not open yet!" note on the page. Sorry! We're still trying to figure out the RFC format. Lawrence § t/e 18:28, 13 March 2008 (UTC)

[edit] Mackensen's ramblings

Broadly, this RfC will fail for the same reasons that it has been brought forth--our current processes don't scale and we lack mechanisms to manage "interests." I've thought for some time the community in its present form was at least ungoverned and perhaps ungovernable. In this RfC, as has been noted above already, you'll hear from every interest who believes ArbCom got it wrong at some point in the past--content editors, administrators, BLP folks, nationalists, deletionists, etc. Heaven knows Arbcom makes non-optimal decisions all the time, but speaking as a party to these deliberations it comes in no small part from an inability to grasp complex situations and determine the best outcome for the project. I'm not convinced any body of 12-15 men and women, no matter how gifted or how well chosen, could accomplish this in a satisfactory manner.

I think solution, sadly, lies in an expansion of the existing hierarchy beyond strictly technical rights. At the moment, the structure is, vaguely, Administrators-->Arbitration Committee, with the CheckUsers floating around on their own. Back in 2005 it was still possible to talk about administrator consensus, but it was a smaller body and shared a common outlook. I miss those days, but they aren't coming back.

A few months ago a couple of us kicked around the idea of delegating block review from ArbCom to a body of users, with final authority/final appeal still resting with ArbCom. What might also make sense is replacing the RfC system with something akin to the American appellate system: groups of users/administrators with powers to ban users from articles or to take other steps short of bans, with appeal to Arbcom. In other words, taking review out of the hands of administrators/the community en masse and investing it in an appointed or elected body. This frees the community to focus on content issues.

These are ramblings. While I think it's an idea worth considering I'm not convinced of its efficacy. I do know that Arbcom cases with talk pages in excess of 300KB symbolize failure. Mackensen (talk) 21:00, 13 March 2008 (UTC)

My 2 cents are that Arbcom is overworked and understaffed. If this RfC does go through, I'd support the idea of splitting up some of Arbs functions (like the fact that users with multiple non-violating accounts can register them with Arbcom!) and maybe even having Arb panels instead of a full body of 15 possible for each case, but that would be a discussion for the RfC. MBisanz talk 06:52, 14 March 2008 (UTC)
I was under the impression arbcom have had internal discussions about this and other issues. It would be nice if they could update us on this, and let us know if there has been progress on this. Carcharoth (talk) 12:12, 14 March 2008 (UTC)
The Committee is, generally speaking, lethargic on any matter having to do with abstract, long-term reform. I think the initiative in any significant reform of the hierarchy and the pseudo-constitutional system we have will need to come from the community as a whole if it's to go anywhere in practice.
(And I do agree with Mackensen that a more developed hierarchy is likely to be the best long-term solution. This goes beyond the failures of the Committee in its normal tasks, incidentally; there are other areas in which we play no role—policy formulation, content disputes, and so forth—in which the size and non-homogeneity of the community is beginning to make traditional approaches infeasible. In the long run, the community ought to consider how it wants such decisions to be made, and set up the necessary infrastructure to make that work.) Kirill 13:28, 14 March 2008 (UTC)
Wikipedia:Dealing with scale? Or can anyone come up with a better name. Do we have a name for long-term strategy pages? Essays? Carcharoth (talk) 16:47, 14 March 2008 (UTC)
I'd call it Wikipedia:Devolution, but that might be begging the question. Mackensen (talk) 23:00, 14 March 2008 (UTC)

[edit] being an arbcom member is status and not a responsibility

In my opinion, a majority of arbcom does not bother to read and understand all the evidence in every case and this has been true for as long as we have had an arbcom. Most are just admins that have leveled up. WAS 4.250 (talk) 11:31, 14 March 2008 (UTC)

Ouch. Whose fault is that? But this comment may be a bit premature. ++Lar: t/c 17:52, 14 March 2008 (UTC)
It's not just laziness--in many cases the combined size of the evidence and workshop pages are simply beyond the capacity of a volunteer to grasp. I stood down at the last election because I didn't have the time to do the job right. Moreover, there's three or four open cases at once, each demanding attention. Again, I think that's an argument for splitting up responsibility. There's no reason for the committee to hear all cases en banc; in practice, one or two people take the lead in a given case anyway. Mackensen (talk) 18:25, 14 March 2008 (UTC)
Note that in the last election I asked a leading question about splitting up to handle more cases, it was not very well received (for example Raul654's answer or NewYorkBrad's answer)... perhaps it's time to revisit that. ++Lar: t/c 22:28, 14 March 2008 (UTC)

On this very note, I had an idea that I mentioned to Brad that I need to try to write up tomorrow. It was for basically two types of AC cases. By default, all cases, period full stop, would a "Closed" format. Arbiters must announce the scope and things they will look at when accepting the case based on opening statements and linked early evidence. Slightly processy, but it would allow cases to be much sharper and focused. To add additional facets to the case, parties or whomever would propose requests to workshop, with why it should expand, to cover what, and with backing evidence. The AC would then decide in public there to expand or not on the valid points and questions. Each case would carry an "In a nutshell..." type header template so people know what questions the AC is considering. If an entire case hinges on, "Did Lawrence Cohen and Mackensen collaborate to frame Jimmy Wales for the Lindberg baby?" only evidence related to this is accepted. Everyone knows the proposed decision will revolve around this alone. But it can expand to a 2nd, 3rd, 4th, 5th question, if a demonstrated need exists. Save us a crapton of time. If the AC decides to accept the case as an "Open" case--the IRC cases and Mantanmoreland cases probably would have, given the scale--it's like now, and anything goes. A "Closed" case is a precise trial, an "Open" case is a Senate investigation and hearing. Open cases would be infrequent to rare. If the AC feels it needs to, it can easily convert a Closed case to Open partway through, if the need arises. Lawrence § t/e 18:46, 14 March 2008 (UTC)

[edit] RFC section survey

OK, I have these sections now in this order.

  • 1: Statements about what works well in the current Arbitration Committee process
Obvious.
  • 2: Statements about what does not work well in the current Arbitration Committee process
Obious. Set seperate from "what works well" section on purpose, to allow more specific statements on both sides.
  • 3: Clarifications on limits of what the Arbitration Committee can't do
This one has funny wording, but it's deliberate I think. The AC as a body I think can try to do anything--whether the community lets them, who knows? I see comments all the time though that the AC can't or doesn't do such and such, often conflicting slightly. Needs clearing up.
  • 4: The Arbitration Committee and scaling
Made this a dedicated section as well--this seems to be a major concern going forward.
  • 5: Arbitration Committee change and reform
Ideas about how to fix, change, or reform the AC and it's processes, aside from scaling matters.
  • 6: Views and statements about this RFC
Views and statements about the RFC itself.
  • 7: Discussion
Standard discussion section for RFCs, directing people to post here on talk.

I'm hesitant to break it up too much more. An RFC with a million top level sections will be a waste, but I think these are general and specific enough to start out. What do you all think? Lawrence § t/e 19:04, 14 March 2008 (UTC)

[edit] Broadening the scope, making this about governance in general

The following proposes a somewhat dramatic shift of focus. I like what Lawrence has started here and am fine with sticking with something along these lines, but I'm suggesting we take it even further.

As Lar says above this is quite ambitious, but if it's going to be ambitious already why not throw caution to the wind and make it extremely ambitious? We could turn this into a community-wide "Request for Comment:Wikipedia Governance." Obviously discussion of the Arbitration process would be a huge part of that, but there's much more we could be talking about. Comments above already suggest a broadened scope would be good. Mackensen, MBisanz, and Kirill suggest devolving responsibilities to whole new subsets of users (which presumably would be "new processes" of the kind we are to avoid discussing in the "Arbitration Committee change and reform" section of this page) and others have raised the issue of election vs. selection which goes to questions about democracy, voting, and Jimbo's role that involve more than simple ArbCom elections.

As Durova and Carcharoth say the overall problem here is one of scale. Wikipedia is just too big to run the way it was run three years ago, and we need to have a big, honkin', community-wide conversation about that and what we should do as a result. Said conversation should last for months and bring in as many people as possible. We could run the proposal by Jimbo and the ArbCom and, if they are game, get their imprimatur so more folks would be inclined to participate. We would maintain very strict civility rules (maybe using a few different admin "clerks" a la ArbCom) and insist that the discussion not drudge up any old grievances, but rather confine itself to analyzing what works now and why, what doesn't work now and why, and what we might do about it going forward (general note: I think we would want to confine the discussion to issues of governance, not of content - they are obviously related but ultimately distinct).

I would also suggest that if we engage in a process like this (and if we don't do it now I think/hope it's inevitable that we'll do it some day) that we rid ourselves of some of Wikipedia's traditional skepticism of "experts" and consult some folks who have ability to shed real light (not heat) on the issues we will be discussing. Obviously we all have our areas of expertise (I can tell you quite a bit about the history of hip-hop and of American foreign policy during the Cold War, just ask!) but few of us know much about how to effectively organize massive online communities devoted to an intellectual project. Actually probably no one is a true expert on that, but there are certainly some people I'd much rather hear from. Why not consult organizational psychologists, anthropologists, sociologists, political scientists, and folks in similar fields? I think we would especially want to hear from those who have studied large online communities, or even large communities in general. The best place to start finding such people would be among Wikipedians, as I'm sure we have a number of folks with expertise in these areas. We could have some of these editors provide us with suggested reading lists that would educate those of us who are not experts and, more importantly, ask them to prepare white papers that deal specifically with Wikipedia in its current state as it relates to questions of governance. Obviously all of this "expert feedback" would not determine what we decide to do - the community would do that - it would just help us to make highly informed choices. Otherwise I feel in a conversation like this a lot of us (myself included) would be shooting from the hip without knowing what in the hell we were talking about.

So that's my BIG IDEA. Maybe it would function better in a different venue, though personally I think simply broadening the scope here would work well. There's been a lot of anger about certain recent ArbCom decisions and I think something needs to be done to address that which is why it makes sense that Lawrence set up this RfC. It seems to me though that this is a good time to have a Wikipedia wide conversation about how this remarkable project will actually function in the years ahead - better sooner than later. Thoughts?--Bigtimepeace | talk | contribs 19:43, 14 March 2008 (UTC)

A complete, community-wide discussion about the future of Wikipedia? If it could be done without becoming a battleground, this could be one of the best ideas I've ever heard. Dr. eXtreme 19:12, 17 March 2008 (UTC)

[edit] Possibly relevant links

  • First ratification vote (talk)
    • Presumably on a revision dated 5 February or earlier. This one as edited by Jimbo Wales adding a link to the vote seems likely.
  • Second ratification vote, by the community
    • The vote is on adopting Wikipedia:Arbitration policy "as it stands in the edit of 14:15, 30 Mar 2004"
      • There is no edit so dated. My guess, based on timezone issues, is that it is actually this edit, dated 13:15 UTC of that day, that was understood to be the subject of the vote.
    • The vote was intended to run from 2 April 2004 to 9 April 2004.
    • The vote was not closed until the end of 2005.
    • At the end of 9 April 2004, the count was 28-4 with one abstain.
    • The final count when it closed was 49-15 with two abstains.
  • changes since then - what legitimacy, if any, do these have if ratification was required to adopt the policy? Are there any substantive changes? I don't see many at the first read through.
  • Wikipedia:Arbitration_policy/Procedure_for_changing_this_policy indicates the existence of several past community efforts to change the way Arbcom does things.
  • Wikipedia:Arbitration_policy/Proposed_amendment_revote - a number of the proposed amendments had broader support than the original ratification. There was a requirement for WP:100 votes with an 80% majority in two weeks that was not met, but the original ratification vote did not even come close 100 votes in the year and a half it was left open.
    • Specifically:
      A1 vote count was 77-4.*
      A3 vote count was 40-3.
      C1 vote count was 54-4.*
      D2 vote count was 46-2.
      *One of the support votes was an account which had no edits before or since outside the vote.
    • Note that there was a strict 500-edit requirement for suffrage in the ratification vote that did not apply to the amendments. I have not taken this into account, but note that users with under 500 edits could note their opinion in the "other support" section of the ratification vote, and only one vote is present in that section (others did, but their votes were moved into the main support section as their edit counts rose).
  • Wikipedia:Arbitration_policy/Proposals may also be relevant

—Random832 17:20, 18 March 2008 (UTC)

[edit] Accountability

One issue I have with the present ArbCom process is the lack of accountability to the community. Committee members are appointed for three year terms, which is nearly an eternity in wiki-time. There has been much discussion about overriding or extending specific ArbCom decisions or indecisions. However, comparatively little seems to have been said about individual ArbCom members. Shouldn't there be some sort of mechanism for dealing with ArbCom members who have lost the faith of the community, and thus should be removed from their positions? I don't see where in the proposed RfC the possibility of instituting recall procedures would be added, but I see this as a necessary and long-overdue addition (as is admin recall, for that matter, but that's another discussion). I believe recent cases provide ample illustrations for the need for this addition, but believe this is not the place to present it at present. Please let me know if I am in error. Jay*Jay (talk) 08:44, 19 March 2008 (UTC)

Well, no, there shouldn't be any means of removing ArbCom members between terms. That would completely destroy ArbCom's ability to rise above petty political squabbles and deliver its unbiased, honest conclusions without fear of retribution from "the community." ArbCom is our judiciary, and while there is precedent for judges to stand for re-election from time to time, there is no precedent for allowing the deliberative process to be placed under threat of popularity contests. FCYTravis (talk) 16:58, 19 March 2008 (UTC)
Of course there should be a means of removing members between terms - and especially since they are so absurdly long. I am not suggesting that it should be a popularity process or a !vote, but if there is a genuine situation where the judgment of a member across a group of situations no longer enjoys the support of the community, that member should be removed. Relying on Jimbo to act in such a case is ridiculous, since this is a community- and consensus-based project. The US has procedures for removing Governors (recall elections) or Presidents (impeachment). Commonwealth countries can (and have) removed judges by parliamentary actions, and governments have even been removed. I am not suggesting a process that is easy to implement or which could be used as a weapon - it would probably require a substantial group to initiate - but none of that changes the fact that accountability is a reasonable check on the scope of the discretion enjoyed by ArbCom, and an appropriate inclusion in a balanced system. In WP terms, a process determined by bureaucrats where reasons for arguing that actions fall outside the reasonable bounds of discretion (and not simply that an editor disagrees with an action) would seem appropriate to me in the abstract. Jay*Jay (talk) 17:39, 19 March 2008 (UTC)
ArbCom members are not Wiki-politicians. They are Wiki-judges. I vehemently oppose any attempt to further politicize our dispute resolution process. It is not for no good reason that U.S. federal judges are appointed for life, and may not be removed except in cases of true malfeasance and corruption, through the impeachment process. We do not remove judges because we might disagree with their conclusions in a particular case. Such a process would be one step further toward mob rule over deliberation and thoughtfulness.
Three years is not "absurdly long." ArbCom is a deliberative body, and any shorter period of time would compromise their effectiveness. ArbCom is there precisely to take long-term views and provide perspective based on experience, not on hot-button fervor.
Just as in the case of administrator recall, which has been repeatedly rejected by the community, there is not a single example which can be pointed to as to why we would need this process. Who would be recalled? Why? What grounds would be sufficient to say something's "outside the reasonable bounds of discretion?" There is no evidence that any ArbCom member has ever done something which would justify being recalled - so why would we invent something to answer a question nobody has asked? FCYTravis (talk) 17:50, 19 March 2008 (UTC)
(edit conflict) I am not suggestion removal based on disagreements over conclusions in any one particular case, as the expressions (empahses added) judgment of a member across a group of situations and actions fall outside the reasonable bounds of discretion (and not simply that an editor disagrees with an action) make clear. Further, US federal judges' decisions are reviewable by the Supreme Court. Supreme Court actions can be overturned by legislative action. All three groups are constrained by a constitution. ArbCom decisions have very few of these sorts of limitations - the discretion allowed at present is essentially unlimited. Imposing a restriction of discretion within the broad region which might be considered reasonable to the community is neither overly restrictive nor unreasonable. Where reasonable people may disagree, discretion is fine - but consistent behaviours that lead to substantial disquiet about the competence of an ArbCom member should not be ignored if they were to exist. You note that federal judges can only be removed for true malfeasance and corruption - at present ArbCom members cannot be removed by the community, or by fellow ArbCom members, or by any other formal process even then. Such a situation is unsatisfactory and (I would contend) unjustifiable. Discussing whether provision should be made for some recall mechanism is perfectly reasonable, and this is not the place to debate what mechanism might be supported for adoption. Jay*Jay (talk) 18:04, 19 March 2008 (UTC)
Supreme Court decisions cannot be overturned by a mere "legislative action" as such. They can only be overruled by amendment to the Constitution - which is composed of legislative actions, but is far more involved of an undertaking than passing a law.
You need to define how the "judgment of a member across a group of situations" being "outside the reasonable bounds of discretion" would ever be actionable. You need to define what those "reasonable bounds of discretion" are.
"Significant disquiet" among a group of people is no reason to remove a judge. Earl Warren aroused "significant disquiet" among people when he led the Warren Court to a series of decisions which shattered precedent, along with segregation and discrimination laws.
I would support a mechanism which allows for ArbCom members to be impeached by ArbCom upon evidence of specific acts of malfeasance, corruption or abuse of office. No popular-vote system based on challenging an arbitrator's "discretion" is acceptable. That goes down the path of mob rule - which is precisely what the ArbCom was invented to prevent. FCYTravis (talk) 18:16, 19 March 2008 (UTC)