User:FT2/ACE2010 extended statement

I have also used the following alternate accounts:

1. Saduski J (talk · contribs · deleted contribs · logs · filter log · block user · block log) (November - December 2009):

   Used to experience being a newcomer of good intention but no experience, the initial learning curve we expect newcomers to absorb, and the extent to which help provided by the community is sufficient. Part for personal learning, part for the WMF strategy project.

   Created brief articles on Venezuelan (ex)ministers Jesse Chacón and Victoria Mata, but 'retired' on discovering how much there is to learn and how confusing Wikipedia is, despite being a quite well meaning and intelligent "newcomer". Convinced me that it is essential to prioritize better handholding and simplification for newcomers on arriving, for the long term survival of the project itself.

2. Ft2 (talk · contribs · deleted contribs · logs · filter log · block user · block log) (December 2008):

   Doppelganger account. Only edits are redirecting its user and talk pages to mine.

3. Account disclosed to Arbcom:

   Used to help co-ordinate real-world editor events. Nothing contentious whatsoever, and no other use or project activity. Unused for more than 1.5 years and I give a commitment it will not be used in future.

4. Update, account disclosed to Arbcom:

   Used for dealing with Grawp attacks. Unused since 2008. (Only other action is a null edit to a user's page at their request).

5. Update, Frank174 (talk · contribs · deleted contribs · logs · filter log · block user · block log):

   Created to experiment with editing as a newcomer 2 weeks before User:Saduski J but never actually used - name was accidentally disclosed hence ability to edit as a genuine newcomer was compromised.

I have only ever edited as an IP or another account on a handful of occasions, and only when accidentally not logged in to this one. Most were obvious from the page history or else I signed with my usual name immediately after. I have not had a previous account.

It is important to place any discussion of the events of 2008 in their proper context. It's also important to note that Arbcom has improved immeasurably since that time.

At 2007, Arbcom's status quo was still strongly bound to the style of 2003 - 2005. Its past traditions had been of the "old school" of Wikipedia: arbitrators were elected to do the best for the project and lengthy rules weren't needed to judge what was best. "What was best" often included avoiding public debate of decisions or maintaining silence in the face of external concerns. For example, appointment of functionaries (Checkuser and Oversight) was never discussed outside the committee (for fear of allowing a user to manipulate the decision) and even users expressing interest was suspect. Appointments were announced without notice and no details of any process were ever given. Users were at times banned with little information given, a strategy which backfired badly when abusive admin User:Runcorn (later known as Poetlister) played on this secrecy to persuade many that his identification as a sockpuppeteer was due to bad faith or the committee's incompetence. Communication was poor.

Internally as the community scale and case complexity grew, Arbcom's main activity was "firefighting". Proposals to improve internal process failed, left to lapse upon the next crisis. It had no formal system of leveraging improvements for this reason. Arbitrator conduct was more variable and at times arbitrators acted or posted inappropriately on-wiki. I did not find it malicious or corrupt, arbitrators tried hard and often burned out, but it was hopelessly entangled in poor self-organization and appeared to be unable to sort itself out. Any attempts were overridden by incoming emails, case work, and stress.

A number of arbitrators were open to change but some changes that might have resolved issues were marginalized due to conservatism (small "c") or doubt by some. As there was no formal decision process, these attempts were almost always dropped either because other issues intruded or due to lack of any system to manage and progress proposals.

In June 2008 I posted on the wiki, a complete decision page related to a ruling by the committee concerning a user, complete with findings and remedies. The user had not been consulted in any way nor did they have any opportunity to make representations. A proper explanation was never given how this came about, and the consequences were extremely disruptive. The brief version is as follows:

During June 2008 a user emailed Arbcom to appeal a community ban. In the course of examining his past conduct, a concern arose that the user had been "hounded" into a ban by having his actions related to Intelligent Design described in a misrepresentative manner to the community, which appeared to have not questioned the matter sufficiently or applied appropriate good faith. The user had sat out his ban and appealed properly. Attention came to focus on OrangeMarlin. The circumstances were reviewed and the issues identified were consulted; consensus suggested there was a concern. So far, all was routine. At this point what happened can only be described as "horribly wrong".

There was a real concern was that the topic area being so toxic, and the evidence perceived to be clear, there was a real question about the merits of yet more on-wiki disruption for months. The norm on Arbcom generated cases was one arbitrator would recuse and present the case. I had that role.

As was customary at that time, the usual handling was followed. The full case and all evidence was circulated to the committee with a proposed action, following which one waited to see if there was consensus.

This was a role I had not taken before, and so I undertook it with great care. I circulated the case and evidence with a description of the proposed action that would be taken - and not one arbitrator commented on it.

This was not entirely untoward. In many matters older Arbcoms operated by "silence gives consent": any arbitrator stating a proposed action, and waiting to see if there was any objection. If not, it was taken to be agreed. That was routine. However I had a concern that such a case needed to be sure of being correct, so I waited, then circulated it a second time. There were still no voices of dissent. Even so, in a level of care unprecedented in that year, I circulated the entire case a third time, with the description and a time by which the described action would be taken. There was still not one expression of dissent or concern.

When the matter exploded, the following 4 issues became clear: -

1. I had been briefed incorrectly or insufficiently earlier in the year by an existing arbitrator. Specifically, I had expressed concern that "silence means consent" was unreliable and had asked several times how to poll the committee for a definitive answer. I was told in no uncertain terms that people would read the emails, if they had objections they would speak, and if they did not speak they understood that to mean there were no objections. If there was silence in response to a check, it was to be taken as non-objection. The implication was that this was an established norm, and it was in fact how most decisions were made. It was strongly implied that not to understand this was a failing and midway between annoying and disruptive. I had checked three times (four including initial limited circulation), then following the given explanation, interpreted the repeated non-objection by every last arbitrator 3 or 4 times over as agreement.
2. The circulated email stated the case would be posted on-wiki for voting. There were differing understandings whether or not this included the proposed remedies. Some arbitrators thought it meant the proposed principles and findings would be posted allowing time for responses. Some, self included, thought it meant the entire case including proposed remedies.
3. I was out of contact for a few hours after the posting. This was not due to "post and run", rather it was because I had posted a large number of process proposals as well, and was under the impression that all of the committee were aware and had endorsed, and voting would take place over time.
4. To compound this, my computer power supply blew on return, leaving me with basic email and little else.

That is how OrangeMarlin's case came to be posted.

The resulting handling was that the case was vacated and blanked. A number of confused and conflicting statements were posted by arbitrators during the first few hours when nobody was sure what had gone wrong. I was not much able to get online to keep up and respond. The arbitrator who had provided the flawed description proved unwilling to explain on-wiki. The consensus of the committee was to defer explanation of "what had gone wrong" (and later let the idea of explaining lapse). A significant number of list members opined that this was a matter of scapegoating and that every arbitrator who could have spoken had been "asleep at the wheel" (the actual words used). A formal closing statement was posted, but once the drama died down no explanatory follow-up occurred. It read in part:

Role of FT2 - It was always an unlikely explanation that FT2, who is known for his careful and thorough work on and for Wikipedia, had wittingly gone outside and deliberately flouted our standard procedures. Part of the blame lies on email discussion as a way to get work done. The Committee takes collective responsibility for what occurred. Inferences that have been made, adverse to FT2's reputation for care, are simply not well founded.

I was disappointed that no public clarification took place. However the committee is a consensus body, and arbitrators are expected to place the project at a premium. There was no way I could explain it myself without either pointing fingers, causing harm, or creating disruption in overriding a consensus of the committee. I accepted the consequences of not speaking in my defense. The harm from overriding consensus to do so would have been considerable and the only person benefiting would have been myself personally.

There were other factors. As described below there was a death at the time, and also I had to weight the question - if I disclosed to the community that each member of the committee made a mistake of this size, what harm might it do? Arbcom should not be shielded if it makes mistakes. But equally nor should it be torn down by one person seeking personal vindication against the consensus of the entire committee. A vindicating statement had been posted. By the time I was reliably online again the matter had largely closed. I could re-open it or accept it. Reluctantly I accepted it. The arbitrator who mis-advised me has since stood down, was otherwise capable in the role, and remains a productive user. I am sure it was an unintended lapse. That was the 1st time I could have spoken, when it would not have led to witch hunting.

I felt somewhat let down - especially having taken care - but there is no guarantee of fairness. It can happen. I could not see a way to subsequently disclose the "how this happened" to the community at the time without adding disproportionate harm, and given that, I felt it better to take the aftermath personally than point fingers at any given person or feed more drama. Instead I dedicated the last 6 months of 2008 to addressing as a priority the systemic internal failings that had allowed the OrangeMarlin case to happen.

This was the other major event of the year requiring disclosure. I have given my word to avoid or minimize commenting on the case or the user, and will do so as best I can, but a few evidence links are needed to substantiate the statements below. The chronological summary includes events discovered later:

1. December 2007

During and after the Arbcom 2007 election, I was subjected to a campaign of defamation and harassment by a user (who was community banned in 2009, in part for still engaging in the same attack campaign 1.5 years later).

The attacking user had made a blog post during the election that referenced a Wikipedia edit and attempted to use it indirectly as a way of making a defamatory comment related to the author of that revision ("The author of this revision..." etc). The blog was taken down but had already been spidered and was "in the wild". David Gerard, an oversighter, oversighted the revision as potentially defamatory on the basis that the edit made the blog defamatory (by linking the blob claims to the author of the edit); with the edit removed, the link would fail and the defamation would not exist when the blog circulated at blog aggregators. Moments later it seems he reversed himself and realized it should have been deleted, which would have been sufficient. Gerard emailed Jimmy Wales and stated openly he had incorrectly used the tool and was in the wrong. Jimmy replied a day or so later that he would ask the developers about getting the edit reinstated again. For whatever reason this never happened.

This took place simultaneous to the 2007 election. My election Q&A was 370 K long - I was answering thoroughly in one of our community's toughest elections, and had almost no spare energy or attention for anything on-wiki except writing answers and skimming emails for whether or not action was required, a factor that most ACE candidates will know well and which turned out to be crucial later on.

The oversighting did not affect the election. The user had already had his concerns roundly dismissed by most of the community long before that happened. The purpose of the blog post was to incite off-site, not to educate Wikipedians, with users told they could "expect to hear MUCH more" [1][2]. The oversighting took place on 7 December 2007 22:02 UTC. By that time about 80% of all votes and personal votes had been cast with the edit being in full view, and even so, the election looked like this (final results). The mood was towards specific rejection of the matter and support for willingness to edit difficult areas [3][4][5][6][7][8][9][10].

From election onwards, I studiously ignored the user (who was banned at this point) and all references to him, in order to not be incited to feed the matter. I occasionally heard defamation was continuing elsewhere and a couple of times was witness or party to discussion of this by other users. I did not track this or read it. I did not visit WR. Perhaps I should have, in which case I would have understood comments alluded to by others.

2. March to July 2008

Between March and July there were a couple of mentions of "oversighting" but a check of records and logs shows that none of these made clear there was a reference of substantive claim of an actual real oversighting. Many wild and spurious claims had been made, and the tenor of the few mentions was vague ("an oversighting" or "the oversight") or seemed like a future concern - "what if he claims oversight was used".

A few emails were sent by an Arbitrator to the arbcom list. But these were sent to the private Arbcom list, where the only people who knew of the oversighted edits (Jimmy and David Gerard) were not members, and where I automatically filtered and archived them due to being related to the banned user. The older oversight log was inaccessible at the time for technical reasons (too long/no paging) so nobody could view the oversight log > 30 days ago either. The net result was that 1/ those who knew of an oversighting never saw the emails about it, 2/ I auto-archived all emails related to the user so I did not read them, and 3/ the old oversight log was uncheckable in any event.

In July a user asked me about the edits. This was the first time I took note of them, since an on-wiki inquiry is noteworthy.

This was 4 July. OrangeMarlin exploded around 27 June while away. My computer power supply failed a day or two after return. Someone very close to me was dying 2-3 July. I was using minimal technology and trying to keep up, taking care of my family, and lacking sleep. I had not registered the oversighted edits as an "issue" and had not had them raised to me as a real matter before; tiredness and lack of tracking the affairs of the banned user caused me to not recall the few brief mentions. I believed my answer was 100% accurate, honest, and complete, that this was the first mention of them to me (in effect it almost was, previous mentions had been vague in the extreme or not seen by me), and I was trying to help. I checked the oversight log (of course) but having little idea what edits to look for, the first thing I noticed was that the oversight log was not working. I said as much on-wiki, said what I believed, and promised to check it up when the log was back. I thought no more of the exchange.

I also gather that some assumptions made by others may have been in error, 1/ it was assumed the emails had gone to a list where I or someone would have recognized their significance, and 2/ there was a separate oversighting of material by the user (100% legitimate) discussed on May 3 which may have been conflated with the previous ones.

3. Sept - Dec 2008

The user was rebanned in September. The Oversight log was re-enabled. I found the edits and spoke to David Gerard who confirmed the matter and copied me his disclosure/apology to Jimbo. I was sent copies of various old (2007 and early 2008) emails between others that had not been sent to me before related to the user's actions.

The user appealed his ban and the entire matter went to RFAR. On December 9, in an unrelated search related to the RFAR case, I found an old email from Jimbo to myself. It was written during the election a year before in 2007 and stated that edits had been oversighted by mistake but this was going to be reversed.

This was a problem. Unknowingly I found from an email in December 2007, that I had in fact misled the community in July 2008. Not that it would have made a practical difference; I still hadn't any involvement or done wrong, and the email had arrived at an immensely busy time, but I'd been sent an email about it and not remembered.

In judging honesty, consider my next action. This was an email that could only put me in a very negative light. Nobody else knew about this email. It was a year old and Jimbo had surely forgotten its existence if (as it seemed) he couldn't remember the entire matter much himself. Not one other person on the project knew. I could have ignored it and nobody would ever have known. Instead I emailed Jimmy and David Gerard the same day, disclosed the discovery, and asked their advice. The replies were from Gerard - to disclose. The problem was that the oversight had been blown up into a question of the proper use of oversight, I think partly by people targeting David Gerard (not 100% sure). A full disclosure would also lead to followup questions focusing on real-world information leaked from Arbcom's list, and that was also a problem.

(Briefly it had been asked whether had he "pulled strings". The election was all but final at the time of oversighting so it had no effect whatsoever. Gerard had indeed met me (unexpectedly) at an "open source" discussion and on learning who I was, said "welcome to hell, you'll need a shovel" or the like, as the election was visibly likely by then. But to discuss or confirm this would confirm people's real-world information and the accuracy of a leak from Arbcom's mailing list. Any comment on this point - whether acknowledge, confirm or deny - had to be weighed against the potential impact on the privacy of individuals and the list.)

I consulted widely. I asked advice from Arbitrators individually and the committee as a whole, from people I trusted and people unconnected. The advice was conflicting. Nobody could suggest a concrete disclosure that would deal with the original matter but not lead to likely breach of the privacy leak issue.

I emailed Thatcher and Jimbo at some point: "But mainly [...] it's already my long standing intention to be open and seek full disclosure [...] in the next few days, and several have tried to talk me out of it. I've declined citing that I want full accountability to the community. If afterwards you are not satisfied, I will have said all that there is to say, and the community and Jimbo may assess me as they will."

4. I stand down.

Privacy policy is not lightly broken. Drama was rife in other areas. Ultimately I couldn't figure out a way to avoid doing harm if I posted. I lacked anyone to reliably consult with. The new Arbcom (AC2009) were appointed at the time; they wanted to pursue their agenda and this was fairly obviously an undesirable problem to adopt. Ultimately the position became untenable. There was no process or body in the community able to examine the private evidence and rule on it, that would be considered neutral and authoritative if (as I expected) I was cleared. I was concerned that Arbcom's trust from the community would be harmed even if I were deemed to have acted well. I stepped down because an arbitrator should be visibly able to show their integrity or step down, and through the nature of the situation I had no route to clear mine. It was the right thing to do. I do not make excuses for the mess and I believe my draft start at disclosure is still on the Arbcom wiki, testimony to the intent. If there was a solution, it was an exceptional matter and I could not find any answer at the time. I did the best I could be sure of. If it was insufficient I am deeply sorry.

On 15 January 2009 I wrote to Jimmy Wales as follows (extract) and explained I was stepping down.

Analysis of problem

Two specific matters went bad. These were Orangemarlin and the present "oversighted edits" issue. The community has a legitimate concern whether the messenger was the creator of these (by substandard care or wilful misconduct), or whether a reasonable or high standard of care existed and in fact the issue was of different origins. There is pivotal information in both that remains rightly non-public at this time. [...]

Community dispute resolution processes cannot resolve this. They may at best see it dissipate, in which case much of the raw data to form a view is unknown and will remain so. This would be unfair to the community, to the committee, and to myself.

Arbcom itself is at risk of being too close to the issue, and hence it rightly may fear being tainted or overshadowed in its 2009 reform agenda by this matter, which it may rightly regard as a 2008 "albatross". It may be torn between wanting fairness, wanting to recognize communal feelings, and wanting to rid itself of the issue entirely by "cutting the Gordian knot". The latter must surely be extremely tempting in terms of resolution.

Although I have asked about self-requested RFAR -- our highest level of dispute handling -- in my view Arbcom cannot easily take on this role (despite being "within remit") as it has a more important job to do, namely serving and focusing on the community. There is a possibility that in fact it cannot take on this case in a practical sense without imperiling its other remits. I may be wrong, but that would be a clear point to consider.

Finally, you yourself cannot take on this issue either. You would usually be the resort for a case Arbcom cannot handle, but it would put you in a difficult position, perhaps transferring some grudges from myself to you if you did decide there was no or little substantive issue. You might also be accused of being too close, due to prior involvement and possibly a prior view. As project leader you may hold and execute whatever view (in that role) you wish -- you may ban a user, change their role or access, or instate them in a role or access. I do not believe on this case that you can readily run the inquiry.

[...]

Some of these are so entrenched that popular wisdom says they cannot be solved. Four well known examples:

• The "civility problem" (views on civility and incivility diverge)
• Content disputes (especially sourcing and weight/fringe issues)
• Ethnic issues (including naming debates which are often tied to these)
• Articles and topics with long term warring

I have had since 2008 to consider these. I believe we need to take them seriously and reach agreement, but views are extremely diverse how to do so. My criteria for content disputes are as follows:

1. Approaches must be community driven. Not Arbcom or some expert panel, but within the community.
2. Approaches must be robust. They will meet our toughest disputes.
3. Approaches must be on-wiki and public. There is no user privacy issues on content disputes so no other venue is needed.
4. Approaches should rely more on peer consensus of collaborating editors than threat of sanctions, and should aim to make that collaboration easy and neutral.
5. Approaches must not in any way arbitrate content - there must never be a "wikipedia official view" or a threat that a dissenting user citing other facts will be seen as disruptive for their concern that a content issue is mis-stated or knowledge has improved.
6. However, stability is needed. The same issues must not be dragged up constantly without good cause, a working consensus or resolution must happen at some point, and issues must be able to be revisited as needed, but not tendentiously.

I hope to trial two fairly gentle proposals next year that meet the above criteria and handle the two most common kinds of problem - a consensus forming approach and a dispute resolution approach. I hope to write both up and will ask for a trial of both during the year. Anyone interested please sign on the talk page.

There is a third proposal that I would like to see trialled in 2011. This is more "heavyweight" and only suited to one specific issue - the very worst kinds of warring where socking, gaming, misrepresentation, tag teams, and tendentiousness are endemic. It would involve a new kind of community restriction (that hasn't been used before) being set on the topic's editing for a limited period of up to than 4 months, but it has the potential to remedy a great many edit warred topics thereafter. Again not "committee" or "expert" based and open to all users. It would probably be a little controversial and cause some initial discussion because it's unusual, but I'd like to see it trialled if the community permits. Insight from such disputes in the past suggests it may do what the community would like to see.