Opabinia regalis (talk | contribs) →Follow-up: re |
David Tornheim (talk | contribs) m →ARCA notice: ce |
||
(One intermediate revision by the same user not shown) | |||
Line 100: | Line 100: | ||
:Yikes, the implementation of the RfC language has been a mess. I think the problem arose because, afterwards, editors decided to reformat the references, and the accompanying punctuation got messed up in the process. The required language and punctuation are at [[WP:GMORFC#Proposal 1]]. I think the comma in question ''is'' there, so it needs to be added back at pages where it got deleted. In any case, your edit is no big deal, and I certainly wasn't finding fault with you for it. I'm just trying to make sure that the RfC results remain stable, after all the work and drama that went into it. --[[User:Tryptofish|Tryptofish]] ([[User talk:Tryptofish#top|talk]]) 00:08, 27 July 2016 (UTC) |
:Yikes, the implementation of the RfC language has been a mess. I think the problem arose because, afterwards, editors decided to reformat the references, and the accompanying punctuation got messed up in the process. The required language and punctuation are at [[WP:GMORFC#Proposal 1]]. I think the comma in question ''is'' there, so it needs to be added back at pages where it got deleted. In any case, your edit is no big deal, and I certainly wasn't finding fault with you for it. I'm just trying to make sure that the RfC results remain stable, after all the work and drama that went into it. --[[User:Tryptofish|Tryptofish]] ([[User talk:Tryptofish#top|talk]]) 00:08, 27 July 2016 (UTC) |
||
:Never mind that, I just looked back and self-reverted. The comma should of course be before the cites, and an extra comma after was indeed the previous mistake. --[[User:Tryptofish|Tryptofish]] ([[User talk:Tryptofish#top|talk]]) 00:14, 27 July 2016 (UTC) |
:Never mind that, I just looked back and self-reverted. The comma should of course be before the cites, and an extra comma after was indeed the previous mistake. --[[User:Tryptofish|Tryptofish]] ([[User talk:Tryptofish#top|talk]]) 00:14, 27 July 2016 (UTC) |
||
== ARCA notice == |
|||
I have filed two actions at [[WP:ARCA]] of which you are named party: |
|||
[https://en.wikipedia.org/wiki/Wikipedia:Arbitration/Requests/Clarification_and_Amendment#Clarification_request:_Genetically_modified_organisms action 1], [https://en.wikipedia.org/wiki/Wikipedia:Arbitration/Requests/Clarification_and_Amendment#Clarification_request:_Genetically_modified_organisms_2 action 2] |
|||
--[[User:David Tornheim|David Tornheim]] ([[User talk:David Tornheim|talk]]) 04:06, 27 July 2016 (UTC) |
Revision as of 04:25, 27 July 2016
Newsletters.
Check RfAs.
WP:ADREV.
Statistics on most-viewed neuroscience pages.
User:Skysmith/Missing topics about Neurology
Commons:Category:Smilies
1, 2, 3, 4, 5, 6, 7, 8, 9, 10 |
Follow-up
Regarding this idea of a "COI investigations mailing list" that's been floating around: I know you've spoken in favor of it, and are usually sane :) I have extremely strong reservations about this and am not sure whether those who like the idea have had a chance yet to think through the back-end logistics or long-term management of this system.
- You'd want the WMF to create, host, and oversee a database or archive exclusively dedicated to storing personally identifying material about alleged COI/paid editors. Inevitably, it will include inaccurate identifications, accusations that would be libelous if public, and attempts at taking out wiki-political opponents, as well as garden-variety private information gathered from amateur internet sleuthing that may not have previously been collected in one place. The reality is that stuff leaks sometimes - technical compromises happen, accidental disclosures happen, and occasionally people do "go rogue". This could be much worse than existing archives leaking, because it would consist almost entirely of off-wiki personal information. There tends to be a great deal of moralizing from some quarters about paid editing, but hopefully we all agree that getting doxxed is not a reasonable consequence of having once possibly violated a website's terms of use.
- You'd want a group of volunteers (existing functionaries? a new set?) to receive this private information and then issue public rulings on whether there is or is not a COI/paid relationship. This is personally and legally risky if they get it wrong, and it's highly unlikely that this group is going to end up with any meaningful amount of training or resources to do the investigations. Wikipedians appointing themselves as "sleuths" and organizing private mailing lists for the purpose hasn't historically worked out all that well.
- You'd want someone (arbcom? the WMF?) to oversee the selection of this group's membership, hear appeals of their findings, deal with disputes that arise, and remove people from the group if needed. This is not a responsibility that anyone currently has extra bandwidth for.
It is not at all clear to me that any of this is doable, or desirable, or worth the level of time and effort that would be required. (Frankly, whenever I skim COIN I think the same thing.) Those who are very concerned about paid editing would, IMO, be best served directing their energies toward changing the notability and sourcing expectations for the topics that are most commonly affected. Opabinia regalis (talk) 06:50, 23 July 2016 (UTC)
- Thanks for taking the time to write this very thoughtful message to me. (But don't be so sure about my sanity!) At this point, I'm in the thinking-out-loud stage, so I'm not strongly committed to any particular approach. And having users like you, who are more plugged in to the logistics than I am, give me advice about things that can be deal-breakers is very helpful to me, and you can be sure that I will take everything you say very seriously.
- No matter what you and some other users with advanced permissions may think, simply trying to revise notability and sourcing guidelines will end up in a lot of heat followed by no consensus. And the fact that many community members are correct, that undisclosed paid editing presents a monumental threat to Wikipedia's mission, cannot be wished away. Nor should it be.
- What I'm trying to do is to get something in place that legitimately allows us to have a bright-line outing/harassment policy, a policy that says that it is never, never, never OK to post outing-type information on-site – but that also allows editors to investigate undisclosed paid editing without having the disruptive user say: you are making a personal attack on me, because you are accusing me of COI but you provide no evidence. There needs to be a private way to evaluate private evidence, that verifies the evidence but makes it credibly possible to never post it on site.
- Taking your points one-by-one, I agree with you that getting doxed is the wrong solution to someone violating the terms of use. I want to prevent that from happening. The way I tentatively see it at this time, such a mailing list would contain information that would not really be a whole lot different in terms of legal liability and risk of doing harm than what Checkusers already do. They call up IP information that can sometimes identify a user as sitting at one very particular computer, and determine the likelihood that two accounts are or are not the same person. To my knowledge, that doesn't mean that a Checkuser can get sued for posting "confirmed" or "likely" on-site, and I think (again, to the best of my knowledge) that Checkusers outing editors has not been a pervasive problem. And I understand that there is a log kept of all Checkuser runs. And I'll make a bet (again, to the best of...) that the Functionaries email list gets stuff that should never be posted on-Wiki, and the ArbCom list even more so. I've had email communication with the ArbCom list in which my real life name is revealed, and I have trusted ArbCom not to go blabbing my name in public. Yes, there was that infamous leak, but I hope that procedures have been improving since then.
- But what you say here gives me the idea that the new whatever that I might propose should be set up in such a way that once a case is dealt with, it gets erased from whatever digital storage there is, not stored or archived.
- To your second and third points, I've been thinking about defining membership on the list as being: those users who are currently on the Functionaries email list who volunteer to participate. It would never be dealing with urgent matters (unlike Oversighting), so there would be no need for a lot of volunteers. If one is already a Functionary in good standing (and I understand that ones in bad standing are kicked off the list), one is already cleared by WMF to have access to private information. And folks who don't have the time or inclination won't have to do anything.
- And here is something very key: under no circumstance should it be permissible to post on-Wiki anything more than that it is "confirmed" or "likely" or "not confirmed" (or something like that) that a COI exists. But not what the COI is (beyond the fact that it pertains to certain pages)! That's not revealing anything even remotely libelous, and the "harm" that could be done to a user would simply be being required to disclose a COI in discussions about a page, something we already ask users to do under the WMF terms of use. I'm not seeing (?) much need for appeals. I don't imagine that any of the Functionary volunteers would be seeing, privately, anything more than what Arbs, Checkusers, and Oversighters already see. I don't think the sky will fall.
- But if you can shoot holes in what I've said here, please do! That's the kind of feedback that I want! And I promise to take it seriously.
- When I have more of an idea of what I think it should be, I'm going to create a draft (with "no index") in my user space, and I'll post about it to Arbs, Checkusers, and Oversighters, and ask for feedback. I'm not going to attempt to make an RfC to the community to adopt a proposal, until after the Functionaries have had a chance to vet it. --Tryptofish (talk) 20:42, 23 July 2016 (UTC)
- Well, to start at the top,
undisclosed paid editing presents a monumental threat to Wikipedia's mission
- unshared implicit premises are part of the problem. I wouldn't agree to that at all - paid editing is a nuisance and a timesink, yes, and certainly some people believe there is some kind of clear and present danger, but there's precious little evidence of it. Yes, we have spam, and so does every other website with user-generated content. We're lucky in that ours is often in the form of new articles on crap topics with little traffic; it didn't much bother the 209k people who wanted to know who Tim Kaine is yesterday. The damaging insidious advocacy cases (Wifione, etc) are a) believed to be "paid" by inference, not evidence, and b) not the cases that would ever be usefully handled by this COI investigation process. - The advantage of approaching this whole problem through notability and sourcing is that the changes work with existing infrastructure and can be done incrementally and reassessed regularly. Sure, there will be no consensus for sweeping all-at-once changes, but the point is to deal with bite-sized pieces of the problem in a flexible way. Possibly this could be coupled with technical improvements like automated classification of new articles to prioritize patrollers' attention. This mailing list is a huge bureaucracy of uncertain benefit that would be extremely difficult to get rid of once it's entrenched.
- I see the temptation, but I think checkuser is a poor analogy here. Checkuser involves the WMF collecting fairly simple routine technical data from everyone, storing it temporarily, and permitting access only under certain conditions to authorized people who are generally familiar with the technical aspects before they start using the tool. The COI investigation process would involve self-appointed third parties collecting random internet flotsam from whoever they decided seemed suspicious, storing it indefinitely on their own computers of indeterminate security, and submitting it to a mailing list for review by people with no particular specialized knowledge. In the CU case the information is provided by the affected users themselves, by virtue of their use of the site; in the COI case the information is gathered from off-site by others only when suspicion arises. Assuming good faith of the reporters, that's a recipe for confirmation bias; taking a more cynical view, it's a good way to try to knock out your opponents in a dispute. (You're much more involved with the GMO topic than I am, and have certainly seen as much of the alt med cluster of disputes, so clearly you know that some editors deploy COI allegations as weapons. Right now they make these claims in public where others can see they're unfounded, but what happens when they gather suggestive tidbits and tie them together with rhetoric, and the report gets read by someone who's just trying to clear the queue and doesn't really understand the subject?) As for legal risk, you're right, nobody gets sued for saying "User:JoeBloggs is technically indistinguishable from User:SpammitySpamSpammer." But that's a completely different claim than "User:JoeBloggs has been engaging in deceptive practices for his own personal profit and in so doing has violated the legal agreement he made by using our website."
- Honestly, it kind of worries me to see comments like
I'm not seeing (?) much need for appeals
- the question mark must be those cases where the determination is factually incorrect, and thus there is indeed harm being done to the user? The functionaries are reasonable people who would be careful, naturally, but people make mistakes and misjudge evidence all the time, and digging around in editors' off-wiki business is not a skill anyone was selected for, and overall you seem to have a much greater degree of confidence than I do in the likely success rate of this venture. (That, by the way, is the catch-22 of the data retention problem - how would you go about appealing a sanction if the evidence on which it was based has been discarded?) - So will the sky fall, probably not. But it is an enormously complicated and uncomfortably invasive solution to a problem more effectively addressed through other means. Opabinia regalis (talk) 06:44, 24 July 2016 (UTC)
- Well, to start at the top,
- Again, thank you, and again, I very much value this discussion. You've given me some things to think about, which I will, and perhaps I will change my mind about some things. Here are some of my thoughts as of now.
- Starting with that question mark about appeals, I did that on purpose, because I'm not that confident myself about what I said, and I wanted you to reply critically. I think you are right in identifying a problem about deleting material while allowing appeals based on that material, and this is something I'm going to have to take some time to work on.
- About what I called the "monumental threat", well, you say tomayto and I say tomahto. I noticed yesterday that your ArbCom colleague DGG commented at the outing RfC in a manner at odds with what you have said to me here, and he did a very good job of making the case. My approach to this is to try to find a way that duly respects both perspectives, and anything I might propose would be premised upon making it out of the question to post sensitive stuff on-site. (But, by the way, another idea that's coming out of those discussions is to have the developers create a simple question that comes up as part of the account registration process, in which a new editor can say yes or no to whether they expect to edit for pay, and a yes would automatically give them advice on their new talk page about how to do things right. I'm going to propose that in the next round of WMF asking for feature proposals, this November.)
- I have trouble seeing how to revise the notability and sourcing guidelines without creating unintended knock-on effects.
- It seems to me that "huge bureaucracy" and "enormously complicated" are hyperbole.
- About individual editors keeping private information about other users on their own computers, they do that already. (Even on mine!) That's not a liability for Wikipedia, and it's not a change. What Wikipedia should be concerned about is what is on WMF servers, etc.
- About editors using this kind of thing in POV disputes, I agree with you 100%. No bleeping way! Something else that I'm going to push for is that this kind of investigation must be strictly for COI that is strictly defined as monetarily based, as opposed to having an opinion. (No investigating someone because they might be a racist, etc. etc.) I've been arguing for that. I would insist that any email submitted to a COI process should be kicked to the curb if it's about POV and so forth. Otherwise, it's a non-starter for me.
- Well, in any case, I take it that if it ends up as a volunteer committee, you won't be one of the volunteers. But thanks again for your feedback. --Tryptofish (talk) 19:14, 24 July 2016 (UTC)
- Uh-oh! Just saw this: [1]. --Tryptofish (talk) 19:46, 24 July 2016 (UTC)
- Damn, people noticed! Now I don't even have implausible deniability! ;) Opabinia regalis (talk) 01:30, 26 July 2016 (UTC)
- Sounds like you've been disappeared! --Tryptofish (talk) 00:02, 27 July 2016 (UTC)
- Damn, people noticed! Now I don't even have implausible deniability! ;) Opabinia regalis (talk) 01:30, 26 July 2016 (UTC)
- Uh-oh! Just saw this: [1]. --Tryptofish (talk) 19:46, 24 July 2016 (UTC)
- I'm continuing to think hard about this, and two more thoughts have occurred to me.
- You are right about what you said about Checkusers, but I really was only trying to say that, in terms of being authorized to see personal information, there's no difference, and in terms of posting on-site things like "confirmed" there are at least similarities.
- About appeals and data retention, I've come around to agreeing that there needs to be a procedure for the accused to defend themselves fairly. One way to do this might be to have the Functionary, only if they decide to support the accusation, take the email they received, scrub it of personal information about the filing editor, etc., and forward it to the accused editor (if email is enabled, otherwise, leave a message on the editor's talk page asking to be emailed back). (This procedure would be explicitly stated, so editors making accusations will know that it will happen.) After doing so, the Functionary would delete everything from WMF systems. I think that would enable appeals even with the subsequent deletion, would make the process fairer, and would still minimize retention of sensitive data. --Tryptofish (talk) 19:30, 25 July 2016 (UTC)
- Yes, I think DGG and I disagree on this - others may yet weigh in. We haven't discussed the paid editing issue as much as last year's committee did, but last year there was a very strong consensus against having arbcom take on that role, and having read those archives I am very much in agreement with that consensus. I'd also say that "comment on content, not contributors" is a long-standing and effective social norm here, as is respect for others' choice to contribute anonymously, and institutionalizing systematic offsite research of other editors' identities and likely motivations is much closer to the description "threat to our mission" than having a backlog of spam to delete.
- I really do think the first thing anyone who wants new processes needs to do is gather some data showing that the types of articles or editors they think are problematic are not being effectively dealt with using existing processes. It doesn't matter so much whether someone uses phrases like "monumental threat" or "chronic nuisance" or something in between; it does matter that we actually understand the scale of the problem at least somewhat quantitatively. In particular, comments on this issue often circle back to the matter of transactions on elance/upwork/fiverr/etc., but my anecdotal experience is that content that comes from these mechanisms is almost always very poor. There is little need for offsite investigations to determine that most of this content is unwanted. The more serious Wifione-style long-term advocacy is not amenable to your system, since there's no evidence of money changing hands. I get the sense that a lot of people think that if they can prove undisclosed paid editing and therefore prove a ToU violation, they've done a better job or caught a bigger fish or something than if they just delete spam for being spam. We should be talking about what the scope of the problem is on a content basis and then designing the best solutions to make sure it gets reviewed efficiently and without wasting volunteer time, rather than trying to design a system based on catching some particular preconceived category of potential violations.
- Another possible issue I didn't think of before is the politicization of the functionary selection process, which has historically been (or tried to be) fairly non-political. Seeding this system with existing functionaries is one thing, but in the future we'd almost certainly see people applying for functionary positions for the sake of joining this COI squad, in addition to (or instead of) doing traditional functionary work. This doesn't really solve the problem of creating a new body because arbcom doesn't have the time/interest/desire/scope/etc., because functionary appointments and complaints are still overseen by arbcom.
- As for anonymized summaries of information as long-term storage mechanisms: better, but I still see this as a disqualifying problem. First, as Fluffernutter posted at length in the harassment RfC, outing problems can be created with long trails of individually non-identifying breadcrumbs. It's really difficult to know in advance what might prove to be a key piece of someone else's puzzle. It's better than storing every detail, but still difficult to justify given the relatively minor nature of the offense. Second, there's no such thing as a summary that doesn't interpret the underlying data. A future reviewer who has only the summary, where the summary was prepared only because the accusation was considered true, can't fully review the matter. Again, you seem very confident that the process would have very few cases of trolling, joe-jobbing, and plain old mistakes. Third, I focused on the data stored by the WMF because that's what we can control, but the existence of such a process necessarily encourages editors to dig and store their results. If the functionary says "not confirmed" but the submitting editor is convinced there is some grave threat to the project, they may well be tempted to publicize what they've gathered. We've already seen examples of similar behavior.
- Now, your new-editor signup idea is smart. The concept of having people deliberately segregate edits so that they have a dedicated and labeled paid-editing account is a sensible compromise... that was suggested, IIRC, by Greg Kohs circa 2006 ;) Opabinia regalis (talk) 00:31, 26 July 2016 (UTC)
- The sign-up idea wasn't mine, but I'm happily supporting it.
- As for politicizing the Functionary selection process, the selection process isn't particularly vulnerable to editor lobbying, and I think unqualified applicants will still be rejected. And I'm not asking ArbCom to do anything new.
- No, I didn't say long-term storage of anonymized data. What I did say seems to me to give the accused a good way to self-defend against trolling, joe-jobbing, and mistakes, and gives them as workable way to present their side to a future reviewer.
- I'd say anyone who tries to publicize rejected accusations is a prime example of someone who should be blocked or even banned for harassment. But I don't think anyone would be prompted by such a system to research opponents any more than the status quo does, because editors already do that all the time: what matters is what gets divulged on-site. --Tryptofish (talk) 00:02, 27 July 2016 (UTC)
- You're right, sorry, I misread your idea about the anonymized summaries - I think you mean that the functionary and the editor in question retain the summary, but no one else does? I had figured the accused editor, at least, would get to see the un-edited submission if they requested it (without the submitter's name, obviously). But maybe that's a bad idea - if the identification is not correct, you've just handed them a pile of links to personal information about someone completely different. This is a lot more complicated with actual personally identifying information than with CU data or on-wiki behavior, which may be distinctive but rarely is personally identifying on its own.
- Yes, I would prefer to just delete spam as spam. (Similarly, I also see no real need to chase up G5-able content unless there's reason to believe it's bad.) That's why I prefer content-level approaches like raising the requirements for notability and sourcing, and would rather see energy invested in progress on that - even if it's slow - than in creating infrastructure for real-world investigations. I recall a proposal back in the days of Essjay to create a similar process for investigating and (sort of) verifying editors' claimed real-world credentials, which is not really so different in principle and I didn't like it then either ;) Opabinia regalis (talk) 02:32, 27 July 2016 (UTC)
- I'd say anyone who tries to publicize rejected accusations is a prime example of someone who should be blocked or even banned for harassment. But I don't think anyone would be prompted by such a system to research opponents any more than the status quo does, because editors already do that all the time: what matters is what gets divulged on-site. --Tryptofish (talk) 00:02, 27 July 2016 (UTC)
AE case
It looks like the community has grown tired with some of the recent things we've seen in the GMO topic, so I've opened up an AE case here. I tangentially mention some evidence where you've shown your concern about casting aspersions, etc. so I thought I'd formally let you know. Here's to hoping this finally settles the topic down after it closes. I for one look forward to being able to focus entirely on content. Kingofaces43 (talk) 17:55, 24 July 2016 (UTC)
- I read everything that you presented there, and I've got it on my watchlist. I'm ambivalent about your doing this, because I'm not sure that it is the best way to get things to quiet down. On the other hand, you certainly made a strong case. I'll probably end up commenting there myself, but for now I'm going to wait and see what comes next. --Tryptofish (talk) 18:43, 24 July 2016 (UTC)
- Yeah, I'll admit I've been on the fence on when/if to go ahead with this for quite awhile. I'm still not happy with going ahead on it, but there have been so many warnings that it looks like things won't improve otherwise. The problem has got to be addressed at some point and the process is never going to be that clean either. Now is probably about as good or bad of a time as any. It looks like others have been commenting elsewhere on this issue that aren't as involved in the topic, so hopefully we may get some input from the not so usual folks. Kingofaces43 (talk) 19:12, 24 July 2016 (UTC)
- You doubtless know this already, but I feel like I have to say it. Please be extra careful and cautious in your own editing, so as not to incur a boomerang. --Tryptofish (talk) 19:17, 24 July 2016 (UTC)
- Definitely on my radar. Part of the reason why I went forward on this was because what should be minor situations are getting escalated so quickly instead of the single message or two it would have taken on an article talk page, so I'm very aware of appearances on that on both ends. I won't ramble here anymore though. I have a lot more I wish I could have said with regards to how battleground behavior from other editors can make those attempting to engage in normal editing look bad too, but I'll deal with that if it comes up in the correct venue. Kingofaces43 (talk) 19:58, 24 July 2016 (UTC)
- You doubtless know this already, but I feel like I have to say it. Please be extra careful and cautious in your own editing, so as not to incur a boomerang. --Tryptofish (talk) 19:17, 24 July 2016 (UTC)
- Yeah, I'll admit I've been on the fence on when/if to go ahead with this for quite awhile. I'm still not happy with going ahead on it, but there have been so many warnings that it looks like things won't improve otherwise. The problem has got to be addressed at some point and the process is never going to be that clean either. Now is probably about as good or bad of a time as any. It looks like others have been commenting elsewhere on this issue that aren't as involved in the topic, so hopefully we may get some input from the not so usual folks. Kingofaces43 (talk) 19:12, 24 July 2016 (UTC)
Recent revert
Hello! Minor thing, but when I look at other articles affected by the RFC (such as genetically modified maize and genetically modified food) the paragraph in question omits the comma I removed in this edit to comply with MOS:REFPUNCT. Is it all right if I remove it again? Me, Myself & I (☮) (talk) 23:46, 26 July 2016 (UTC)
- Yikes, the implementation of the RfC language has been a mess. I think the problem arose because, afterwards, editors decided to reformat the references, and the accompanying punctuation got messed up in the process. The required language and punctuation are at WP:GMORFC#Proposal 1. I think the comma in question is there, so it needs to be added back at pages where it got deleted. In any case, your edit is no big deal, and I certainly wasn't finding fault with you for it. I'm just trying to make sure that the RfC results remain stable, after all the work and drama that went into it. --Tryptofish (talk) 00:08, 27 July 2016 (UTC)
- Never mind that, I just looked back and self-reverted. The comma should of course be before the cites, and an extra comma after was indeed the previous mistake. --Tryptofish (talk) 00:14, 27 July 2016 (UTC)
ARCA notice
I have filed two actions at WP:ARCA of which you are named party: action 1, action 2 --David Tornheim (talk) 04:06, 27 July 2016 (UTC)