Wikipedia talk:Arbitration/Requests/Case/Rich Farmbrough/Proposed decision

Main case page (Talk) — Evidence (Talk) — Workshop (Talk) — Proposed decision (Talk)

Case clerk: TBD Drafting arbitrator: TBD

Behaviour on this page: Arbitration case pages exist to assist the Arbitration Committee in arriving at a fair, well-informed decision. You are required to act with appropriate decorum during this case. While grievances must often be aired during a case, you are expected to air them without being rude or hostile, and to respond calmly to allegations against you. Accusations of misbehaviour posted in this case must be proven with clear evidence (and otherwise not made at all). Editors who conduct themselves inappropriately during a case may be sanctioned by an arbitrator, clerk, or functionary, without further warning, by being banned from further participation in the case, or being blocked altogether. Personal attacks against other users, including arbitrators or the clerks, will be met with sanctions. Behavior during a case may also be considered by the committee in arriving at a final decision.

Clarification on principle 6

I've seen bots get blocked for being malfunctioning, and their owners fixing the problem and then unblocking the bot. Would this no longer be permitted? --Rschen7754 20:25, 5 May 2012 (UTC)

As far as I know, no. Rich Farmbrough has inappropriately been unblocking his bots when he wasn't supposed to such as unblocking while he was technically blocked or unblocking without fixing the issues that were brought up by the community.—cyberpower ^Chat_Online 20:29, 5 May 2012 (UTC)

It seems that !X changed the wording back in 2008 - with no link to any discussion. But it was news to me. Certainly most, if not all, of the unblocks I did had explicit permission from the blocker to unblock. And it's not clear that wording of a MediaWiki message is policy - well actually it's pretty clear that it isn't. Rich Farmbrough, 21:11, 5 May 2012 (UTC).

(edit conflict)I don't know the specifics. I'm just going off of what I am seeing.—cyberpower ^Chat_Online 21:14, 5 May 2012 (UTC)

(Yep, it's not a big deal. Rich Farmbrough, 21:22, 5 May 2012 (UTC).)

Well, it would just be very unwise to unblock your own bots from now on - your bot gets blocked for some unanticipated mistake (which according to ArbCom is already a crime, you should have anticipated mistakes) - then you do your best fixing the problem, changing the code, etc. etc. - if you then unblock, the bot edits on, and the fix is not a perfect fix, and the bot makes a very related, or even the same, unanticipated mistake, then you would have unblocked your bot while not fixing the problem. And no-one can see whether you REALLY tried to fix your bot or not. --Dirk Beetstra ^{T C} 05:34, 6 May 2012 (UTC)

Interesting

Here I gently chide an admin and BAG member for using evidence of manual editing (saving something then realizing the mistake and fixing it) to support a claim of automatic editing. Citing a <facepalm> as incivility worthy of a year's ban, desysopping and banning from using automation is cazy. Rich Farmbrough, 21:28, 5 May 2012 (UTC).

And again for suggesting a user be polite and friendly in reporting issues.

And for suggesting a user take the advice of another to work colliagally. Rich Farmbrough, 21:46, 5 May 2012 (UTC).

3.3.2: "Ban on automation"

Regarding 3.3.2: how would someone on WP:AE determine whether Rich was using automation? It's notoriously difficult to tell what method was used to make an edit, and if Rich states that a sequence of 1,000 edits (say) was all made manually, there would be no evidence that could contradict this. Even 1,000 edits could be made manually, using tabbed browsing and patience. — Carl (CBM · talk) 00:08, 6 May 2012 (UTC)

It looks as though this is the main reason behind the proposed site-ban; unfortunately, unless he really chugs away at something (as noted in some of the contributions lists I posted in evidence, and posted in the findings here), there really isn't any way to definitively tell short of checkuser, and even that's not 100% reliable. What may work better is limiting him to a very low edit rate, such as no more than 4 edits/minute, such that using automation would actually become more tedious than simply doing it by hand. Hersfold ^(t/a/c) 00:37, 6 May 2012 (UTC)

Really I can't believe I bothered to do stuff manually, I must be an idiot. Not only does everyone assume it's automatic anyway, they also get the "evidence" completely arse-about-face.

• Assisted editing - fast, because you have to be there.
• Bot editing, fast or slow, as long as the job gets done.

If people think a 4 edits/minute speed restriction on my human account is a good idea, I could cope with it. It does nothing though. Rich Farmbrough, 01:16, 6 May 2012 (UTC).(Using some automation)

Indeed, before bots I was using tabbed browsing and a Firefox plugin called Linky to open 100 tabs per window. It was much faster than AWB, because it loaded 100 pages simultaneously, the speed limit was how fast you could change tabs, fix the page and hit save - which in turn was CPU limited, due to the complexity of Netscape/Firefox and the overhead that Wikipedia pages carry. Rich Farmbrough, 01:21, 6 May 2012 (UTC).(Using some automation)

Rich, the point is not whether you use automation or not - the point is that you are banned from using it. They do not have to prove that you are editing automated, if you do 3 edits in a row where you manually clean up three things (e.g., you used Google to find a strange misspelling in three Wikipedia pages), and you save those 3 edits (well within a 4-edit per minute level), you will be 'editing automated'. What, you might even be told off if you do those three edit on 3 different days, and editing in between respectively 53 and 69 other pages - you know how these edit restrictions are and can be used, you know what happened to Δ. And don't worry, because you are now here once, you will re-appear before ArbCom within a year after you are unbanned, and they will cite recividism, and even if they then also can't find anything substantive, they will ban you because you were banned before. There is probably only one way to handle with this - leave. --Dirk Beetstra ^{T C} 10:38, 6 May 2012 (UTC)

And assuming good faith has gone straight out the window... yes, the evil ArbCom is seeking to ban everyone on the project. The secret's out now... :-/ Hersfold ^(t/a/c) 17:22, 6 May 2012 (UTC)

3.2.5 Rich Farmbrough's undisclosed use of automation

•  Fixed all my edits will now reveal if they have automation assistance. Rich Farmbrough, 01:36, 6 May 2012 (UTC).(Using some automation)
  

Which is rather moot, especially if you say that every edit you make. Headbomb {talk / contribs / physics / books} 21:31, 6 May 2012 (UTC)

Comment

Dirk Beetstra asked me to comment here after a comment I made on Rich's talk page.

I've never had a problem with Rich's edits. I know people have been complaining about him for a long time, but the proposed solution seems awfully harsh for the offense. The incivility is barely uncivil, and half of the offenses date from 2010. It would thus seem to be a low-level irritant, hardly the kind of thing to ban someone over. The argument that you need to ban/etc. him because how would you know if he evades his restrictions seems backwards to me. I would expect that you would impose the restrictions, and if he evades them, then ban/etc. him. Well, I've never been on ArbCom, so I don't know how to do your job, but that's the impression I get as an outsider.

Also, Rich has been unfailingly helpful the times I've dealt with him. Currently I have a bot request that's been languishing for over a month. I wanted to lay the groundwork for a Wikiproject project to properly format the references of our thousands of language articles, by adding the necessary parameters to transclusions of the infobox so they can be quickly reviewed by hand. It got hung up on whether the bot should add a reference section, or whether a different bot should clean up afterwards by adding a ref section, but even after I removed that item from the request, it just sat there. (I've just posted it for the third time.) Rich was willing to do it, but his month block was imposed while we were still debating whether it was appropriate to add the reference section. He'd be willing to do it now, but for the threat of being banned. And now I'm getting busy enough with other things that I don't know if I'll be able to start the project if the request is ever approved—which would mean leaving c. 4,000 language articles without overt references, so that editors continually mis-tag them as unreferenced despite the fact that they are referenced.

No-one else steps in to take over from Rich when he's not here. — kwami (talk) 07:35, 6 May 2012 (UTC)

The request is not hung up because of the reference section, the request is hung up because the task's bot operator is at ARBCOM, partially because of his behaviour during that task's BRFA. Headbomb {talk / contribs / physics / books} 07:51, 6 May 2012 (UTC)

Rich Farmbrough's responsiveness

@Kirill - Fram is part of the community, Fram is not the representative of the community, Fram is not the community. I ask, again, can you show that the community finds problem with Rich's edits, not only that Fram finds problem with the edits. --Dirk Beetstra ^{T C} 08:58, 6 May 2012 (UTC)

For example, there is the discussion that led to the first edit restriction [1]. Here is a particularly apropos comment from it:

Your dismissive attitude ("No one cares..."; "Forget it and go and write an encyclopedia."; "There is no controversy here. Nothing to see, keep walking.") is a major part of the problem. I don't know how many people care about this, but I can tell you that I do. When I consult diffs to evaluate the edits, your bot's inconsequential changes waste my time. I've gone to your talk page to raise the issue, only to be reminded by the existing complaints (and your [non-]responses thereto) that you routinely ignore/dismiss such criticisms. So I don't bother to add my voice to the futile chorus (and I assume that others act in kind). —David Levy 23:56, 24 October 2010 (UTC)

— Carl (CBM · talk) 09:17, 6 May 2012 (UTC)

Is there no alternative to a siteban?

I understand that Rich's attitude to previous sanctions is telling against him here, but is there no other solution than sitebanning him. Could one maybe just restrict him from making any edits in Article space? He could contribute at places like Wikipedia:WikiProject Perl, he's been giving tutorials in coding which could continue, he just needs to stop contributing using automated tools himself for a bit. --Elen of the Roads (talk) 10:55, 6 May 2012 (UTC)

+1. A site ban should not be necessary. Given that the problems have been with automated tools, ban on use of those is quite enough. I understand there are enforcement concerns (particularly in relation to 3.2.5, undisclosed use of automation), but I think that can be worked out. (If no better solution can be found, Elen's suggestion of restriction from article space would work; or maybe a lowish daily limit on number of edits in mainspace would work.) I said some time ago that Rich's skills can be very usefully applied supporting others on bot issues, and that should not be prevented, if he's willing. Rd232 ^talk 11:51, 6 May 2012 (UTC)

If ArbCom told me not use tools, I wouldn't. Nothing would get done though. Rich Farmbrough, 12:22, 6 May 2012 (UTC).(Using some automation)

I don't think thát will be a problem, Rich. --Dirk Beetstra ^{T C} 12:40, 6 May 2012 (UTC)

I'm not necessarily opposed to resolving this without a site ban. However, for that to be viable, we need some way of enforcing the ban on automation; and, given Rich's unwillingness to disclose whether he is using automation tools (and which tools, and for what purpose), I don't see how that can be done. I certainly don't want to just dump the problem on the lap of the administrators working at AE; they have better things to do than monitoring the timing of Rich's edits.

I don't think a ban from article space would be sufficient, given that one of the major complaints was Rich's mass creation of categories. More generally, banning him from particular namespaces doesn't seem like an effective strategy; virtually every namespace is amenable to the use of automation (e.g. template redirects on talk pages, meta-template invocations on template pages, etc.), and I see no fundamental reason to believe that Rich would be any less willing or able to use automation outside the article space. Kirill ^[talk] 12:49, 6 May 2012 (UTC)

A bot could always check up on Rich's edits. If suspicious automated behavior happens, the bot can always notify AN and ArbCom about potential automated activity.—cyberpower ^Chat_Offline 12:57, 6 May 2012 (UTC)

Do we have a bot that can check for "suspicious automated behavior"? If not, who is going to create one? Kirill ^[talk] 13:04, 6 May 2012 (UTC)

Don't think a bot would be necessary, no doubt Fram and CBM will be happy to continue monitoring Rich's contribs. Jenks24 (talk) 13:27, 6 May 2012 (UTC)

@Kirill, maybe you could consider to ask Rich Farmbrough or Δ to write that bot for you? --Dirk Beetstra ^{T C} 13:31, 6 May 2012 (UTC)

User:Cobi could write the bot. By slightly tweaking ClueBot NG analytic scripts it could be used to analyze Rich's Contributions using ClueBot RF. Since ClueBot NG is artificially intelligent it will learn off of Rich's editing style and would be able to determine more precisely if he is using automation or not.—cyberpower ^Chat_Online 14:04, 6 May 2012 (UTC)

No, this is not a viable option. Cobi is a very capable programmer, yes, but unless he's able to design a true artificial intelligence system (thinking Skynet or HAL 9000 here, and look how well those turned out), the judgment of what merits "suspicion" can only be made by a human. Yes, we have vandalism bots, such as ClueBot, that have a high rate of accuracy, but they're designed to look for multiple blatant criteria; if something doesn't go over a minimum threshold score, the bot won't touch it. They are not, AFAIK, designed to "learn." Besides that, if a human can't readily determine if a set of edits were made manually or not, how could a human write a program to make that decision for them? Bottom line is, computer programs are really quite remarkably stupid, and (until we have true AI) incapable of exceeding or in many cases reaching a human's ability of judgment. Hersfold ^(t/a/c) 17:29, 6 May 2012 (UTC)

According to Cobi, ClueBot NG is an AI bot with learning algorithims. And we don't need a perfect bot, just one that is able to detect obvious automation scripts being used and report it.—cyberpower ^Chat_Online 19:46, 6 May 2012 (UTC)

All I can say is that reading this does not reflect well on either ArbCom or Rich. One can think of all the metaphors (Sledgehammer <-> Nuts; Baby <-> bath water; etc.). I would encourage both the ArbCom and Rich to step back, to both get over their respective righteousness and self-importance, resolve this problem, apply mutual respect and all get back to work. What a waste of good time and electrons. — billinghurst sDrewth 13:51, 6 May 2012 (UTC)

I'm remaining impartial about this. I'm only throwing suggestions and possible solutions to potential problems.—cyberpower ^Chat_Online 14:04, 6 May 2012 (UTC)

Why are some of you trying to bend over backwards to create a means of checking on an editor's contributions when they have apparently shown themselves incapable of living up to the trust that their privileged position already afforded them? Leaky Caldron 14:08, 6 May 2012 (UTC)

That's the problem. When it gets down to talking about drone-slapping someone, you know it's not going anywhere. I do think it would be worth one try - no automated edits, no scripts, no bots, no repeating the same task faster than once every ten minutes, no mass creations of anything. Talk about things, discuss them, make content creating edits to articles, offer tutorials on coding, but if a problem that needs automated editing appears, just ask for someone to do it at Village pump or wherever. One chance. But would it drive Rich mad, and is it reasonable? Elen of the Roads (talk) 14:39, 6 May 2012 (UTC)

It's apparent that Rich is very attached to automation like I am too (my status and signature). I would certainly go mad if I I'm editing in an environment where I can write assistance scripts and so forth and not be able to use them.—cyberpower ^Chat_Online 19:46, 6 May 2012 (UTC)

I think the whole automation thing is a red herring. Is it a good edit? Is it a bad edit? Some edits we can have a debate about, but that is not what is being discussed here. People have been kind enough to say that I "do a lot of good work" - maybe someone takes exception to that or maybe we should construct a balance and put grinding through hundreds of elementary particle templates on one side (as good), and removing markup (per guidance) on the other side (as bad). This is the model Wnt tried to introduce - and I think there is something to it, except that it is too complex to actually execute. I would encourage people to try it as a thought exercise. We have maybe 750 edits to this case. Maybe they are neutral, maybe they go to community building, (plus) maybe time wasting (minus). 1500 edits to tag some maps with GFDL, that was before automation at all "open in new tab., edit, ctrl F, ctrl V ctrl S (that was the save keystroke in those days), tab back, repeat, repeat..." - they are all on Commons now, so maybe that's neutral. Fixing up 30,000 US places demographics to be readable and correct text, including 3000 manually. Probably 100,000 typo fixes. Maintaining some 4,000 dated maintenance categories. That's all gotta be good. Being sharp when people are rude to me. Yeah, that's kinda negative (in some ways) - but I'm generally pretty laid back. I got called a "fucking liar" and didn't respond, and far worse things. I think I come out ahead, or at least break even on civility.

So the real question is are we looking backward or forwards? If we are looking backwards I think I have a substantial net contribution to the project. The worst that is being brought against me is saying "Tosh" or <sigh>, or deleting a few trailing spaces. And remember some of the things that have been fussed about, I have gone off and got consensus on, with no objections - as it seems to me was bound to happen.

And then if we look forward, what have I to offer? Well consider that on 24th March I addressed practically every incomplete item on the WP:BOTREQ, and they would all be done if not for blocking (over something we all now agree was pretty harmless, and most people would probably call useful). So as Kwami says, what has happened since? Basically sweet Fanny Adams, except that part of his task came within the purview of a previous BRFA and so could be completed by HPB. Does the community want me around, submitting BRFAs like they're going out of fashion, and solving issues for editors and readers, or would they rather BOTREqs get archived undone?

TLDR - I know. Rich Farmbrough, 01:17, 7 May 2012 (UTC).(Using some automation)

Rich, the reality is that - no matter how much work there is to do - none of us are indispensible. If we go under a bus/move to a desert island, there's always someone else will stop the world from ending. The best you can hope for is avoiding a siteban (and I don't think the situation has reached that point yet), but you need to start looking at how else you can contribute, if you avoid the problem areas in Anomie's explanation of 'automated' editing. Marking all your edits (automated) just because you use that fancy javascript is making it look as if you're not prepared to recognise the difference, or - worse - that you're taking the piss (don't think you are, but that is how it is looking to some people).Elen of the Roads (talk) 10:21, 7 May 2012 (UTC)

You're right, Elen, it is just that the rubbish will now stay there longer. Sure, we are all indispensible, but losing an editor, any editor, is a loss. However you (pl.) want to wiggle your way out of it, saying that what is happening here is not-so-bad for Wikipedia, saying that it is only a ban of one productive, knowledgeable editor. But if that keeps you (pl.) happy. --Dirk Beetstra ^{T C} 10:30, 7 May 2012 (UTC)

I'm not saying that, so please don't put words in my mouth. I'm saying that the argument about indispensibility and workload has never succeeded before, and probably won't cut much ice now. It would be better for Rich to engage with the discussion about a restriction. He's already said that if asked, he wouldn't make automated edits, but then going and marking all his edits as automated (while I can see that it is honest in one sense, give the amount of javascript it takes him to make a talkpage post) hasn't persuaded people that he understands the distinctions that Anomie explains below. Which I'm sure he actually does understand. Elen of the Roads (talk) 10:41, 7 May 2012 (UTC)

The you was aimed at the community (hence the 'pl.'). But, as usual, who cares whether high volume editors who do a lot of good work leave (or are forced to (temporarily) leave - though this is simply a one-way ticket to eternity). Editors could actually choose to fight and keep the editors here who are not physically 'forced' to leave. --Dirk Beetstra ^{T C} 13:55, 7 May 2012 (UTC)

Reset you calendars!

As I understand, ArbCom has no trouble to find that Rich has used undisclosed automation under 3.2. But still someone makes the argument that Rich needs to be site-banned because it's impossible to tell if he uses automation under 3.3? And people vote in favour? Galls (and guys), April 1st has been over for more than a month. I would also suggest you read up on Turing test. If you cannot figure out if certain edits are automatic, who cares? And why? --Stephan Schulz (talk) 14:58, 6 May 2012 (UTC)

It is a problem if the community feels someone is "getting round" a restriction by "gaming the technology". You get bad feelings and arguments. In a previous rerun of this very argument, a different editor persisted in making lots-of-very-fast-repetetive-edits-with-a-ton-of-mistakes. In the end, it didn't matter whether he used AWB or a team of monkeys to achieve the effect, he ended up sitebanned. In this case, it's "very fast repetitive edits that people disagree with the consensus of" that's the problem. Rich is an intelligent guy - but so far he consistently refused to stop making that kind of edit. Really he needs agree to stop making lots of edits very fast with marginal consensus, whether that means abandoning the bots and scripts, or sacking the monkeys. --Elen of the Roads (talk) 15:12, 6 May 2012 (UTC)

I think the experience with the Delta edit restrictions showed some of the problems when comes to assessing speed. Assessing similarity of edits could similarly be a problem. I think it would be a lot simpler to just trust Rich to stop doing that kind of edit, whilst making a provision that if he appears to be breaching that trust (with some reasonable scope for clarification on what is permitted - any clarifications should be absolutely respected), that he be banned (via WP:AE discussion) from all edits except for talk namespaces and the Wikipedia namespace. Rd232 ^talk 15:28, 6 May 2012 (UTC)

(Re. Elen): That's a reasonable argument. But a) it's not the argument ArbCom is making, and b) it's not really responsive to my concern. The argument at best supports something like an edit frequency limitation. And I'm really not very susceptible to arguments that "the community feels" something. We don't have a feel-o-meter, and a large part of the community probably has no particular feeling at all. I only ever noticed this case (and even became aware of Rich) by accident because of a caricature on Jimbo's talk page. Most users probably know nothing of this case. I would also maintain that if someones feelings are hurt by a change from {{Reflist}} to {{reflist}} (or vice versa), they are in dire need of a LART application and we should not base serious decisions on their concerns. --Stephan Schulz (talk) 15:31, 6 May 2012 (UTC)

Stephan, I think you may have slightly misunderstood my argument in favor of 3.3.3. My assertion is not that it's per se impossible to determine whether Rich is using automation; as you point out, we've done exactly that in 3.2.5. Rather, my assertion is that identifying undisclosed automation by way of a subjective analysis of edit timing is neither sufficiently easy nor sufficiently reliable to form an effective method for enforcing a ban on automation (3.3.2).

The situation would be different if the evidence showed that Rich was consistently open and forthright regarding whether he was using automation; were that the case, I would be willing to accept his assertion that he was no longer using automation tools at face value. As 3.2.5 shows, however, Rich has not been open regarding his past use of automation; and thus, to put it quite bluntly, we cannot simply trust him to comply with 3.3.2 voluntarily, and must instead provide a mechanism to forcibly ensure compliance—a mechanism which we are, unfortunately, lacking. Kirill ^[talk] 15:43, 6 May 2012 (UTC)

Thanks for the feedback. Yes, I've oversimplified the argument a bit to show its absurdity. And I still see this absurdity. We either can determine if someone is editing automatically, or we cannot. If we can, we can enforce an automation ban. If we can't, it does not matter if someone is editing automatically or not. And if we want to ban automation "just because", not because of any concrete problems, I strongly maintain that we are very much on the wrong track. Isn't "prevention, not punishment" still policy? --Stephan Schulz (talk) 17:23, 6 May 2012 (UTC)

And you have rather oversimplified this "hiding automation". If the community says I must not edit using AWB, or stand alone client programs of my own devising, or both, so be it. No such injunction has been made, therefore I have nothing to hide. I have never stated that the edits referred to were not made with AWB, I have not even been asked, for the record (E&OE) they were. Rich Farmbrough, 23:39, 6 May 2012 (UTC).(Using some automation)

Comments by Anomie

As a BAG member, but not speaking for BAG as a group, I'd like to make a few comments:

• Regarding the "Automation tools" principle, the line between a task requiring BAG approval and a task not requiring BAG approval is very fuzzy, but in general there are three factors to be considered:
  1. The level of automation involved. If edits are made without human intervention, it's considered "fully automated" and requires approval. If the edits are prepared by the bot or script, but then a human must review the diff and manually hit the "Save page" button, it's considered "semi-automated" or "script-assisted manual editing" and may or may not require BAG approval. Editing without any scripted assistance is considered "manual editing".

     The line between "semi-automated" and "script-assisted manual", in my opinion, is another fuzzy line based on the level of human involvement is required in selecting and preparing the edit before presenting it to the user for review. If the script chooses the page and the edits to apply and the human just approves the edit, it's "semi-automated". If the human selects the page and then uses a button to activate the particular script, it's "script-assisted manual". WP:BOTASSIST addresses this issue, as well. In any case, the distinction rarely matters much except when someone is trying to wikilawyer around an accusation of running an unauthorized bot (e.g. the kind of wikilawyering WP:MEATBOT is intended to prevent).

  2. The number of edits involved. If a semi-automated or (supposedly) manual task is going to be affecting a large number of pages, BAG approval may be desirable or required if only to avoid the hassle of being accused of being an unapproved bot. OTOH, in some cases a discussion demonstrating consensus at WP:VP or in an RFC may serve as well, particularly for a manual task. And in some cases, as in the oft-ignored WP:MASSCREATION, the community has explicitly decided that BAG approval is required regardless of the level of automation.
  3. The speed of edits involved. For semi-automated or (supposedly) manual tasks making edits at a high rate of speed for a sustained period of time, BAG approval may be desirable for the same reasons.
  None of this has anything to do with the technology used for automation. AWB may be used both for tasks requiring BAG approval and tasks not requiring BAG approval. A fully-automated bot could be written in Javascript and would require BAG approval, although most user scripts don't affect editing in any way (and are thus entirely outside the remit of BAG; WP:BOTSCRIPT addresses this point) and those that do are often "script-assisted manual editing" or "semi-automated" at most. In my opinion, both the first and second drafts of this principle accurate reflect the situation, although the second suffers somewhat from focusing on the tool rather than the task. Anomie⚔ 16:56, 6 May 2012 (UTC)
• Regarding the "Identifying the use of automation tools" principle, a related point in WP:BOTPOL, specifically WP:MEATBOT, states that even an ostensibly manual editing process may be in effect a fully-automated bot with a "meat" component if the human is not exercising their human judgment; the intent and wording is specifically aimed at nullifying attempts at a "this was entirely manual" defense when the WP:DUCK test (as mentioned by Jclemens) indicates an unapproved bot. Anomie⚔ 16:56, 6 May 2012 (UTC)
• Regarding the "Unblocking a bot" principle, ArbCom may also want to consider the not-uncommon case where the blocking administrator explicitly gives "permission" for the bot operator to unblock when the problem is fixed, usually justified per WP:NOTBURO. Anomie⚔ 16:56, 6 May 2012 (UTC)

As an editor, I'd also like to comment on something unrelated to Rich Farmbrough's bot activity. In regard to the "Rich Farmbrough's administrator status revoked" remedy, besides Rich's unblocking of his own bots, there have also been issues in the past regarding Rich's making untested, controversial, and/or contra-consensus edits to highly-visible fully-protected templates, with the same attitudes discussed more fully here with respect to his bot operation. I see Fram touched on this in his evidence, and it was mentioned in passing by Fram, CBM, and Elen of the Roads in the workshop. Wikipedia:Administrators' noticeboard/Incidents/Rich Farmbrough/October 2010 may also be relevant. Anomie⚔ 16:56, 6 May 2012 (UTC)

Indeed, I think the previous principle automated editing broke it down relatively well with no such issues as there are on 3.1.3; Anomie's breakdown is much more thorough and very well explained. :) — madman 03:56, 7 May 2012 (UTC)

Comment on 3.2.3

This section reads... "Examples include cosmetic changes to non-rendered whitespace (A, B, C, D, E), cosmetic changes to template invocations (F, G), removal of comments (H, I), and unapproved mass creation of categories (J)."

Regarding

• A: This edit was done before the restriction.
• B: Removing whitespace at the end of lines is not a contentious thing, as long as it's not done on its own, I'm unsure if this comes with stock AWB (so it could be a violation of the edit-restriction), but it certainly could be implemented (AKA, if it's a violation, it's not an egregious one). The main purpose of this edit was to remove a linked date fragment. I don't know if this task had consensus, but it is not a case of WP:COSMETICBOT.
• C: Same as above, except with ISBN hyphenation (which is an approved task).
• D: Same as above, except with typo fixing + tagging the article with dmy template (unproblematic, except the possible violation of the editing restriction regarding whitespace edits, assuming these aren't part of stock AWB).
• E: This one is a legitimate edit-restriction violation, as it changes the whitespace inside infoboxes, and changes <references/> to {{reflist}}, neither of which AWB does, or would do in future versions.
• F: Legitimate edit restriction violation
• G: Pretty sure infobox templates are automatically capitalized by AWB (and if not, they reasonably could be without being considered disrupted). The whitespace to the infobox however, are beyond the scope of AWB genfixes.
• H: Legitimate edit restriction violation
• I: Legitimate edit restriction violation
• J: Legitimate edit restriction violation

I exhort ARBCOM to use better supporting evidence for this resolution, if it is to pass. Headbomb {talk / contribs / physics / books} 20:46, 6 May 2012 (UTC)

Thank you for this evaluation. --Dirk Beetstra ^{T C} 05:14, 7 May 2012 (UTC)

Comment on [Unblocking of SmackBot]

I want to note that admins sometimes unblock their own bot, but when they do so it's usually understood that the cause of the block was addressed, and it's usually done with the another admin's consent. Headbomb {talk / contribs / physics / books} 21:01, 6 May 2012 (UTC)

Comment on [Rich Farmbrough banned]

That seems rather extreme. The problems are related to script/bot-based editing. I see zero benefits in banning Rich from editing articles 'normally', nor in depriving ourselves of his technical expertise in discussions. Rich, for example, could do a lot of good if he could work with the AWB team to engineer and tweak additional AWB fixes. Since these fixes would be vetted and implemented by the AWB team, we would not run into the problems that led to this case. As for problems of "enforcement", why not do some WP:AGF here and trust RF to keep his word, and trust admins to have a certain level of clue. Distinguishing this behaviour, from this isn't the hardest of things to do. A ban regarding "bot-like editing, largely construed", should certainly be considered before a scorched earth remedy. Headbomb {talk / contribs / physics / books} 21:09, 6 May 2012 (UTC)

I would go with that, but as Cyberpower says above, that might drive Rich mad. You can see in what Rich says that he has "high anxiety" about tasks not getting done. I think Anomie's set of definitions are very useful here, and make it clear that you can define a level of editing that is verboten without actually needing to know which tool is in use(and as Rich appears to use some fantastically complex javascript even to make a comment on a talk page, the tool effectively is irrelevant). Elen of the Roads (talk) 10:15, 7 May 2012 (UTC)

I really doubt restricting him to edit "normally" would drive him mad. Even if it did, we're neither a kindergarten or a psychology clinic. Headbomb {talk / contribs / physics / books} 15:15, 7 May 2012 (UTC)

Grammar

"implicit indication made that automation was used to performed them." should be "implicit indication made that automation was used to perform them."--Felix Folio Secundus (talk) 08:47, 7 May 2012 (UTC)

Thanks, fixed.  Roger Davies ^talk 09:07, 7 May 2012 (UTC)

Finding 3.1.6 - unblocking own bot

Finding 3.1.6 is badly written, and ignores points made by arbs elsewhere on the page.

Administrators may not unblock their own bot if another admin has blocked it. As Special:Unblock says, "Administrators: If you or your bot have been blocked, you must not unblock yourself even if you believe the block is unfair, inappropriate, or in error. Instead, contact another administrator through e-mail, IRC, the mailing list, or by leaving a note on your talk page." (emphasis added)

What Special:Unblock (which isn't itself policy, of course) means in relation to bots is "don't unblock instead of discussing with others (especially the blocking admin)". That leaves room for the common WP:NOTBUREAUCRACY practice of a blocking admin giving permission to the bot operator to unblock after fixing the problem that was the reason for the block. By contrast, the Finding, in its first sentence, leaves no such room; it's translated into a blanket ban on unblocking your own bot. Whilst there may be a case for such a policy, it's not ArbCom's job to make it, so please don't. Rd232 ^talk 10:49, 7 May 2012 (UTC)

Seconded! --Stephan Schulz (talk) 11:38, 7 May 2012 (UTC)

Indeed, badly written. --Dirk Beetstra ^{T C} 11:41, 7 May 2012 (UTC)

Don't really know what to say...

Just gut-wrenching. Our priorities are fucked up. Jenks24 (talk) 13:26, 7 May 2012 (UTC)

Well, you must be realistic. I've been on the project for 7.5 years and have collected just short of 20000 edits. Rich has 50 times more edits. Assuming we make mistakes and enemies at the same rate, there will be 50 times more whiners with 50 times more material to sieve through for minor missteps and thing that can pe presented out of context. How can ArbCom resist a 2500 times higher rate of "evidence" for "misbehaviour"? --Stephan Schulz (talk) 13:45, 7 May 2012 (UTC)

I share, obviously, that opinion. But who cares that someone leaves, 'the reality is that - no matter how much work there is to do - none of us are indispensible'... Regarding your hidden remark: I am sorry, but maybe you (plural) should. --Dirk Beetstra ^{T C} 13:56, 7 May 2012 (UTC)

Probably pointless adding my support here, but there you go.--Gilderien Chat|List of good deeds 13:59, 7 May 2012 (UTC)

Bot operators need to run bots within consensus and within policy. It should come as no surprise that this is what happen when you don't. Headbomb {talk / contribs / physics / books} 15:11, 7 May 2012 (UTC)

Don't you find it absurd that edit such as these [2], [3], [4], [5], [6], [7], [8], [9], [10] are being used as evidence to consider banning someone from the project? Jenks24 (talk) 15:23, 7 May 2012 (UTC)

Jenks24, it is fucked up, just like you said. Don't you find it absurd that blocking the bot (generally accepted as 'an alternative account') of an opponent in a case is not even worth mentioning? Don't you find it absurd that when others run approved scripts on unapproved tasks on massive speeds (well over 100 per minute) and massive number of pages (thousands) get utterly, completely ignored (it may even be against earlier established consensus ..) by the community. Ah well, it is just pointless. --Dirk Beetstra ^{T C} 15:28, 7 May 2012 (UTC)

It is. Ridiculous. Surely then we should ban everyone who uses scripts such as advisor.js, (which I have used in the past) which is bascially for correcting formatting in accordance with WP:MOS.--Gilderien Chat|List of good deeds 15:32, 7 May 2012 (UTC)

Who runs scripts at 100EPM without BAG approval? Because I know of no one, and no bots, that edit at that speed. Headbomb {talk / contribs / physics / books} 15:38, 7 May 2012 (UTC)

Dirk just made that one up I think, out of sheer desperation. I wish he would realise that argument doesn't work even if it's true, and if it's not true, it just cuts off even more sympathy (I used to collect the rates many years ago. The number of times I was told "there are loads of people on this street who never pay a penny", well I'd be rich if I had a fiver every time someone said it. The number of times it was true was precisely zero.) Elen of the Roads (talk) 16:03, 7 May 2012 (UTC)

Ban really required?

If I understand correctly, the reasoning for the site ban stems from a combination of the difficulty in enforcement of a compromise and Rich's history of breaching his edit restriction. Now, I would like to point out that Rich was of the belief that the restriction itself was illegitimate. Maybe, MAYBE, if he was asked to make a statement acknowledging the legitimacy of an arbcom ER he could be trusted to abide by it. There may need to be a sidebar that says that even people he does not care for *cough*Fram*cough* are able to bring complaints against him without disturbing that legitimacy. I would hope that Rich would get hung for being Rich and not for what betacommand did. Let's not kid ourselves... The siteban is because BC could AND WOULD defy any restriction. Rich isn't there yet, is he? 206.47.78.150 (talk) 15:39, 7 May 2012 (UTC)

It's possible this would work. Do you think you can talk Rich into it? --Elen of the Roads (talk) 16:05, 7 May 2012 (UTC)

Not me. I'm just the peanut gallery. It needs to come from an authority he respects and it needs to be genuine. Anything less and it will end up as him and his detractors playing 'gotcha' with each other and that just ends back here. An effort to update bot policy would help. It needs to be clearer and it needs to reflect actual practices. The self unblocking thing in particular is too big a deal to be left subject to inferred interpretation like that. 206.47.78.150 (talk) 17:54, 7 May 2012 (UTC)