Wikipedia talk:Arbitration/Requests/Case/Magioladitis/Proposed decision

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

Case clerk: Amortias (Talk) Drafting arbitrator: Opabinia regalis (Talk)

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.

BU Rob13's section

Utterly insufficient, but I have little hope of convincing you of this, and I don't care to spend time arguing this. No details of enforcement for the restrictions? No restriction related to cosmetic editing outside of the AWB genfixes? No apparent provisions for the restriction to account for the fact that the community will likely never evaluate all AWB genfixes/CHECKWIKI fixes (partially due to lack of interest, partially because these are decentralized processes with no weight behind them, partially because the documentation is so poor that they're impossible to review)? Not even a reminder related to unblocking your own alternative account? No finding of facts nor remedies related to clear editing against the consensus to discontinue these comsetic-only edits at AN over a year ago? See you in three months, after issues continue and we're required to return, either due to Magioladitis or others. Not only have you not handled the situation here, but by failing to enforce the community's policies, you've empowered all those with dissenting views about how semi-automated editing should be conducted to act however they please regardless of existing policy with the comfort that they will receive, at most, a slap on the wrist. ~ Rob₁₃^Talk 01:03, 10 February 2017 (UTC)

@Bgwhite: I never asked you to file 50+ BRFAs. The Magioladitis BRFAs were largely declined due to (i) process spam, and (ii) lack of consensus. The declines came fast and heavy because he filed 25 BRFAs in a ~ 48 hour period, utterly destroying the BRFA process. The ~3 active BAG members can't keep up with that, clearly. I asked you to have a discussion with me about the different aspects of CHECKWIKI and identify any that might benefit from additional community scrutiny, not to file 50 BRFAs. You declined to have that conversation at the time. Sadly, I'm no longer able to have it, as I've largely withdrawn from administrative aspects of the project. When I last looked through the CHECKWIKI list, the number of problematic fixes were relatively small. Note that I've supported several of Magioladitis' proposed tasks out of the current set of BRFAs; I similarly support most of what your bot does. As for the desire to avoid community discussion, all I have to say is that no single editor has the right to unilaterally dictate what edits a bot may or may not make. 100% of bots on the site require consensus. Your bot is no different. ~ Rob₁₃^Talk 03:06, 10 February 2017 (UTC)

Your approvals weren't rescinded. Had you chosen to have that conversation, you would have learned quickly I was looking at less than 10 fixes which were potentially problematic and intending to ask you to send those through BRFA. Not 50. The reason Magioladitis has to do things this way and is facing so much skepticism is because the community made a determination that his bot was editing disruptively and his approvals must be revoked. Since you aren't a disruptive editor, I imagine you'll have a much better time. (I'd recommend RfCs before BRFAs because the latter need consensus.) ~ Rob₁₃^Talk 12:23, 10 February 2017 (UTC)

@Opabinia regalis: See Wikipedia:Bots/Requests for approval/Yobot 48, where Magioladitis continues to ask for approval for A but claim the approval lets him do B. I believe Ramaksoud2000's question needs a real answer, because we need a path forward when (not if) these issues recur. The normal solution would be to revoke AWB access and block the bot account. As arbitrators here haven't even deigned to consider desysopping, we're left with an admin whose AWB access cannot be revoked and who's unblocked his own bot in the past. ~ Rob₁₃^Talk 11:04, 11 February 2017 (UTC)

The threading is a little hard to read there, but if you mean "asking for approval to replace doubled http://http://www.whatever.com and then suggesting expanding to also fix http://https://www.whatever.com" and the like, then I'm not sure I see a serious problem there. Yes, the request should specify the full scope rather than adding stuff as it comes to mind, but running those as separate tasks or requiring a new review might be swinging too far in the other direction from the previous vague BRFAs.

As for the scope of the remedies, I'm personally not a fan of proposing harsh remedies if they aren't reasonably likely to pass. As I mentioned below to Ramaksoud2000, I think it's important to keep in mind that the failure mode here isn't so bad relative to other controversial wiki issues. Opabinia regalis (talk) 08:57, 12 February 2017 (UTC)

When a bot operator says "I will do A", we take that at face value during approval as the scope of the task. I invited him to clarify that he wants to do "A and B" many times, which I would also be supportive of, but he instead argued for a good long while that it's ok to say "I will do A, but not B" and then do B. That's just downright silly, and it will lead to problems. ~ Rob₁₃^Talk 16:16, 12 February 2017 (UTC)

@Euryalus: The last time I contacted an admin regarding Magioladitis' cosmetic edits, they declined to act because it's a big hassle to block an admin. And ArbCom is declining to act on the blocking policy or involved, which effectively invalidates it; a policy which isn't enforced is no policy. I hope it won't get to the point where any of this matters, and I hope if it does, those steps will be applied, but I have some serious skepticism. ~ Rob₁₃^Talk 02:37, 12 February 2017 (UTC)

I don't quite get why people see the current blocking policy as unclear. Bot accounts are alternate accounts. You are not permitted to unblock yourself, including any alternate accounts. This has (or should have) universal application, even for alternate accounts of bot-owning admins. However we'll see how the discussion progresses. -- Euryalus (talk) 00:13, 15 February 2017 (UTC)

Oh, I fully agree, and I think that's one building block (among many, including general WP:ADMINCOND) for why desysopping should be considered here. I've just stopped arguing that because this Committee seems intent on weakening policies by failing to enforce them. ~ Rob₁₃^Talk 00:34, 15 February 2017 (UTC)

Why exactly are we passing a remedy about AWB developers when there exists a full-proof option in AWB titled "Skip if genfixes only"? This skips an edit, with zero errors, if you're not applying whatever main task the bot operator has designated the bot to do. What more do you expect AWB developers to do? The problem here is that a bot was run with this option turned off. Also, no mention that one of the parties in this case is an AWB developer who works heavily on general fixes? Ok. ~ Rob₁₃^Talk 16:15, 12 February 2017 (UTC)

See Wikipedia:Bot_requests#Lt._Ayala, where a BAG member has reprimanded Magioladitis for running an unauthorized bot on his main account within the past couple days, including making several clearly cosmetic-only edits and edits with misleading edit summaries. [1] (doesn't make the fix the edit summary claims) [2] (doesn't make the fix the edit summary claims) [3] (cosmetic only edit) [4] (cosmetic only edit). Those four problematic edits are out of 93, an error rate far above that expected by any editor using AWB in semi-auto mode, where editors are supposed to look over and "Save" every single edit. I continue to disbelieve that these issues will magically go away with no sanctions, and I struggle to understand how the Committee believes the same in the face of all evidence to the contrary. ~ Rob₁₃^Talk 16:21, 13 February 2017 (UTC)

And now this. ~ Rob₁₃^Talk 04:02, 14 February 2017 (UTC)

• @Newyorkbrad: You are correct that a relevant village pump should be the place for any consensus discussions. That is where BAG members send a botop who has insufficient consensus a good 90% of the time. The remaining 10% is sending them to a relevant WikiProject if the bot task only affects one project (i.e. WikiProject tagging, etc). It would be wise to leave at least some discretion with the BAG members and just leave the restriction in the form of "Seek consensus at the Village Pump or another broad community venue, if a different venue is considered appropriate by a BAG member." ~ Rob₁₃^Talk 00:33, 19 February 2017 (UTC)

• This case never ceases to amaze. We now have a threat to block admins for "abuse of admin tools" if they block his bot. Still, no consideration for desysopping. The mind boggles. Whether the "hypothetical" was intended to have a chilling effect or is a genuine lack of understanding of WP:INVOLVED and what constitutes abuse of admin tools, it's stunning that there's no recourse even being considered here. ~ Rob₁₃^Talk 20:50, 22 February 2017 (UTC)
  • I neither personally attacked you nor called for you to be blocked, Magioladitis. ~ Rob₁₃^Talk 21:48, 22 February 2017 (UTC)
    • I have your BRFAs on my watchlist. I do the same for many BRFAs. I have this case on watchlist, as I've been involved in it from start to finish. I've responded to other editors on your talk page to defend your edits as being non-cosmetic when someone questioned them. I have every user talk page I've posted to watchlisted, since I have the setting checked to watchlist every page I edit. I have most pages related to bot policy/requests/BAG/etc. on my watchlist because I'm a bot operator. I'm not sure what more you want from me here, Magioladitis. Lately, I've been defending your edits/proposed bot tasks just as often as I've been critical of them. If you're expecting me to decline to look over what you're doing as you attempt to reform from over half a decade of cosmetic-only editing, that's not realistic. Trust, but verify has an important second step. ~ Rob₁₃^Talk 02:54, 23 February 2017 (UTC)

• The assertion by Magioladitis that ArbCom should create policy by fiat by "clarifying" (or, less generously, "dictating") that a change in watchlist behavior results in a policy being thrown out must be wholly rejected. The watchlist behavior is not the majority reason why most editors are opposed to these edits, and only the community can change policy or the AWB Rules of Use. ~ Rob₁₃^Talk 04:29, 25 February 2017 (UTC)

• @Doug Weller and Opabinia regalia: Please don't send people to BRFA for a consensus discussion. BRFA is for technical approval. Consensus discussions are held elsewhere. They always have been. Typically, if anyone objects to a bot task at a BRFA, it's declined or placed on hold and sent elsewhere for consensus seeking. We can't hold an effective RfC within a BRFA. Trying to mash together the consensus and technical discussions into one nightmarish chimera will make it near impossible to approve any tasks, and that's obviously not the goal here. AN was suggested as the best place to reach a wide audience (not just technical editors). BRFA itself for consensus seeking is worse than no remedy at all. ~ Rob₁₃^Talk 09:51, 26 February 2017 (UTC)
  • @Opabinia regalis: Fix ping. ~ Rob₁₃^Talk 09:52, 26 February 2017 (UTC)
  • @Headbomb: I'm not sure you've read the consensus remedy the same way I did. As the title is written, it appears to attempt to seek actual consensus following a discussion from the community that the edits in question should be done and should be done by a bot. That seems to be at odds with the later wording that suggests just a courtesy notice. Would your opinion on venue change if the remedy was worded to require consensus seeking rather than a simple notice? ~ Rob₁₃^Talk 09:58, 26 February 2017 (UTC)
    • @Headbomb: If "seek broad input" is what ArbCom is looking for, I have no problem with that. I objected to the original wording of the title, which was "seek consensus", being held at BRFA. BRFAs assess the consensus for a task. They don't generally involve broad consensus seeking at BRFA, though, which is what I was fearing ArbCom was trying to do. ~ Rob₁₃^Talk 17:20, 26 February 2017 (UTC)
• When determining enforcement, it's important to take into account that much of the problem here is the party's attitude toward testing. Before I implement a change to any of my bot tasks, I rigorously test them on many different types of pages in semi-automatic mode, where I can catch errors, or on my own account with semi-auto edits. It's a common and expected best practice to do this sort of testing before pushing any changes "live". It seems clear this type of rigor hasn't been applied to some changes on the Yobot account. No, we shouldn't punish anyone for a simple bug with escalating blocks. But yes, there should be consequences or additional oversight if a bot operator doesn't demonstrate sufficient rigor in testing their bots. Personally, for the bot business, I'd prefer an alternative to blocking the operator altogether. Why block an editor when they could be working to fix the task? Instead, I would suggest that any uninvolved admin may block Yobot as an arbitration enforcement action upon noticing any cosmetic-only actions, with the block only to be lifted upon the issue being fixed and the blocking administrator (or a consensus of uninvolved admins at AE) giving their go ahead. If the same type of error is made a second time, then the BRFA authorization may be revoked by any uninvolved administrator as an arbitration enforcement action. This would force more thorough oversight, testing, and trials through the usual BRFA process to weed out a persistent error. The major benefit is no blocks of Magioladitis whatsoever for his bot's actions, addressing Opabinia regalis' concerns. The penalty of having to go through the admittedly annoying/slow BRFA process would encourage testing out changes pre-emptively on the operator's own initiative, providing the "backbone" DeltaQuad sees as necessary. ~ Rob₁₃^Talk 00:06, 28 February 2017 (UTC)

@Ramaksoud2000: We have the BRFA process and BAG for that. Scrutiny is high, so if problems recur, they can be dealt with. Another remedy dealing with bots wouldn't be remiss, but this is still a great improvement from the original toothless PD. ~ Rob₁₃^Talk 22:07, 1 March 2017 (UTC)

@Ks0stm: There are no current Yobot tasks being re-reviewed. The tasks' authorizations were entirely revoked and cannot run (i.e. former tasks), and some of them have been submitted for new BRFAs. If you meant to exclude current BRFAs from suddenly needing broad input, which would probably be wise to prevent a flood of noticeboard notifications, then the best wording would be "With the exception of bot tasks under review when this remedy is enacted, ...". As written, nothing would be excluded, based on how I'm reading it. ~ Rob₁₃^Talk 11:57, 2 March 2017 (UTC)

@Opabinia regalis: That would read "failures or misuse of the stop feature" which isn't what you're going for. I'd add "recurring issues" to the comma-separated list instead. ~ Rob₁₃^Talk 02:10, 15 March 2017 (UTC)

Actually that is what I was thinking, as it's clearly a misuse of the purpose of your bot's stop feature to ignore a good-faith stop and simply restart the bot without addressing the issue. I don't love "recurring" and so forth in an edit that's meant to stopgap possible future wikilawyering because that opens space for still more of it. But I think we're hitting diminishing returns on fine-tuning the wording. Opabinia regalis (talk) 05:32, 16 March 2017 (UTC)

@Opabinia regalis: When I see "misuse of the stop feature", that sounds like misuse on the part of the person who stopped the bot, not on the operator for restarting the bot. The operator doesn't actually do anything to restart the bot except look at the message and hit start; they're never actually doing anything with the stop feature itself. ~ Rob₁₃^Talk 13:05, 16 March 2017 (UTC)

Maybe, but we've already had the inevitable ARCA request in miniature here and Euryalus has already made a clear statement of the intended interpretation. I'm not trying to be dismissive of your suggestion or Amanda's concerns, but I think there's more value at this point in closing the case and letting everyone move on with their wiki-lives than there is in additional wording-polishing. (My impression from past cases is that last-minute tweaks and clarifications often introduce as many points of ambiguity as they resolve anyway, just by the inevitable fact that they're available for review for a shorter period.) Opabinia regalis (talk) 22:45, 16 March 2017 (UTC)

Bgwhite's section

So, before the case is over....

1. Magioladitis has yet to get a BRFA thru with ~25 denied or in progress.
2. Bots we use are being denied to fix any issue with "general fixes" is on.
3. Obvious CheckWiki errors are already being denied such as adding http:// (google.com vs http://google.com) and DEFAULTSORT fixes.
4. Anybody can now call for an RFC or a village pump discussion on each individual CheckWiki error as the decision says "soliciting broader community input". I get todo 70+ RFC's in which non-cosmetic CheckWiki fixes will be denied.
5. BU Rob13 already wants me to go thru the same BRFAs in which everything is being denied. I get todo 50+ BRFAs in which they will be denied.

Sorry BU Rob13, you are already getting everything you wanted without the ARB decision. Magioladitis can't run a bot. I won't be able to run a bot. CheckWiki errors are being castrated. How can we be here in three months when we can't do anything? Now I get to be tied down in 50+ BRFA's and 70+ RFC's for months. I've been saying it all along to you and Spinningspark, even though the case isn't about CheckWiki or me, we will be kicked out of Wikipedia too. I'm at the door waiting to be shoved. Bgwhite (talk) 02:59, 10 February 2017 (UTC)

@BU Rob13: You asked me to redo my BRFA process. I said after the ARB case was over when the lay of the land will be known. Magioladitis is currently redoing his BRFA process and every CheckWiki error is required to be a separate BRFA. Therefore, I will have to do the same and have 50+ BRFAs. Your "withdrawn from administrative aspects" causes you not to help me, but it has not stopped you from making comments in almost all of Magioladitis' BRFAs. You've supported 4 of the ~25 BRFAs out of 4 total supports that have been given by everybody in all the BRFAs. People are saying no to any BRFA that requires genfixes and no to some non-cosmetic CheckWiki fixes. My AWB bot uses genfixes to fix the issues.

Remember when I said nobody has asked me a CheckWiki question? It's still that way. In the Arbs rush to vilify CheckWiki, they didn't know it's on 30+ wikis with 25+ people involved and not a small technical group. The last error expanded came from the WMF. Arbs don't know what "management practices" CheckWiki has, but now I have to change it and do RFCs for every error that's added.

I'm not avoiding community discussion. There's a difference between avoiding and not willing to go thru 50+ BRFAs in the environment Magioladitis is doing and going through 70+ RFCs. It takes 2-4 weeks for an individual BRFA, Magioladitis can only submit 5 at a time, therefore 5-10 months. Would you really want to go thru all that, plus have a hostile crowd? As you have withdrawn from administrative aspects, I'd say your answer is no too. I'd think you would also agree that one's sanity is more important than going thru that for months on end. Bgwhite (talk) 07:24, 10 February 2017 (UTC)

I don't see why this decision should be read as requiring existing bots without the history of problems Yobot has had to be re-authorized. If the community or BAG separately wants to do that, they could, but this case is called "Magioladitis", not "Bots that annoy people sometimes" ;)

As for checkwiki errors, I don't see that there's a call for RfCs of every individual error. Assuming you mean remedy 1, the suggested approach there is to classify and separate purely cosmetic error categories so that those errors are not selected for fixing unaccompanied by other changes. Asking for broader review of that classification hardly seems onerous. Accessibility fixes are by definition not "cosmetic-only", so these relatively high-priority fixes shouldn't be affected. Opabinia regalis (talk) 22:44, 10 February 2017 (UTC)

In passing, Proposed Remedy 1 is also only "an encouragement." Actual outcomes would be a matter for the community. -- Euryalus (talk) 00:29, 11 February 2017 (UTC)

MSGJ's section

@Opabinia regalis: Minor comment: FOF 10 is not a finding of fact but is written as a principle. — Martin (MSGJ · talk) 09:35, 10 February 2017 (UTC)

Echoing BU Rob13's comments above that the proposed remedies will not solve any of the problems and will likely make worse. Hopefully some other arbiters will come up with something stronger! — Martin (MSGJ · talk) 14:13, 10 February 2017 (UTC)

@MSGJ: The question was raised before I posted this whether FoF 10 was necessary at all, and I could still go either way on that, but the purpose is to make note of the fact that the concerns have actual practical consequences, rather than purely being an issue of technical violations. Opabinia regalis (talk) 22:42, 10 February 2017 (UTC)

47.222.203.135's section

• Remedy#6 — "Magioladitis reminded to refrain from cosmetic edits. Magioladitis has begun the process of resubmitting BRFAs... He is reminded to use caution when performing future edits..."

• What is the definition of cosmetic? WP:COSMETICBOT? Outcome of the "Community encouraged to review policy on cosmetic edits" discussions? And is this a sanction, or a reminder? Making cosmetic edits *is* permissible, as long as one also makes a substantive edit simultaneously, is my understanding. Is this a sanction that prevents Magioladitis (and his bots) from making cosmetic edits under any circumstances? Or just a reminder to be cautious? Whichever it is, please clarify the wording. 47.222.203.135 (talk) 14:57, 10 February 2017 (UTC)
  • Fair point, this should really say "cosmetic-only". And yeah, the outcome of a community discussion would be the best definition, but as a matter of practice it would be best to use the most expansive of the common definitions until a better one is formulated. Opabinia regalis (talk) 22:42, 10 February 2017 (UTC)

• Remedy#7 — "Magioladitis restricted from making general fixes or checkwiki edits. ...temporarily restricted from making any [genfix or checkwiki]... until such time as the community discussions recommended in the preceding sections conclude, except as explicitly authorized by an approved [brfa]."

• WP:NOTFINISHED suggests that discussions will never conclude. Better either set a specific time-limit for which Magioladitis is temporarily restricted, i.e. "for six months" or "indef" or "for six days" or whatever is considered appropriate, or be specific about which discussions are being referenced and what counts as them being concluded, i.e. "until the Feb'17 RfC about WP:COSMETICBOT is closed by an uninvolved admin" and whatever other specific discussions are being referenced. In my experience discussions are never really concluded, so again I would ask the arbs please clarify this decision-language.
• Also, side point, please ALWAYS say genfix (or employ italics/scarequotes/etc) in the proposed decision rather than saying general fixes, so that nobody is confused by the lingo to come to the mistaken belief that "fixes in general" are what is begin discussed here, when genfix is the actually-intended meaning. 47.222.203.135 (talk) 14:57, 10 February 2017 (UTC)
  • Heh, the original version of that proposal had a time limit and I changed my mind, figuring that allowing such edits only where an independent review had occurred was a reasonable way to provide specific feedback to Magioladitis. A time limit doesn't have any provision for feedback. Opabinia regalis (talk) 22:42, 10 February 2017 (UTC)

• My congratulations to the arbs for not giving in to the temptation to solve every problem with a WP:BANHAMMER, and for doing your best to send these complicated discussions back to the community for further RfCs and such. Sometimes the only winning move is not to play, as the old saying goes, or in the Americanized metaphorical sports-equivalent, sometimes punting is exactly the right thing to do. But I'm slightly dismayed that the arbcase has not explored the question of tool abuse more carefully: is it the sense of the arbs, that blocking a malfunctioning bot, when a simple edit of the bot's talkpage or a click of a non-admin-dev-provided button on the bot's userpage could also stop the bot... does such a block, constitute tool abuse by the blocking admin? (I would hope not... but the proposed decision is silent on this matter.) Is it the sense of the arbs, that when an admin in the good faith belief that the underlying cause of said bot being blocked, has been corrected, goes ahead and unblocks the bot, that this is wheel-warring? (I personally think not but apparently an RfC is needed! So I understand why the arbs may remain silent on this particular question in the proposed decision.)
• But here is the thing which was left unsaid, that I wish the arbs would consider saying: In a situation where a bot is blocked by admin X, and that selfsame bot is later unblocked by admin Y, does it constitute wheel-warring for admin X to immediately block that selfsame bot again? To me this answer is 100% clearly "no it is fine". My understanding was that the reason for arbcom's accepting the case, was to try and help sort out the broader questions, but that the particular question arbcom was "supposed" to answer, was whether or not there was any tool abuse. I tend to side with Rich_Farmbrough in the belief that unblocking one's own bot, when one believes the software bug has been corrected, is 100% fine and correct. And I believe at least one arb agrees with that view, i.e. that there is no need for uninvolved admin Z to perform the actual unblock, since only the dev can know for sure whether they believe in good faith the bug has been squashed and unblocking is now in order. Further discussion is already ensuing, on this point.
• What about admin X, though? In the hypothetical case where admin Y is unblocking without real consensus, and thus is letting a rogue bot free again, can admin X re-block with impunity? I think the answer is clearly "absolutely as long as the initial block and any subsequent re-blocks were performed in good faith rather than in anger". Similarly, I think that admin Z can re-block the bot, with impunity, just as much as admin X can do so, if they (admin Z) in good faith believe that the bot is malfunctioning, even if the bot in question was unblocked 30 seconds ago by admin Y with the edit-summary "everything is fine now DO NOT re-block this bot". Does this make sense? So my request, specifically....

• Can the arbs please say, something a bit more concrete, about what the wheel-warring policy is (addition of new principle #3.1.6 perhaps summarizing WP:MOPFIGHT),
• ...this portion of my request retracted, I'm really more interested in hypothetical admin Z here and then say whether wheel-warring did or did not occur, in the 19 specific cases where Yobot was blocked (expansion of existing FoF #3.2.2.4 perhaps),
• ...plus ideally add a small but methinks-crucial amount of prose to remedy#4, specifically: "...the blocking[,] and unblocking[,] and re-blocking of problematic bots..."

• Thanks 47.222.203.135 (talk) 14:57, 10 February 2017 (UTC)
  • One of the long-standing challenges with the arb-decision format is working out how to document that an issue has been considered but ultimately wasn't judged to be problematic enough to appear in the decision. It is the sense of this arb - and, I think, broadly representative of our discussion of the case - that the matter of unblocking your own bot is not settled in policy and is not so self-evidently wrong that it deserves to be called out in the decision. In the previous case that dealt with the issue - the Rich Farmbrough case - there was quite a bit of disagreement over whether the bot-unblocking was a serious issue, and the finding was ultimately vacated, and there's been little reason to consider the issue since. My own view is that a problem involving two people separated by five years is not really one that demands a policy be created to address it, but that's just my preference for policy minimalism. The community could of course discuss the matter and conclude that judging on a case-by-case basis works fine, or it could develop some guidance on the subject, as is being discussed here. I don't really think that analyzing such things in terms of "wheel warring" is likely to be helpful; that just introduces charged terminology without adding any clarity. Opabinia regalis (talk) 22:42, 10 February 2017 (UTC)
    • Appreciate your responses, thanks. I agree that use of wheel-warring terminology, as a descriptor of the events in this specific arbcom case at least, could be seen as inflammatory... the practice of unblocking one's own bot was de facto pretty common, but not as clearly backed by consensus as I would have thought. I'll be watching the broader discussion. But my I'm not trying to corner the arbs into taking a stance on whether unblocking one's own bot is-or-ain't considered to be a wheelwar... I'm trying to corner the arbs <grin> into saying whether or not it would be seen as wheel-warring in a hypothetical case, vis: admin X blocks bot, admin Y unblocks own bot, admin Z re-blocks same bot three minutes later.
    • Is admin Z in the clear, if they in good faith believed the bot was still malfunctioning? Or can admin Y file an arbcom case against admin Z, for wheel-warring? (Compare and contrast with the scenario where some fourth *human* editor is being blocked/unblocked/reblocked, rather than a bot-account being blocked/unblocked/reblocked.)
    • Because my understanding is that the gist of this whole arbcom case, was #1) some people firmly believed unblocking one's own bot was wheel-warring, but also more subtly #2) various people also believed that any re-blocking of selfsame bot, would constitute wheel-warring. It is okay if this is out of scope, or something considered but not added, or whatever. My argument for mentioning the re-blocking scenario, is that I think the perceived power-imbalance, is simply that when admin Y unblocks their own bot, any hypothetical admin Z may well be fearful of re-blocking, lest the arbs be angered. *Would* the arbs be angered at good-faith admin Z, and accept a tool-abuse case against Z, filed by Y or filed by some uninvolved party? If not, might be worth explicitly saying so, to lessen the perceived power-imbalance that unblocking one's own bot is perceived to create.
    • The answer I seek, may also be applicable to Ramaksoud2000's what-if question, below: in August 2017, if admin X is named Opabinia and blocks Yobot for unauthorized-template-bypass-surgery-buggy-ness, and admin Y is named Magioladitis and unblocks (presuming the RfC outcome permits such!), can some hypothetical admin Z re-block Yobot three minutes later with edit summary "please discuss further"? Or is admin Z wheel-warring? 47.222.203.135 (talk) 10:59, 11 February 2017 (UTC)
      • We can hardly pre-commit all future arbs to our interpretation even if you succeed in cornering the current set ;) This is a good example of why I don't think talking about "wheel warring" adds anything useful: presumably the reason you want to know whether Z is "wheel warring" is that someone whose actions are classified as such might be desysopped. (And surely we'd have people here insisting that Policy Says So, So It Must Be Done.) But that's not nearly as useful as finding out why Z is doing what they're doing - because they genuinely believed Y hadn't fixed the problem? they were trying to adjust the block settings and had an "edit" conflict? they were making a point (or a WP:POINT) about unblocking your own bot? they just wanted to stick it to Y? If the first reason, was Z's belief well-founded? If so, were they right? And so on.... You get the point. In most cases I don't think it makes sense to pre-judge the hypothetical behavior of hypothetical people and then try to map a real and much messier situation back onto the hypothetical. Opabinia regalis (talk) 08:57, 12 February 2017 (UTC)

Admin Z would not be wheel-warring if they reasonably believed that further discussion was needed. They are, after all, blocking for a different reason. Broken -> Fixed -> Discussion needed. It might be questioned whether blocking is necessary in either case. Despite the two schools of thought (that (un)blocking bots is trivial or that is is serious business) it seems unwise, unhelpful and unkind to block where an alternative mechanism is available.

All the best: Rich Farmbrough, 10:52, 12 February 2017 (UTC).

Ramaksoud2000's section

One question: What should occur when Magioladitis runs bots/bot tasks without approval again? Ramaksoud2000 ^{(Talk to me)} 15:33, 10 February 2017 (UTC)

A recurring point in this dispute is that Magioladitis appears to have held the good-faith belief that his tasks were approved even while others felt the BRFAs should have been interpreted more restrictively (most clearly demonstrated in the discussion about the list of, and definitions of, the checkwiki errors listed in task 16). Since the problem is more closely related to vagueness than to deliberate flouting of a well-defined bot task scope, the problem will hopefully be addressed by more specific approvals following requests that get broader review. Opabinia regalis (talk) 22:42, 10 February 2017 (UTC)

Respectfully, this answer avoids the question. We may disagree on the cause of the issue, but what should happen if/when he runs bots on his main account or bypasses template redirects again? Ramaksoud2000 ^{(Talk to me)} 23:50, 10 February 2017 (UTC)

Rely on community policies and procedures regarding user conduct such as dispute resolution, noticeboards, and if necessary, community or administrative actions. ArbCom should not replace these community mechanisms. If calling to attention deficiencies in process and policies, and possibly other remedies, will re-equip the community to successfully handle future incidents via these venues, then that's arguably the best way forward. Mkdw ^talk 23:34, 11 February 2017 (UTC)

Policy already prohibits running bots on main accounts. A bot acting out of scope or disruptively should be blocked until the issue is resolved. Policy also already prohibits admins from unblocking their own accounts, and (at least in my view) this should be applied re admin bot-owners as it is with admins of any other kind. -- Euryalus (talk) 00:30, 12 February 2017 (UTC)

Thank you for your response. Ramaksoud2000 ^{(Talk to me)} 23:59, 11 February 2017 (UTC)

Not so much "avoiding the question" as "questioning the premise". But Mkdw is right, I don't see any need here to specify non-standard enforcement processes and think the community is perfectly capable of responding proportionately. I do hope that problems are handled promptly, but we don't see people monitoring Magioladitis' contribs after the case closes to pounce on occasional, minor errors. One thing we need to keep in mind is that, while it's clear that a lot of people are frustrated with this situation, a failure mode that consists of making "cosmetic" edits is in the scheme of things pretty benign. There's no BLP violations, no sneaky spam, no source misuse or POV-pushing, nothing that would directly affect the product we show readers. There's also no harassment or abuse or even incivility. And to the extent that cosmetic edits might cause some of those things to slip by undetected, there's already a proposed technical solution just waiting on phabricator for more comments. Something keeps happening that's a nuisance, yes, and the community should take steps where needed to minimize it, and we appreciate the efforts of the people who have encountered the problem, but we should still be careful to keep things in perspective. Opabinia regalis (talk) 08:57, 12 February 2017 (UTC)

It seems like you are saying that operating outside of BRFA approval in a manner that does not seriously harm the encyclopedia is not considered serious enough for Arbcom. I won't act like I know more than you about what Arbcom should take up. You also seem to be saying that the community can handle this on its own. Again, I won't claim to know more than you about this matter. However, if both of these are true, why was the case accepted? Each person presented evidence just to support their initial statement. No surprises came up. It sure would have saved everyone a lot of time if the case was declined from the beginning. Ramaksoud2000 ^{(Talk to me)} 21:57, 13 February 2017 (UTC)

Fair question. I did vote to accept the case, after all :) A few points:

• A case is in effect an investigative process. The decision to open one means there's enough in the request to warrant further investigation, not that sanctions are a foregone conclusion.
• In my opinion, the problem looked a bit different after the evidence was in than it did at the request stage. Substantive allegations of admin tool misuse are worth investigating, but speaking for myself (I think, though, broadly in agreement with the sense of the committee in discussions prior to the PD) those allegations proved weaker than they first seemed. On the other hand, the broader difficulties in defining and recognizing the specific type of "cosmetic-only edit" error Magioladitis makes most frequently became a recurring theme. I think it was also recognized by some commenters even at the request stage that the case had potential for precedent-setting in the area of bot operation. While we're not really a "precedent" kind of group, it's hard to avoid the fact that we do de facto have precedents. IMO, in a case that has that kind of potential impact, it's worth trying conservative remedies first. I think the discussions we're recommending will provide the community with better tools and more leverage to manage this and related situations.
• Having already collected evidence and spent some time discussing the nature of the problem means that the community has already made some headway on discussing the broader issues. In fact, those were motivated in part by Headbomb's workshop post indicating that preparation work for similar discussions was already underway. Sometimes, I think, the primary value of an arbcom case is in the provision of a structured forum in which interested editors can collect evidence and share possible solutions - the influence of the actual arbs is smaller than we'd sometimes like to think :)
• I'm worried that some people following this case are getting too focused on the narrow question of "is Magioladitis obeying the rules". It's true that he's not doing a great job of it*, but he's also failing in a way that is nondestructive. Harsh personal sanctions for nondestructive behavior would prioritize rule-following as an end in itself over addressing the practical concerns that motivated the establishment of the rules in the first place. I've always been pretty squishy on sanctions for behavior that isn't directly harmful, and pretty skeptical of enforcing rules for the sake of making sure the rules are enforced. That's not meant to be a judgment of whether it's "serious enough", but rather a desire to stay focused on what's practical. Opabinia regalis (talk) 00:49, 14 February 2017 (UTC)

  *Magioladitis, don't take that as an endorsement or an encouragement to continue what you're doing, because you shouldn't. As I said yesterday, you need to slow down and watch what you're doing, because you are still making too many mistakes that should be easy to avoid.

Thanks for the explanation. Ramaksoud2000 ^{(Talk to me)} 01:29, 14 February 2017 (UTC)

Remedies 7.1 and 10 adequately address my concern expressed above, except that 7.1 should be amended to include both semi-automated and automated actions. Ramaksoud2000 ^{(Talk to me)} 19:39, 1 March 2017 (UTC)

Section for User:CBM

Perhaps relevant to the decision: Magioladitis today made a series of edits under AWB with an edit summary which was usually nothing more than "fixing stuff using AWB ([5]). In one case, in the 3 seconds after 18:37:13 [6], he both edited a comment on a BRFA [7] and approved an AWB edit [8] simultaneously at 18:37:16. Magioladitis says the edits were saved manually, and I believe him, but from my perspective the edit rate was so high that manual review would have been challenging. For example, there are comments written in much less time than would be expected, for example between 23:33:28 [9] and 23:33:31 [10]. It is not clear how someone could write and save a comment in 3 seconds after approving an AWB edit, if they had just given full attention to the AWB edit. Moreover, many of the edits appear to be Yobot tasks that have not yet been reapproved, such as Wikipedia:Bots/Requests_for_approval/Yobot_34. My point here is that the proposed decision should approach AWB use as something that needs to be carefuly delineated. — Carl (CBM · talk) 23:58, 12 February 2017 (UTC)

To be honest, I think it's more useful to worry about the content of the edits than the timestamps, unless the rate or volume is itself a problem. Opening separate tabs to comment in multiple threads and then saving the comments rapidly is pretty common.

Part of the reason for a review process for bots is the fact that some edits are fine for humans to do in small numbers, but require more careful review when done at scale. On that premise, there's nothing inherently wrong with doing a task manually and separately filing a BRFA for the same type of edit. That said, Magioladitis, it is not a good idea for you to do this now. And as I said in the thread on my talk, these edit summaries are insufficient, and I think that should by this point be obvious to you. You need to be much more careful. Opabinia regalis (talk) 07:11, 13 February 2017 (UTC)

Recently, Magioladitis made additional series of cosmetic edits from his main account, such as [11] today. He is also making series of edits from his main account for bot tasks that were denied, such as this BRFA [12] and 500 edits [13]. I believe that a remedy such as "Magioladitis restricted" may be the only way to affect this editing pattern. The BRFA process can only be effective if, when a task is denied, the user does not proceed to run the task on their main account.

Another way to remedy these series of large numbers of edits on the main account - which should be run on a bot account - would be to limit the total edits per day from the main account, perhaps to 50 edits in the main namespace with no limits in other namespaces, so that it would be necessary to respect the bot approval (and denial) process. — Carl (CBM · talk) 12:58, 27 February 2017 (UTC)

NativeForeigner's Section

Agree with NYB's comments on Remedy 5. I find remedy 8 to be rather toothless and redundant to best practice. A bit odd for a remedy, in my opinion. NativeForeigner ^Talk 08:25, 13 February 2017 (UTC)

RF's section

I find it very unfortunate that statements like this "What should occur when Magioladitis runs bots/bot tasks without approval again?" are made.

It is not here a proposed FoF that Magioladitis has done that. Even if it were the wording "when" rather than "if" smacks of WP:ABF to say the least.

This is begging the question and poisoning the well in one fell swoop.

All the best: Rich Farmbrough, 20:04, 13 February 2017 (UTC).

Oh and reverting another editor's contributions to a discussion without informing them is also a sign of WP:NOTHERE. All the best: Rich Farmbrough, 20:08, 13 February 2017 (UTC).

Anomie - every unblock I made was completely "clean", with one exception, where I made a simple mistake, relying on what BAG member had said elsewhere. All the best: Rich Farmbrough, 17:42, 18 February 2017 (UTC).

Anomie's Section

(annoying method of commenting you have here)

@Opabinia regalis: Re WP:TFD/O, since I tend to take a strict line on keeping to what was actually approved in my bots' BRFAs, I would have insisted on submitting a new BRFA for that if I hadn't had the foresight to include "Other tasks affecting only WP:TFD and subpages as determined by consensus at WT:TFD." in AnomieBOT 59 and the good reputation to be trusted with such vagueness. Anomie⚔ 03:50, 14 February 2017 (UTC)

Also, a general comment regarding unblocking of bots (which I've previously posted at WT:BOTPOL) since I see that getting some discussion here: I note that Wikipedia:Blocking policy#Temporary circumstances blocks specifically mentions "Blocks of unapproved or malfunctioning bots should be undone once the bots gain approval or are repaired" without restriction against an admin-operator doing the unblocking of their own bot. Wikipedia:Blocking policy#Unblocking does say that unblocking yourself, which IMO includes your own alternative account, is "almost never" acceptable, but an unblock of your own truly repaired bot could easily be seen as falling under the "almost". The conflict here, and in Wikipedia:Arbitration/Requests/Case/Rich Farmbrough for that matter, largely seems to come down to repeated insufficiency of the repairs (as Newyorkbrad also observes). Anomie⚔ 03:50, 14 February 2017 (UTC)

Magioladitis' section

There is also the case that the blocker is OK with me unblocking. And also I wonder what happens to the admin that blocked an AWB bot because they don't know they are other ways to stop it. In fact they bot policy suggests that the bot should have not been blocked. -- Magioladitis (talk) 12:32, 22 February 2017 (UTC)

I also would like to know why an admin to block only the bot account and not the main account to prevent any unblocking and there is the need of remedy for this. The thing is clear: If the block was done only to stop the bot the block should not have been done. If the block was done for other reasons then all accounts should be blocked. -- Magioladitis (talk) 17:08, 22 February 2017 (UTC)

I also would like to know what do you we do in the following case: An admin blocks me bot, I block them for abusing their tools. How do we resolve from there. -- Magioladitis (talk) 17:13, 22 February 2017 (UTC)

Magioladitis, if you block another administrator for blocking your bot, you will be desysopped. Kirill Lokshin (talk) 18:58, 22 February 2017 (UTC)

Kirill Lokshin I maybe a little pointy here but it was never explained why an admin can block a bot without blocking the bot owner. -- Magioladitis (talk) 19:27, 22 February 2017 (UTC)

That's basically wheel warring, and pointy or not, you should know that leads to a near-immediate desysop. Ks0stm ^{(T•C•G•E)} 23:57, 22 February 2017 (UTC)

Ks0stm True. I ask what happens to the other admin that incorrectly blocked only one of the accounts instead of blocking all the accounts. -- Magioladitis (talk) 00:01, 23 February 2017 (UTC)

Blocking just the bot account when the bot account is not working properly is, from what I understand, standard procedure. If you'd like your main account blocked when your bot account is blocked you technically have the tools to do that yourself, or else I would recommend trying Category:Wikipedia administrators willing to consider placing self-requested blocks. Ks0stm ^{(T•C•G•E)} 00:17, 23 February 2017 (UTC)

Ks0stm OK. So when a block is technical, admins could be allowed to unblock their own bots after they fix the technical issue. Am I right? Moreover, since AWB bots allow themselves to be stopped by just leaving a message to their talk page, blocks are just an unnecessary drama. Am I right on that? -- Magioladitis (talk) 01:41, 23 February 2017 (UTC)

• Magioladitis, you getting the permission of the blocking admin has failed before, as there you ask if you can unblock, Materialscientist says they would prefer you not to, and you go ahead and unblock it anyway, post admitting that there are potential solutions to the problem at hand, but seemingly don't do anything about it. So no, I don't trust you to unblock your own bot after the blocking admin authorizes it.
• As for other ways to block the bot, you can inform the blocking admin of that, and if they believe that is appropriate, they can unblock and stop the specific task. The task should try to be stopped first, but when it takes you three days to reply to it and the bot is still running, that is extremely problematic.
• Also causing the drama of blocking the main account is not necessary in the slightest, it would add to the stacking pile of drama this bot has caused. There is zero policy requirement to do so, and contravenes WP:PREVENTATIVE which you should be well aware of as an administrator.
• Admins should be allowed to unblock their own bots after they fix the technical issue, yes. But your history shows that the same issue still keeps poping up, which is abuse of that privilege, which is why I proposed taking it away.
• I am appalled by your comment that you would block someone for abusing their tools if they blocked your bot. That is so stupidly bright line that I absolutely agree, a desysop would be immediate. It's a clear violation of WP:INVOLVED, which is linked or mentioned around 20 times on the workshop, so unless you didn't read the workshop or the basic administrative policies of the project...you should know what it means.
• Also going to propose a change in policy because you don't get a response here within 24 hours, is unneeded and disruptive. I am really starting to get the feeling that you are smoke screening a lot, and attempting to distract people from the issue at hand. There is only a certain tolerance I have for that, and Arbitrators are watching. Remember conduct during a case is also fair game in regards to remedies. -- Amanda (aka DQ) 07:30, 23 February 2017 (UTC)

It was almost certain that BU Rob13 will continue the personal attacks. Now they want me blocked for asking questions. My question is still on: If any editor used their bot to do something they should not, should or should not all accounts be blocked at once? What does the action of blocking only one of the accounts in question mean? The answer ofcourse is that sometimes the bot account is only blocked for malfunction this is the reason the bot owner can, under certain circumstances to unblock. If we are about to add additional bureaucracy then we should be very careful about the wording and the implementation. From my perspective the case that I block my own bot is one case. This was the case when xeno had to block the bot because the talk page trick did not work. But, I would like an answer to the case that an admin uses the block button to only stop the bot. -- Magioladitis (talk) 21:38, 22 February 2017 (UTC)

I note once more that BU Rob13 comments in every page I leave a message and even replies to other in my talk page. -- Magioladitis (talk) 01:56, 23 February 2017 (UTC)

Since my questions failed to make my point. I left a comment here. -- Magioladitis (talk) 00:09, 23 February 2017 (UTC)

• My mainly concern is that other bot owners do not get in trouble after this ArbCom. DeltaSquad's comment that "Admins should be allowed to unblock their own bots after they fix the technical issue, yes" is very important and in a good direction. This should not be forgotten after the this case is closed. I suugest that this should become part of the decision in the part where the principles are written or somewhere else.
• I still believe that the block button is overused against bots and I would like to see some change on that. It's not the first time I suggested something like that. My comment in the WP:BLOCK is only a continuation of Wikipedia_talk:Bot_policy#Changes_to_.27dealing_with_issues.27_section.

-- Magioladitis (talk) 07:51, 23 February 2017 (UTC)

So let's get this straight: I would like that the decision is written in a way that wont encourage unnecessary blocking against bots, it won't lead to use this case against other bot owners, it won't imply that unblock policy needs to change.

I would like also to underline the asymmetry here. Any bot can be blocked without any warning even if the bot allows other ways to be stopped. So if an admin goes and blocks an editor without final warning there is a problem but if an admin does the same to a bot there is no problem.

Magioladitis (talk) 10:53, 23 February 2017 (UTC)

Let me be clear that while I made the comment that bot owners normally should be able to unblock, in response to your questions/comment, this will not be the case for you with the now passing restriction.

Also other bot owners will not get in trouble for violating the restriction, as it is only for you. There is no need for any policy clarifications, but the community can on their own start that discussion. We can't prohibit that. But i've opposed "Community encouraged to develop policy on bot unblocking" and it's currently failing with the majority of the votes in the oppose section. If your worried about particular votes, you would have to speak to those arbs.

I will not be proposing any principle on "Admins should be allowed to unblock their own bots after they fix the technical issue, yes" because that is my personal opinion and is not backed by consensus. ArbCom can not create policy.

Also bots that we have had onwiki do not have their own mind and decision making process that is independent of their orders to do a task. Warning them would have zero effect. Warning their bot owners may have a positive affect, but as I outlined above, that's clearly not the case with you. So the asymmetry will exist. -- Amanda (aka DQ) 22:47, 23 February 2017 (UTC)

DeltaQuad That's good to hear. I already left a comment in the Workshop about my case. So, I am OK with this restriction. I think the only real change that I would like to see, but this is not part of this ArbCom case, is that the community realises that blocking a bot, when there are other ways to stop a task, is general not a good idea. Thanks for the reply.

On the other issue of the "cosmetic" edits you can see that people still debate on which edits should be done or not. Even the concept of people endorsing or opposing some of the CHECKWIKI tasks shows this. But I a not going to pursuit further. It's been more than a month that we discuss this issue. -- Magioladitis (talk) 23:25, 23 February 2017 (UTC)

Watchlists

User:Evad37/Watchlist-hideAWB.js provides a script to hide AWB edits. Is this the wanted thing and we are all OK now? -- Magioladitis (talk) 19:28, 22 February 2017 (UTC)

I tried this script and it doesn't seem to work for me. Headbomb {talk / contribs / physics / books} 20:46, 22 February 2017 (UTC)

Headbomb Maybe we sould ask the person who created this how it works. I have not tested. it. It's still a good idea to create such a script as it was suggested anyway. I bet this will help T111663. -- Magioladitis (talk) 21:40, 22 February 2017 (UTC)

Per User talk:Evad37/Watchlist-hideAWB.js now we have a tool to hide AWB edits. -- Magioladitis (talk) 09:43, 24 February 2017 (UTC)

Thanks for posting this. I haven't tested it myself, but it looks useful. However, it isn't really a solution to the "cosmetic edits cause watchlist problems" issue. That problem is: if you don't use the expanded watchlist, and you hide bot edits, and a vandal edits a page on your watchlist, and a bot edits the same page after the vandal, then you will not see any changes on your watchlist for that page, and you will therefore miss the vandalism. That problem can't be solved by a script that hides more edits. Also, while this script would help reduce watchlist clutter from AWB edits, it somewhat misses the point that people might want to see and review AWB edits to pages on their watchlist, and the extra cost of reviewing very small changes might make them a net negative even if the change itself is unobjectionable. Opabinia regalis (talk) 01:28, 25 February 2017 (UTC)

Reminder on "Cosmetic"

Extended content

Section 3.3.6 needs change since it uses the term "cosmetic" without quotes which could imply that this term really exists while the term was never defined in the Evidence page. The correct is to use "cosmetic" in quotes. -- Magioladitis (talk) 09:35, 24 February 2017 (UTC) We need a sentence that if the watchlist problem is solved, and it seems that it may be solved soon, any restrictions on general fixes editing become outdated. This will even affect the AWB rules of use. -- Magioladitis (talk) 12:11, 24 February 2017 (UTC) I absolutely disagree with this interpretation. -- Amanda (aka DQ) 20:00, 25 February 2017 (UTC) DeltaQuad then you need to add more why the general fixes are a problem. -- Magioladitis (talk) 20:40, 25 February 2017 (UTC) A restriction is not invalidated because it becomes "outdated". Cosmetic edits also have other reasons for being limited and restricted, but that's not to go into here. -- Amanda (aka DQ) 20:50, 25 February 2017 (UTC) Just to draw a line under this: the change you suggest is not arbcom's job. You need an RfC. "A major practical consideration justifying this policy has been mitigated by technical changes A and B" is a reasonable argument to make in an RfC that proposes relaxing or deprecating COSMETICBOT and/or AWB rule 4, but it is not within arbcom's remit to decide the conclusion of that discussion. Opabinia regalis (talk) 10:13, 26 February 2017 (UTC) Well, we should go because the restriction will be based on the "Problems with cosmetic edits" section I guess. -- Magioladitis (talk) 20:54, 25 February 2017 (UTC) Per ArbCom's draft: Cosmetic or inconsequential edits can be problematic because they clog watchlists and page histories, incur time costs for performing and reviewing trivial tasks, may reduce the likelihood of detecting vandalism and damaging edits when performed using bot-flagged accounts. They also write that "a phabricator task (phab:T11790) has been open since 2007 seeking a change in watchlist behavior that would mitigate some of these problems." A script that hides AWB edits will cover the first one and if the edits are not done by bot-flagged accounts we cover the third point. -- Magioladitis (talk) 12:44, 24 February 2017 (UTC) WP:HIDEAWB is now available. AWB Rules of Use is not a policy. It was created mainly because Smackbot was highly active so there was no need at the time to ave many editors doing general fixes. -- Magioladitis (talk) 09:56, 25 February 2017 (UTC) Whatever you call it, it's longstanding common practice and if you want to change the rules of use, you should propose your changes in an RfC. Opabinia regalis (talk) 10:13, 26 February 2017 (UTC) Still not a policy. There are more tools out there that do similar things. Someone even created AWB-script which uses the word "AWB" but is not related to our product. -- Magioladitis (talk) 10:19, 26 February 2017 (UTC) The phab ticket was created on Feb 17 2016 by me. This is a year ago. Long before this case started. -- Magioladitis (talk) 15:48, 26 February 2017 (UTC) Oops, I misread the date; thank you! (And thanks to Doug for removing that bit.) Opabinia regalis (talk) 18:44, 26 February 2017 (UTC) Re on Carl: You raise two different series of edits here. The invisible characters was discussed in my talk page recently. It was proven to be a valid task that it even changes in the visual output. -- Magioladitis (talk) 13:16, 27 February 2017 (UTC) To ArbCom: Headbomb now questions other bots. Not just my actions. The Wikipedia:Arbitration/Requests/Case/Magioladitis/Workshop#Proposal_by_User:Headbomb proporsal had only my name. Now it asks about Menobot. -- Magioladitis (talk) 20:57, 27 February 2017 (UTC) Wikipedia_talk:Bot_Approvals_Group#Trials_and_subtasks. BRFA procedure is about to change. -- Magioladitis (talk) 07:14, 28 February 2017 (UTC) It's totally positive that there is an effort to describe "cosmetic" as "a variety of minor errors, formatting problems, accessibility issues, wikicode syntax irregularities, and other inconsistencies across Wikipedia" which means are valuable edits i.e. edits that can be done. -- Magioladitis (talk) 09:34, 2 March 2017 (UTC) To ArbCom: Remedy 7.1 means I can fix anything that was consensus or not? -- Magioladitis (talk) In the meantime, this is why the patrollers in fact need help by CHECKWIKI. -- Magioladitis (talk) 16:13, 2 March 2017 (UTC)

So, since the term is not used in any passing remedy this is not really needed now. -- Magioladitis (talk) 00:07, 15 March 2017 (UTC)

I would like to emphasise the fact that obviously that changing {{Template:foo}} to {{Foo}} obviously changes the rendered HTML because the "Skip if only cosmetic changes" in AWB does not skip the page. Just to underline the fact that not everything is that obvious. -- Magioladitis (talk) 13:00, 16 March 2017 (UTC)

First of all, I just tried this in my sandbox, and I'll be damned if I can find any difference in anything but the commented-out performance data.

More substantially, to put a stop to this developing meme, which I see on your talk page as well: there is no consensus that "cosmetic" can be defined as "changes the rendered HTML". You may want to propose that definition as part of the RfC recommended by this decision, but you cannot act as if that definition already exists and has consensus. (I personally, as a fairly literal-minded person, would strongly prefer to see an RfC result that provides a technically specified definition of "cosmetic". But the community is hardly obliged to decide in favor of your, my, or anyone else's personal preferences.) Opabinia regalis (talk) 20:56, 17 March 2017 (UTC)

Ramaksoud2000

I think Ramaksoud2000 should also be reminded to refrain from actions that cause drama and use the normal way to contact other editors. This part is totally missing from the final text. It's clear that if this thing is not dealt then there s a chance that Ramaksoud2000 go after other editors after that. -- Magioladitis (talk) 08:58, 13 March 2017 (UTC)

No, it does not make sense to wag our fingers at someone for making a request for dispute resolution which was accepted. Opabinia regalis (talk) 00:08, 15 March 2017 (UTC)

Opabinia regalis They may continue their crusade in the future in not discouraged. Ofcourse, the may not since CHECKWIKI projects stopped for English Wikipedia. No updates for the last 2 months. -- Magioladitis (talk) 00:15, 15 March 2017 (UTC)

There is no evidence of anyone being on a "crusade". It would be much more helpful, going forward, for you to deal with critical feedback without making that kind of assumption about the motivations of the person offering it, even if you find the method of delivery annoying. Opabinia regalis (talk) 05:25, 16 March 2017 (UTC)

AWB interface

"in order to use technical means to avoid problematic edits more effectively" What does that even mean? What are the problematic edits? AWB is open source. I think WP:SOFIXIT applies too here. Moreover, if anyone has a request there is a Phabricator project open. What os the meaning of ArbCom requesting something like that from volunteer programmers? -- Magioladitis (talk) 08:56, 17 March 2017 (UTC)

Minimally, it means "please pay attention to the results of the cosmetic-edits RfC, if there is one". Opabinia regalis (talk) 20:56, 17 March 2017 (UTC)

Thryduulf's section

@DeltaQuad: your proposed remedy 10 needs an exception to allow Magioladitis to unblock his bot if they are the person who blocked it. This could be construed as permissible under the "discuss with the blocking administrator" clause, but it's probably better to make it explicit. Thryduulf (talk) 12:18, 22 February 2017 (UTC)

Agree that would be permissible, but no evidence they've ever blocked their own bot so I'm not sure there's a use-case for the clarification. Please let me know if there's something I'm missing, however. -- Euryalus (talk) 12:21, 22 February 2017 (UTC)

I haven't looked to see whether they've done this or not. My goal is to reduce the chance for confusion and wikilawyering from the start - a single sentence (at most) now could save paragraphs (or more) of discussion later (which could confuse any other issues that arise). I know some bot owners have blocked their own bots (for a variety of reasons) and I see no reason to cause headaches and unnecessary issues should Magioladitis want/need/choose to do this (one possible reading of the restriction could be that if Magioladitis is the person who blocks the bot then it can never be unblocked while this restriction is active). Thryduulf (talk) 12:45, 22 February 2017 (UTC)

Xaosflux's section

@DeltaQuad: regarding specify the venue for discussing bot unblocks, "it depends". If you want to list some though I suggest adding Wikipedia:Bot owners' noticeboard. — xaosflux ^Talk 14:02, 23 February 2017 (UTC)

Headbomb's section

• Given that several policy-related discussions and clarifications are ongoing, and possibly changes to how AWB genfixes & checkwiki processes are handled, I would like ARBCOM to have a motion/proposed decision that any restrictions/decisions can be reviewed after (and if) policy changes, or rescinded if policy now make the ARBCOM decisions/restrictions redundant. Headbomb {talk / contribs / physics / books} 12:04, 24 February 2017 (UTC)

• @Magioladitis: the default watchlist interfering with vandalism detection is only ONE of the reasons for WP:COSMETICBOT, or restrictions on general fixes, not the only reason. Headbomb {talk / contribs / physics / books} 12:37, 24 February 2017 (UTC)

• @DeltaQuad and Kelapstick: related to "Problems with cosmetic edits", you objections seems to be based on reading more into the proposed FoF than it was actually says.

"Cosmetic or inconsequential edits can be [not are] problematic because they clog watchlists and page histories, incur time costs for performing and reviewing trivial tasks, and may [i.e. in some not all cases] reduce the likelihood of detecting vandalism and damaging edits when performed using bot-flagged accounts. A phabricator task (phab:T11790) has been open since 2007 seeking a change in watchlist behavior that would mitigate some [not all] of these problems." (emphasis mine, bracketed comments mine)

This is a very reasonable FoF, with more than enough room for proper context and interpretation. I don't see the grounds for objection. That bots are flagged allowing some editors to unclog (or clog) their watchlist doesn't alleviate the 'problem' with page histories, but also for editors who choose to see bot edits because bots aren't perfect, these cosmetic changes still incur opportunity costs. You may argue that it's not a big opportunity cost, but there is an opportunity cost. It's also not because you can used advanced options to see more than just the most recent edit that it changes the fact some editors will not see the most recent non-bot edits. This does mean that the likelyhood of vandalism being caught is reduced. You may argue that it's not a big reduction, but there is a reduction. Headbomb {talk / contribs / physics / books} 14:26, 24 February 2017 (UTC)

I don't disagree with the facts, but the remedy as written puts undue weight on cosmetic edits and genfixes as being the reason why things are broken, without describing solutions or presenting other factual information. There will always be opportunity costs, regardless if cosmetic edits are in the way or not. -- Amanda (aka DQ) 19:56, 25 February 2017 (UTC)

• To all arbs Re "Community encouraged to review policy on cosmetic edits". I don't really see what these two phabricator tickets have to do with the remedy. While, if implemented, they would be solutions to some issues that have been identified here and elsewhere, these two tickets are absolutely not the place to discuss what is or is not a cosmetic edit. Remove mention of the tickets, and the remedy is fine.Headbomb {talk / contribs / physics / books} 14:35, 24 February 2017 (UTC)

• To all arbs Re "Enforcement of restrictions", I would like this to be specifically addressed, rather than adopted by default. Particularly, if Yobot makes a cosmetic edit, it seems ridiculous to block the bot for 1 month, or in a different reading, to block Magioladitis for one month because of it. I think you need enforcement remedies tailored to the situation. In particular, I would propose "If Yobot (and other bots operated by Magiodilitis) violates restrictions, it may be disabled per WP:BOTISSUE. Magiodilitis may not resume operation of the bot before the issue has been addressed, and if blocked, without the consent of the blocking admin. If Magiodilitis violates restrictions, the standard enforcement principles apply." Headbomb {talk / contribs / physics / books} 14:52, 24 February 2017 (UTC)
  • I'll have to think about this and discuss this with my fellow Arbs. I've never been a big fan of standard enforcement, but that said, the sanctions if Magioladitis violates need to have a strong backbone. -- Amanda (aka DQ) 19:56, 25 February 2017 (UTC)
  • This is a good point. Prescribing escalating blocks for repeated but non-escalating problems seems unlikely to help. Opabinia regalis (talk) 19:47, 26 February 2017 (UTC)
• To all arbs Re "Magioladitis requested to seek broad consensus for new tasks", WP:AN notices would make sense if you want to solicit the input of admins, but bot approvals are part of WP:BAG's mandate, not WP:ADMINs (and WP:BAG already monitors bot-related noticeboards). If you want to mandate notices to the community, posting a link to the BRFA on one of the village pumps would be the place to do it, rather than WP:AN. I say this as a BAG member, so you can save Magioladitis the confusion about where to put such notices, and the community confusion about where they should expect them, if you deem those notices necessary. You might also want this remedy to apply to any Magioladitis-run bots rather than only Yobot. Consider instead "With the exception of existing Yobot tasks currently being re-reviewed, Magioladitis is requested to seek broad input into new Yobot tasks, such as by posting a link to the BRFA on AN the VPs, or by another method accepted by BAG or any other noticeboard deemed relevant by BAG."

I note, however, that WP:BAG is already empowered to request such notices to the community, if we deem it necessary. As such, the remedy might alternatively be construed as BAG is reminded to ensure that broad community input has been sought when reviewing BRFAs, such as by posting a link to the BRFA the VPs, or any other relevant noticeboard."

I further note we recently wrote Wikipedia:Bot Approvals Group/Guide#Guide to BRFAs, which outlines what BAG members should do when reviewing BRFAs, which includes ensuring that the community has a chance to give its input on the tasks. Headbomb {talk / contribs / physics / books} 02:29, 25 February 2017 (UTC)

This outlines some of the many problems with this remedy, and I know that there is talk of a rewrite for this, so i'll make sure these points are addressed. -- Amanda (aka DQ) 19:56, 25 February 2017 (UTC)

• @Opabinia regalis: I meant "one of the Village Pumps" rather than "all the VPs". VPT is probably the most relevant in many cases, but there are some cases where another VP would be good. It mostly depends on the nature of the bot task proposed. Headbomb {talk / contribs / physics / books} 10:37, 26 February 2017 (UTC)

@BU Rob13, Doug Weller, and Opabinia regalis: This is a reply to BU Rob13's comment above. BRFAs are there to evaluate both the technical soundness and consensus of a task, it is not only a 'technical approval'. BAG routinely requires BRFAs to be advertised as a way of ensuring said consensus, or as a way of making sure the community has a chance to give its input on the implementation details. (And this should clearly NOT be done at WP:AN. AN is for admin matters, not for soliciting input or holding discussion on the specifics of a proposed bot task.) If an RFC is required before a BRFA, BAG is more than capable and perfectly willing to let Magioladitis know one is needed. We do not, and should not, however, require a discussion / RFC before every proposed bot task. Let's say Yobot 52 was proposed as a new Yobot task. It's patently obvious that this is a legit task that should be done, and no RFC needs to be held for that. Likewise if Magioladitis decided to tackle this BOTREQ. Let BAG do its job. Headbomb {talk / contribs / physics / books} 10:58, 26 February 2017 (UTC)

@Headbomb: Are you suggesting rewording the wording you suggested "requested to seek broad input into new bot tasks," so that it doesn't mean literally all new bot tasks? Doug Weller talk 11:53, 26 February 2017 (UTC)

The original wording is to seek broad input, for example, by posting a link to the BRFA at AN, or other noticeboards per BAG discretion. My proposed revised wording it to seek broad input by posting a link on the VP, or other relevant noticeboards as per BAG discretion. That can apply to all BOT tasks if ARBCOM deems it necessary (and that's how I read either wordings). However, "requested to seek broad input" doesn't mean "hold a discussion/RFC for every bot task before filing a BRFA" as BU Rob13 seems to think it means. Headbomb {talk / contribs / physics / books} 12:10, 26 February 2017 (UTC)

@Magioladitis: "BRFA procedure is about to change." No it is not, it's the same procedure as before. Many, if not most bots were properly trialled, including those with several subtasks in their BRFA. The only thing "different" is that BAG will fix an occasional blind spot in the BRFA trials. Headbomb {talk / contribs / physics / books} 11:28, 28 February 2017 (UTC)

Stevietheman's section

I had not intended to participate again in this case, but there are some aspects I see in the proposed decision I felt a need to comment about. Overall, I can live with (so to speak) where this is going, so I have no strong objections to anything.

• Re: FoF #10, I would expand the problems to include "the potential for tidying conflicts/wars". As long as we have the requirement to have one substantive change in a bot/AWB edit, this potential is far reduced. If this requirement goes away, the skies the limit with back-and-forth tidying. These conflicts exist because in some aspects editors have their own ideas as to what tidying should be done, or whether an atomic tidying has any value at all. Requiring something substantive spaces these tidyings out, usually, to a point where few such conflicts will occur. To wit, if phab:T11790 is resolved, that is certainly a positive development that will address real problems, but it doesn't take away the need to avoid wholly non-substantive edits. Stevie is the man! ^{Talk • Work} 20:10, 28 February 2017 (UTC)
• Re: Remedy #3, I would like to nudge arbcom to a majority vote for that. After all, it is just an encouragement, and per my evidence about "AWB limitations", issues related to this case go beyond whether AWB provides enough skipping options. For those who don't understand AWB enough, I get that as it is complex, but if you assume good faith with my and others' technical explanations, I hope this would be an easy support. Stevie is the man! ^{Talk • Work} 20:10, 28 February 2017 (UTC)
  • On this point, I would like to clarify that "skipping genfixes" doesn't apply in this case because genfixes are used to fixed the checkwiki issues. It is correct, however, that a module can be used to run specific genfixes (although apparently that approach hasn't been used by Yobot so far). I will reiterate, though, that there have been issues in AWB that have helped to trigger some of the problems related to this case, such as not treating template bypasses as cosmetic in the sense of no rendered HTML being changed. So, if a user has a setting to "skip only cosmetic changes", it wouldn't have worked as far as these bypasses are concerned. "AWB limitations" in the evidence goes over this and much more. Stevie is the man! ^{Talk • Work} 13:30, 1 March 2017 (UTC)

Latest progress

What's the latest with this case? It'll be one month tomorrow since an outcome was due. Thanks. Lugnuts ^{Precious bodily fluids} 14:39, 6 March 2017 (UTC)

@Amortias:, @Opabinia regalis: Lugnuts ^{Precious bodily fluids} 14:22, 7 March 2017 (UTC)

@Lugnuts: To clarify one point: the dates listed for each case specify the ends of each phase and the target date on which a proposed decision is expected to be posted. A proposed decision is not an "outcome", and further proposals can be added as voting proceeds; the case is finalized only after the motion to close passes. We don't currently include a target date for completion of voting and closure of the case. Arguably we should do that, but it often happens that behavior during the voting phase ends up influencing what people think will be effective in solving the problem.

Recently we have been discussing a small modification to the enforcement provision, which I meant to post yesterday but got sidetracked. I will do that this afternoon. I'd expect things to wrap up fairly soon, and the currently passing remedies to be stable. Opabinia regalis (talk) 19:59, 7 March 2017 (UTC)

Thanks OR. Lugnuts ^{Precious bodily fluids} 20:01, 7 March 2017 (UTC)

(Hope it's fine to post outside my section this once.) I'd like to encourage the Committee to wrap this up as soon as possible. While this drags on, edits that would violate the restriction remedy continue to be made. That remedy isn't yet in place, but the AWB rules of use and WP:COSMETICBOT still are, so issues are ongoing. At some point, the marginal benefit of continuing to refine the encouragements, reminders, etc. is far outweighed by the marginal cost to the community of the ongoing issues. My hope is that the problematic edits will stop entirely once an actual restriction is in place, but if they do not, at least there will be enforcement procedures in place. (Unlike the encouragements, getting proper enforcement procedures in place is important, so this isn't prodding anyone to skip over that important step.) ~ Rob₁₃^Talk 23:48, 7 March 2017 (UTC)

Which edit do you find now problematic? I find problematic the fact that you keep calling my edits problematic. I wonder that the problem is. Using the word "problematic" solves the problem of explaining what the exact problem is, -- Magioladitis (talk) 23:58, 7 March 2017 (UTC)

You've requested I not discuss this with you, so I haven't. But if you'd like me to clarify: See the big section on your talk page pointing out some problematic edits. Some with non-breaking spaces were fine (and I explained that to slakr on IRC, to ensure he was being fair to you) but others were clearly cosmetic-only. As for what's problematic, feel free to refer back to this, where I highlight in great detail that you know exactly what people find problematic about your edits. Continuing to pretend you've never been told is not endearing. ~ Rob₁₃^Talk 00:51, 8 March 2017 (UTC)

If I'd planned to write this from the beginning I would've dispensed with the "sectioned talk page" thing, which I hate. But it's only my third-least-favorite feature of case page formatting... ;)

Magioladitis, there are lots of things in need of gnoming on the wiki. It would be a good idea to focus for now on the ones that are unambiguoussly not "cosmetic", especially pending the outcome of (what will hopefully be) an RfC to clarify the matter. Opabinia regalis (talk) 05:42, 8 March 2017 (UTC)

RF's second section

I had not seen, I think, proposed remedy 10 until recently. It is clear that this is something that has been discussed in the past, with an opposite conclusion reached in general. I do not believe that there is a substantive issue relating to unblocking here, I think we went through the unblocks in the workshop, and found them all, or almost all to be sound.

Therefore I see no reason for this proposed remedy. I can see that it might be wise for Mag to avoid doing that, in future, in order to be "squeaky clean" (I think he has agreed to that), but passing this remedy, without either clearly stating that the unblocks were within policy, (or identifying at least a number that were not) seems to risk implying that policy was broken when it wasn't.

All the best: Rich Farmbrough, 23:20, 14 March 2017 (UTC).

It's specifically not a statement of general policy, and only a requirement that Magioladitis in particular not do it. Opabinia regalis (talk) 00:11, 15 March 2017 (UTC)

A requirement based in no policy and in no evidence that my actions ever violated something. -- Magioladitis (talk) 00:13, 15 March 2017 (UTC)

In the workshop page I suggested the following:

Magioladitis is reminded that unblocking Yobot really annoys some people. As such he is requested not to do so, ...

In doing this I gave a reason, and a perfectly sound one, to not unblock his bot. (And you will see that he agreed to the request, with good will.)

The Remedy given here has no basis in any FoF - nor indeed does it provide any reason. It is just a fiat.

I know Arbcom has the power to impose remedies, and does not need to (and in some cases should not) subscribe to a good faith model of editor behaviour. But this is not one of those cases.

All the best: Rich Farmbrough, 22:25, 19 March 2017 (UTC).

I would also add that I am wary of "remedies" that are added after the nominal expiry date of the proceedings. All the best: Rich Farmbrough, 23:49, 25 March 2017 (UTC).