Following a discussion among the CheckUsers, I'm happy to announce that QEDK is promoted to full SPI clerk. On behalf of every one of us, thanks to all the clerks for their hard work. We couldn't do any of this without you. :-) Katietalk16:31, 8 October 2017 (UTC)
Thanks a lot, Katie and every checkuser who chipped in. I am glad to be of assistance and hope to continue with my good work. --QEDK (愛 • 海)17:34, 8 October 2017 (UTC)
I was thinking in terms of a more central place where input could be entered. User talk pages are not optimal for that purpose for various reasons. Coretheapple (talk) 17:27, 14 November 2017 (UTC)
Welcome to the club. There is no club, we have no moneys like them admins. But what we have is a cool fez. --QEDK (愛 ☃️ 海)09:46, 23 November 2017 (UTC)
Ah, I see admins at one noticeboard have passed you off to some other noticeboard, in our fine tradition. Well, let me join in by suggesting you would be better off to post this at WT:SPI rather than this subpage! Right helpful lot we are. In seriousness, I see that at least one of these is a case that I left hanging last month when I semi-retired from SPI, so I'll deal with that. But not tonight, I'm going to bed. Ivanvector (Talk/Edits) 04:41, 16 January 2018 (UTC)
Sir Sputnik promoted
Following a discussion with the functionaries, I'm pleased to announce that Sir Sputnik has been promoted to full clerk. Congratulations to them, and on behalf of all of us, thanks to all the clerks for the work they do. I've said it before and I'll shout it again – we could do none of this without you. :-) Katietalk22:58, 30 January 2018 (UTC)
It seems we don't really have a procedure for this and maybe I'm overthinking it, but I'd like to get back into active clerking. I've been away from formal clerking for a few months, although I'm never really away from SPI. Since promotions are supposed to be the purview of the CheckUsers and functionaries, it doesn't seem right to just put myself back on the active list, so I guess I'm reapplying? Let me know, anyway. Ivanvector (Talk/Edits) 14:35, 12 February 2018 (UTC)
Just so clerks know for the future, you don't need any "approval" to come back from an absence where you are merely marked temporarily inactive. It's another story if you're on the fully inactive list, which we usually only move clerks to after a year or more of no activity. That requires approval. Y'all have lives, as do the CheckUsers, and we fully expect that there will be some months where you need to take a break. Glad to see you back, Ivanvector! ~ Rob13Talk14:26, 14 February 2018 (UTC)
Cases with large histories
I'm running into a problem (most recently at Wikipedia:Sockpuppet investigations/Bambifan101) where merging into cases with very long histories is getting problematic. Bambifan101 now has just over 800 revisions on the main case page, which is well over the suggested limit for history merges (500) and indeed I encountered a database error trying to merge into it today. My off-the-cuff solution was to selectively restore the history in two batches, which works fine, but should we be thinking about a better solution? Something like moving part of the history off to a subpage when the case gets large enough? I'm not even really sure how that would work, just putting it out there. Ivanvector (Talk/Edits) 14:36, 14 February 2018 (UTC)
Oshwah promoted
Following discussion on the functionaries mailing list, Oshwah has been promoted to full SPI clerk. Congratulations to him. I'd also like to thank all the clerks for the work that you do on behalf of the CheckUser team. You make SPI more manageable, and we couldn't do it without you TonyBallioni (talk) 23:54, 6 July 2019 (UTC)
Hello all. In the process of doing an ongoing cleanup task, I found Wikipedia:Suspected sock puppets. I was not an active editor during the time when that page was active, so from what I'm seeing, most of those subpages (Special:PrefixIndex/Wikipedia:Suspected sock puppets) look like the precursor to this process. As an editor who has made several SPI reports in the past (and even as a editor who has had interest in being an SPI clerk ... a few times), I am very familiar with the fact that touching any part of the SPI archives is incredibly controversial. With that being said, here's my proposal: Merging the cases of Wikipedia:Suspected sock puppets into the SPI archives. I see what would have to be done to make it happen without breaking anything, and I see the potential value in merging the archives of Wikipedia:Suspected sock puppets into this page, especially the ability to more easily track down old cases in the event it is necessary, such as if a sockmaster from the pre-SPI forum says decides to pop up again; it would make the cases on them easier to locate from the main SPI page, such as with a subpage search, and in addition, keeps the cases in order, including chronological order. If this task is approved, I would be more than willing to perform the task myself and/or discuss/create/submit a workflow for the steps necessary to accomplish this task. Also, if I can do it, if I have to technically be a SPI clerk to perform this task, I am willing to undergo the training to accomplish that status. Steel1943 (talk) 23:07, 12 July 2019 (UTC)
I'm also not entirely sure what we're talking about here, but it seems like a lot of work for little benefit. None of those accounts would have any CU data, and if they've been active this long, we likely know enough about them that adding a few more accounts to the archive from years ago wouldn't help much. TonyBallioni (talk) 21:27, 16 July 2019 (UTC)
I get that, but on the same token, it seems related to a point where it could be helpful, considering the SPI site seems to not have a method at the present time to search through the archives of the page I referenced. Anywho, I'm just inquiring potential interest in this ... and I could get it done since I tend to work with archives a lot here ... but if there isn't any, no worries and I'll just be on my merry way. Steel1943 (talk) 21:46, 16 July 2019 (UTC)
Ah, I misunderstood. You just moved the conversation; you didn't do the merge. I got confused by what you said at the top. Then it's easy. Don't do the merge. Thanks.--Bbb23 (talk) 22:02, 16 July 2019 (UTC)
The sockmasters whose activity spans both SSP and SPI usually have links to the SSP archives on their SPI page. "Whatlinkshere" from a master/sock's userpage or user-talkpage could also turn up anything else, though it could be drowned in other links. If there is actually some uniquely useful information in the SSP forums/archives/subpages beyond that, I would suggest just making a custom search link for them. Absolutely no need to move thousands of pages. Someguy1221 (talk) 00:58, 17 July 2019 (UTC)
Yep, there are links to SSP already, as well as the historical page which points to all of the other archives. Sure, we could merge and sort them, but it isn't useful, as other editors have said above. --qedk (t 桜 c)06:41, 17 July 2019 (UTC)
Some IP ranges are so busy, the servers choke if you try to do a sleeper check. Other ISPs dump people randomly on a /12, making any kind of range block or sleeper check impractical (that's sixteen /16s). It's not just VPNs that can cause problems. I usually leave a note when the tool isn't much use. However, I've seen cases where the tool wasn't much use one time, but it helped the next time, so I try not to speak in absolutes ("do not endorse requests in this case because the tool is guaranteed to be useless"). Although VPNs can be a pain, other times I can still confirm a few accounts. I guess it depends on the VPN used, the technical skill of the sock puppeteer, and how paranoid they are. Zzuuzz probably knows more about dealing with annoying LTAs on proxies and VPNs than me, though. Many of the editors that I block are fairly unsophisticated. NinjaRobotPirate (talk) 02:14, 15 September 2019 (UTC)
(Not pinged, but I'll comment) Use your judgement. Any administrator can block an account for socking without CheckUser evidence. If you feel it is justified without CU, you are free to block even if an SPI filer has requested CU... Turkey I've never had problems with (people tend to use the same VPN service, so we can find sleepers, etc.)More generally, interpretation of CU results requires judgement, and some CUs are more comfortable with different VPNs, networks, regions, etc. If you think it is useful, there's no reason to not endorse. We may run it without posting results at the SPI if we aren't confident in them, and leave it for another CU to see if they can come to a different conclusion, because of familiarity with the case or with how a network works. If you want an example: I refuse to run CU on anything remotely related to Nepal because I will only return Inconclusive because of how the internet in that country works. Others don't mind running it. That's fine, and reflects acceptable judgement on all parts. TonyBallioni (talk) 02:22, 15 September 2019 (UTC)
Sir Sputnik, neither account is using a VPN or proxy of any other sort. They are not anywhere near Turkey either. One is using a sticky IP that acts like a static line with no other socking associated on that IP. The other editor appears to be hundreds of miles from that one in the same large country. They appear unrelated by locations and useragents. Does this help with the general question? — Berean Hunter(talk)03:37, 15 September 2019 (UTC)
After discussion with the functionaries team, Cabayi has been promoted to full SPI clerk. Congratulations! On behalf of the CheckUsers, thanks to Cabayi and to all the SPI clerks for the hard work and dedication they give to the SPI project. We couldn't function without you all. :-) Katietalk17:30, 23 September 2019 (UTC)
What if the master (the first registered account) is not the first blocked? It could be that G5-able articles are being overlooked.
A tool to scrape the usernames from the {{SPIarchive notice}} and {{checkuser}} templates on the SPI page and its archives, and find the earliest indef would (IMHO) be useful.
Cabayi, Interesting! I could probably work on something. It would be pretty easy to get the names from an spi or archive with a regex like: /\{\{(SPIarchive notice|checkuser)\|(1=|)(.*?)\}\}/iSQLQuery me!15:34, 20 October 2019 (UTC)
There's no previous report under that name - you can check the page history to see (also no deleted contribs). Probably the previous socks were so obvious that whoever blocked them didn't bother filing a report. Ivanvector (Talk/Edits) 13:06, 1 November 2019 (UTC)
I would prefer (1) not, because the output of {{checkuser}} is already pretty long, or (2) if we must add such a link then link it to the user's contributions which are page creations locally (example) rather than on xtools, since xtools breaks all the time. Ivanvector (Talk/Edits) 20:23, 22 December 2019 (UTC)
A couple of years ago, I offered to clerk at SPI; as far as I knew, you had to be a clerk to help out there. My offer was never accepted (although requests after mine were), and it was eventually removed without notifying me. Although I'm no longer interested in SPI, it might be helpful to other editors wishing to clerk if SPI would clarify what kind of apprenticeship is expected before becoming an "official" clerk. Miniapolis15:12, 15 February 2020 (UTC)
Well, the notice at the top does state: "Applications are reviewed by the CheckUser team periodically, and clerks are added as needed. Applications that go without comment or where the user has become less active on Wikipedia or at SPI may be removed by a CheckUser. Applicants are encouraged to actively participate in SPI cases as this will increase their chances of being selected." Sro23 (talk) 17:21, 16 February 2020 (UTC)
And it wasn't there when I volunteered. If you need help, you shouldn't turn your back on competent editors who volunteer in good faith. The instructions are still vague: "Clerk for us, and maybe we'll consider allowing you to clerk". Nonsense. Miniapolis00:20, 17 February 2020 (UTC)
Miniapolis, I went through a number of those listed as wanting to clerk and used a rule of thumb that if they had not participated in any SPI cases within the last 12 months then I removed them for inactivity. Your statement, "...as far as I knew, you had to be a clerk to help out there" seems inconsistent with your activity both before and after your clerk request. You seem to have been aware of the instructions for admin patrollers and assisted on six cases before the clerk request, two cases afterwards and one case after I removed your name. You have participated in nine cases total by my count but if I left something out then I welcome correction.
Request removed for inactivity on October 9, 2019 with summary, "removed due to inactivity at SPI, please reapply and work cases if you are still interested".
(edit conflict)If anyone can clerk at SPI (as seems to be the case, whether "helping out" or formally clerking), there should be no need to formally request clerkship. SPI seems to be quite different (and yes, vaguer) than ArbCom, where clerking is formal and structured. As I mentioned above, the acceptance of applicants after my request led me to conclude that favoritism may have been a factor. All the best, Miniapolis17:16, 4 March 2020 (UTC)
Not just anyone can clerk. Per this section, "Any user in good standing may ask to be considered for clerkship." Only clerks, checkusers and admins may close a case as you did here. The admin instructions are separate and I linked to them above. Clerking is formal and structured at SPI. It was not favoritism as far as I was concerned but rather inactivity and a lack of overall participation which is why I worded the summary with "...work cases if you are still interested". Those cases help give checkusers something upon which to form a basis of opinion. Less than ten cases isn't enough SPI experience. If you work cases then you are more likely to get picked provided that you do a good job. — Berean Hunter(talk)17:47, 4 March 2020 (UTC)
Berean Hunter, as an interested potential clerk, I'm curious - how can interested editors participate in SPI (other than filing cases) without stepping on the toes of clerks? I for one would be happy to participate in more SPI cases, but I'm plenty aware that even well-intentioned "helping" in areas like this can end up getting in the way of the clerks. creffpublica creffett franchise (talk to the boss) 17:58, 4 March 2020 (UTC)
Help filers by adding evidence as well as parsing the existing evidence to help form a conclusion if possible. Study what the clerks do and you will be able to assist them by pointing things out when new cases come up. If you observe enough cases then you will be able to assist by clarifying who the sockmasters are for new cases i.e. "I think this is really <sockmaster name> and here is why..." Actively hunting socks and filing cases as you have mentioned is a very good activity to build experience and will help clerks/checkusers form an opinion as to whether you might make a good clerk. My SPI team colleagues will likely have more suggestions. — Berean Hunter(talk)18:28, 4 March 2020 (UTC)
It is not really the CU "team" that picks up trainees but individual CUs who decide whether to train a new clerk. That decision I'm sure varies from CU to CU, but to allege favoritism is absurd. Based on Miniapolis's comments, I personally would never agree to train her, and if someone suggested it at, for example, the functionaries list, I would oppose it. Clerking is tough, and we certainly don't need attitude. Finally, training a clerk, although it varies from editor to editor, takes time and energy. We are all volunteers, and sometimes we may feel that we simply don't wish to take that time or expend that energy. I've trained clerks before, and I've either been lucky or really good at picking, but they performed in an exemplary fashion during their training, making my job easier.--Bbb23 (talk) 19:17, 4 March 2020 (UTC)
The instructions at the top of this page say that requests to be a clerk may be removed by CUs in various circumstances. Wouldn't it be preferable to archive them instead? As far as I can tell, we have "subpages" that are archives but not for this page, or at least we don't archive the clerk requests to the archives. I don't like having no record of requests except through the contribution history.--Bbb23 (talk) 17:11, 4 March 2020 (UTC)
Seems a good idea. Allows a user who wants to ask again to not have to trawl through the contributions page for this page, but instead find their old request in the archive. Also would allow a CU to more easily check for previous requests. Dreamy Jazz 🎷 talk to me | my contributions23:30, 4 March 2020 (UTC)
I'm guessing I was pinged about an archiving bot? These are such low-volume requests that you could just use OneClickArchiver or do it manually. If we want a bot, I might recommend User:ClueBot III. You can set it to archive only after {{done}}/{{not done}} (or what have you), and to look at only level 3 headings. If we need specialized archiving functionality like at WP:PERM, MusikBot might be able to help. At current my bot isn't built for general purpose. At any rate, I agree archiving clerk requests is sensible. — MusikAnimaltalk00:32, 5 March 2020 (UTC)
I envisioned archiving being done only when needed, not in any automated way. The only changes would be (a) clerk requests wouldn't remain here "forever" and (b) instead of being removed, they would be saved in an archive.--Bbb23 (talk) 00:41, 5 March 2020 (UTC)
or, like Bbb23 says, we can do it on case by base basis which can be deemed worthy for being archived; either way, not many requests come in here. Thanks for the input MusikAnimal, it is appreciated a lot :) —usernamekiran (talk)22:01, 5 March 2020 (UTC)
At first I was just going to move the case, but then I took a closer look and realized the clerk instructions were to make a separate copy. You can see why redirecting the CollegeMeltdown SPI to the Mrelving SPI might be problematic in the future. If someone wants to open another SPI case on CollegeMeltdown it just gets confusing. This has happened a few times before with other SPIs. Also you can see the note I left on the Mrelving SPI stating where it was copied from. Sro23 (talk) 15:57, 21 April 2020 (UTC)
@Sro23: You missed it in the CollegeMeltdown case btw, Moved case to Wikipedia:Sockpuppet investigations/Mrelving... might as well change that, since that misled me about the case being moved (and not copied). --qedk (t愛c)16:14, 21 April 2020 (UTC)
After consultation with the CheckUser team, I’m pleased to announce that 1997kB has been appointed a full SPI clerk. Congratulations and thanks for your hard work :) TonyBallioni (talk) 16:59, 10 May 2020 (UTC)
The Checkuser team are also pleased to announce that JJMC89 has been promoted to a full SPI clerk. We would like to extend our congratulations and a big thank you for all of the hard work that you have done. — Berean Hunter(talk)12:48, 12 May 2020 (UTC)
I've gone a ahead and cleaned up the content of the relevant pages. I don't think there's anything we can do fix the relevant editing history. it's all somewhere related to the case, so no big deal. Sir Sputnik (talk) 16:25, 18 May 2020 (UTC)
It is because there is no evidence of socking but CU is being requested. We have seen such disruptive SPIs before too by the same group of socks who had been sanctioned for filing such SPIs before. Why they are being entertained now is completely beyond me. Aman Kumar Goel(Talk)01:18, 9 June 2020 (UTC)
@Aman.kumar.goel: Clerks only request or decline CUs but the final authority for any checking is the responsibility of the individual CU, so complaining here is pointless as the decisions are not up to any clerks; if a CU feels like it's WP:FISHING they will decline to conduct one, also noting that creffett did not request a check on your account, so I don't get why you're so agitated (I'm not saying anything else because I want to WP:AGF). --qedk (t愛c)06:13, 9 June 2020 (UTC)
I believe due to the canvassing done by Aman.kumar.goel here, creffet became a bit uncomfortable and closed the case hastily without any action to avoid controversy. The reason I'm saying that is the complete reversal of opinion of creffet about the merit of the evidence. There were in fact enough evidence of behavioral overlaps that not only prompted creffet to endorse the CU twice but also made another uninvolved editor to support it. But in their closing message, creffet said there is not enough behavioral evidence to persuade me. I have, however, gathered additional evidence and asked them to reopen the case to request another checkuser with an additional account. I also believe a check on sleeper account is needed because of some valid reasons. If not sock puppetry, I believe there was enough reasons to suspect meatpuppetry (which was also supported by creffet) that might also require an action against the suspects. I'm posting this message here so that it doesn't get archived before the reopening. I.Bhardwaj (talk) 15:00, 9 June 2020 (UTC)
@I.Bhardwaj: A check was done, it was mostly "Unlikely" and that's it, behavioural evidence was inconclusive. Just let it be. --qedk (t愛c)15:06, 9 June 2020 (UTC)
I would like to have the case reopened and reviewed over the merit of evidence (including the additional part that I will post there) and see if additional actions are needed. I would also want to make it clear that I'm not complaining against anybody and creffet is doing an awesome work at wikipedia, rather it's the situation that has led to the issue. I.Bhardwaj (talk) 15:12, 9 June 2020 (UTC)
The evidence was reviewed by a clerk who also happens to be an admin and data reviewed by a checkuser. Hence it's closure and archiving. If there's off-wiki evidence, send it to functionaries, otherwise it just seems like a desire to see the user(s) blocked based on disagreements. If there are behavioral merits aside from a suspicion of sock puppetry, there are more appropriate venues for that.Praxidicae (talk) 15:14, 9 June 2020 (UTC)
And just to clarify, I'm not suggesting you take it to ANI so much as just letting it go. Just that SPI is not the place to rehash or relitigate, well, anything. Praxidicae (talk) 15:26, 9 June 2020 (UTC)
Looking for some beta testers for spihelper.js modifications
Hello fellow clerks and checkusers! I'm working on a couple of modifications to the popular spihelper.js script to add a couple new features. I'm looking for people who would be willing to try it out and give feedback on the design and whether you encounter any bugs. To use my version, you just need to disable the existing spihelper.js and install mine, you can either use script installer from User:GeneralNotability/spihelper.js and then disable your existing spihelper import, or manually comment out importScript('User:Timotheus Canens/spihelper.js'); in your common.js with two forward slashes and add importScript('User:GeneralNotability/spihelper.js');.
So far I have added functionality to request a global lock from the block menu (though I'm waffling on whether to make it its own top-level menu option instead) and am planning to add altmaster tagging functionality to the block menu as well. Comments, feature requests, and criticism are welcome! And for everyone else - I plan to send a request to Timotheus Canens to merge my changes into his master script, so I'm hoping that these changes make their way into the "official" spihelper eventually. GeneralNotability (talk) 16:15, 7 July 2020 (UTC)
GeneralNotability, would you mind adding something to the script which allows a warning to be displayed if the page has been modified since it has been loaded / the case status has changed since the page was loaded. I have "closed" a couple of closed cases because the case was closed by a CU while I was writing a closing comment. Dreamy Jazztalk to me | my contributions22:46, 7 July 2020 (UTC)
Dreamy Jazz, that's a good idea! I'll look into that - I think I can grab the page ID on load and then again when you hit submit and bail out if they're different (Twinkle does something similar but in reverse, so I can probably borrow from there) GeneralNotability (talk) 23:32, 7 July 2020 (UTC)
I've added simple edit conflict detection, so it now tracks the version of the page you saw and submits that to the API with the edit requests (which will make it bail out with the warning "editconflict" if the edit fails because of that). There's still some potential for race conditions, but it should significantly reduce the chances of us stomping each others' edits. GeneralNotability (talk) 12:51, 8 July 2020 (UTC)
Shopping list
Taking a dive through TC's talk page back to the start of 2018, here's the requests I spotted (plus a new one):
warn if the case has been modified since it was been loaded - User talk:Timotheus Canens/Archives/2020/6#spihelper.js - Dreamy Jazz ( Partly done by GN - we now grab the page revision on load and submit that with our edit requests so it will bail out if there's an edit conflict instead of stomping)
xwiki reporting - occasionally reports are filed on enwiki as if enwiki has authority over other wikis. These need to be passed to those wikis for attention or, for those wikis without checkusers, to m:SRCU - Cabayi
If you've selected multiple options (say, "Note/comment" and "Close case") which both provide text boxes, just show a single text box.
In the blocking form, provide some shortcuts like, "uncheck all", "uncheck all that are already blocked", and a way to fill the same Tag choice in for all the socks, so you don't have to pick "Suspected sock" N times. ( Partly done by GN - added select-all and fill-all options, haven't figured out "select unblocked" yet)
Provide a preview for the text box.
Not sure if this is possible, but automatically purge the page instead of redirecting you to Special:Purge.
Hi. I couldn't open the tool just now. I got an error and the SPI screen did not open. The error from the console was:
Uncaught TypeError: Cannot read property 'parse' of undefined
at spiHelper_getPostExpandSizeNew (index.php?title=User:GeneralNotability/spihelper.js&action=raw&ctype=text/javascript:1081)
at HTMLDocument.spiHelper_addLink (index.php?title=User:GeneralNotability/spihelper.js&action=raw&ctype=text/javascript:1151)
at mightThrow (load.php?lang=en-gb&modules=jquery%2Coojs-ui-core|jquery.ui&skin=vector&version=11cba:49)
at process (load.php?lang=en-gb&modules=jquery%2Coojs-ui-core|jquery.ui&skin=vector&version=11cba:49)
On narrower monitors (I'm running 1280x1024) the req lock label and checkboxes spill over the edge of the main box. To mitigate this, consider removing the word "has" from the alt master dropdown options. The options are plenty clear without it, and we can conserve a bit of width that way. Another option might be to dynamically adjust the width of the username and duration textboxes based on screen width, though that might be a bit harder to implement. Sir Sputnik (talk) 20:57, 9 July 2020 (UTC)
Another possible idea would be to do a sanity check on CU endorsements. If there's no non-stale socks listed in either the current case or the archives, put up a "are you sure?" alert, ask for confirmation, and offer a text box into which you can explain why you think CU would be useful. Maybe also a button that pre-fills the text box with, "Just check for sleepers". -- RoySmith(talk)14:13, 10 July 2020 (UTC)
Bugs and features
Since this turned into a much bigger thing than I planned (started out with "let's add a couple features" and we're currently at "rewriting large parts of it"), probably best to move feature and bug discussions off of the clerk noticeboard. If you have either, please post at User talk:GeneralNotability/spihelper.js. GeneralNotability (talk) 01:00, 11 July 2020 (UTC)
in your common.js (vector.js, whatever), each {{checkuser}} on an SPI page will get annotated with information about how the user has been tagged. "P" means puppet, "M" means master. The colored background indicates blocked/proven/confirmed. Hover over the box for details. Eventually maybe I'll show things like alt masters and users which are socks of somebody other than the master of the current case.
I wrote this mostly to speed up checking a case before archiving. Note: this just tells you what is. You're still in charge of figuring out if that's right or not. Feedback appreciated. -- RoySmith(talk)00:12, 17 August 2020 (UTC)
Clerks, checkusers, and patrolling admins, I present spihelper version 2.0.0! This has been the product of a month or so of development, initially intended as a small project to add a couple new features, and ending up as a pretty large overhaul of SPIHelper.
New features list
The following discussion has been closed. Please do not modify it.
Changed to per-section editing, so now spihelper can deal with multiple cases on the same page
Redid a bunch of workflow steps, combined the request checkuser/handle checkuser request options
Combined all textboxes into one, added a preview button
Drop-downs over the text box to insert commonly used templates
SPI-Archive now only archives closed reports
New "global lock" option on the block and tag page (to request a lock at SRG)
Support for the altmaster field in the sockpuppet template
Added "set/unset all" options to the block and tag page
Support for IP ranges (and IPv6) where single IPs were previously required
Abort edits to a case page on edit-conflict
Automatic purging of the case page after any edits
Ability to move either a section of a case or the whole thing (the latter will also handle adding an spiarchive notice and, if needed, do a histmerge. The former does a copy-paste with attribution in the edit summary)
Settings file support:
Set watchlist behavior when creating/editing case pages and archives and when tagging blocked users
Mark yourself as not a clerk (hides clerk-only options)
Option to log all SPI actions (great for trainees who are bad at filling out their clerk logs!)
Now giving releases clever nicknames for geek cred
More technical changes:
Redid a lot of calls to use the mw API instead of manually creating GET and POST requests
Using async/await for asynchronous calls, much nicer than then-chaining and lets us decide at call time whether we want to wait for the function to return
Lots of code documentation
Using ts-check to get typescript-like type checking
General code modernization and cleanup
Changed the previous iffy (and IPv4-only) regex to identify IPs, now use mw.util.isIP instead.
@GeneralNotability: I have two relatively small qualms, 1) can we get a back button to get back to the first menu and 2) I think the preview should be disabled for tabs it shows nothing on (like the "block/tag" screen). Everything else looks sleek and finally glad to be able to pick the section! --qedk (t愛c)09:19, 17 August 2020 (UTC)
QEDK, I've got a back button queued in my dev version. Not sure what you mean for 2), every tab should have a comments box at the bottom and so should have a preview button. GeneralNotability (talk) 17:06, 17 August 2020 (UTC)
Not content to let GeneralNotability have all the fun, I've been entertaining myself with some tool writing of my own. SPI Tools is a meant to automate some of the common and repetitive analyses when working through a SPI report. Please take a look at https://spi-tools.toolforge.org/. There's something of an instruction manual you can peruse, or you can just click around and explore. The tool is read-only, so there's nothing you can break by exploring. -- RoySmith(talk)03:09, 21 July 2020 (UTC)
I've installed a new version of this. There's a bunch of tweaks, but the major addition is a G5 search, which attempts to find G5-able pages related to the SPI. That's only partially complete; it comes up with a list, but the scoring is primitive at this point. -- RoySmith(talk)01:39, 27 July 2020 (UTC)
I just uploaded a new version of this. It fixes a bug which caused the active case menu to not get properly updated. It also fixes a bug in the IP range calculation code, which now doesn't barf on IP addresses with the high bit set. I've also added a little javascript shim that lets you get directly to the SPI Tools app from any SPI case page. If you install the shim in your common.js page:
I notice that DENY gets applied inconsistently. For example, in WP:Sockpuppet investigations/0lesson, the latest sock got tagged. Looking at the archives, there's a note on 23 November 2019 to stop tagging per DENY, which seems to have been largely ignored. Does it matter? I see two possible outcomes. 1) It's important, in which case the tools should start enforcing it. 2) It's not important, in which case we should just stop using it. -- RoySmith(talk)13:52, 30 August 2020 (UTC)
Personally, I have always seen tagging as a matter of judgment that is best approached on a case-by-case basis. Just because the previous sock was not tagged per WP:DENY does not necessarily mean the current sock also should not be tagged, though it does make it more likely that the new sock also should not be tagged. The reason is because we some sockmasters engage in both simple vandalism/clear trolling and more subtle sock puppetry. When deciding to tag, we are balancing the informational benefit (facilitating the identification of future socks) against the cost of potentially feeding the trolls, and it's not always so black-and-white. Mz7 (talk) 18:20, 1 September 2020 (UTC)
Mz7, I won't object to someone reverting it if people think that telling clerks to DENY is, itself, a failure to DENY. I completely agree that tagging is a judgment call, but the problem is that there's no real documentation or examples of when to tag or not tag. Since it's largely "institutional knowledge"/unwritten rules, being told "use your judgment" doesn't really help when you don't have anything solid to base that judgment on, and so adding a "don't tag this" notice will at least help people learn what sort of cases should and should not normally get tags. I think this view is supported by the fact that the three people to support this (me, RoySmith, Dreamy Jazz) happen to be the three most junior SPI trainee clerks, so we haven't developed the aforementioned institutional knowledge like other clerks and CUs have. GeneralNotability (talk) 18:47, 1 September 2020 (UTC)
Looking for institutional memory about case templates
Hello fellow clerks! I was going through Wikipedia:Sockpuppet_investigations/SPI/Indicators and there are several there I don't recognize/have never seen used. Anyone know what the following are used for?
I've used "unnecessary" (as a clerk) for a case where the reporter requested checkuser but the case was too obvious to bother, though I came to use Clerk declined with an explanation instead. The others I could guess at their use but don't have any experience with them. Ivanvector (Talk/Edits) 17:19, 12 August 2020 (UTC)
Nah, writing out what you mean to say is pretty much always fine. The templates are like optional shortcodes for the most common responses but there's no requirement to use them. Ivanvector (Talk/Edits) 19:00, 12 August 2020 (UTC)
'Thrown out / Rejected' is traditionally reserved for those special occasions when a stronger variation of 'declined' is appropriate. It basically means 'I'm not even going to read this', often combined with 'I'm probably going to block the reporter'. I recently used it here. -- zzuuzz(talk)17:46, 12 August 2020 (UTC)
I don't believe I've ever used those templates, though I think I've seen some more "old school" CU's use "completed" and "not possible" (as in CU results completed, or CU is impossible) before on rare occasion. Also, does anyone use Delisted? Sro23 (talk) 18:16, 12 August 2020 (UTC)
Is there an official policy about when to tag and when not to tag? This is a confusing area. Yesterday, for example, I got corrected in opposite directions on different cases (sorry, I don't remember the exact cases). In one, I added a tag to a blocked user, which was removed because the block was not indef. In the other, the exact opposite; I queried another clerk if they were sure the tag they had applied to a non-indef blocked user was correct and the response was, (paraphrasing) "Yes, I tag temporary blocks". So, do we tag? Do we not tag? Is it up to the individual clerk? -- RoySmith(talk)14:38, 3 November 2020 (UTC)
Given that the master tag explicitly says This account has been blocked indefinitely, it shouldn't be used when that statement is inaccurate. Sir Sputnik (talk) 14:41, 3 November 2020 (UTC)
I'm not sure I understand where your comment is coming from. The tag would apply and I didn't say otherwise. I'm also going to courtesy ping @QEDK: as the other clerk who corrected Roy on this. Sir Sputnik (talk) 15:18, 3 November 2020 (UTC)
QEDK I generally don't even bother tagging masters only blocked temporarily, it always seemed like a banner of shame to me. But if you must, the correct template to put on MosMusy's userpage is {{sockpuppeteer|checked=yes}}. Sro23 (talk) 15:41, 3 November 2020 (UTC)
Personally I don't ever not block sockpuppets indefinitely, I'm not sure why anyone would. Sro23 is right about the template for non-indeffed (or non-blocked, even) sockmasters, except only use |checked= if they have been. Using {{sockpuppeteer}} with no parameters gives "this account is suspected of abusing multiple accounts", which is right, but I don't know why you'd bother tagging a sockmaster if they're not blocked. There's also a timeblocked parameter for sockmasters with temporary blocks, which I also usually don't bother with. Actually there's a whole list of parameter examples with the messages they produce on the template's documentation page, it's pretty flexible for different situations. Ivanvector (Talk/Edits) 16:13, 3 November 2020 (UTC)
@Sro23 and Sir Sputnik: The block was made with the SPIhelper tool as was the tag, I had assumed that it was applying the correct tag which would mean including 1) checked=yes and 2) timeblocked (note that the template works but there's a formatting error). The tag is merely procedural and also accurate, whether it's shameful or not is not something clerks should think about. In any case, I'd say tagging any non-indefinitely blocked sockmasters can be left to discretion but there's nothing wrong with doing so (and we have been doing so afaik). Pinging @GeneralNotability: regarding the tool. --qedk (t愛c)16:41, 3 November 2020 (UTC)
QEDK, I inherited spihelper without timeblocked functionality and nobody's ever asked for it (and I've always been told not to tag temp-blocked users anyway). GeneralNotability (talk) 19:36, 4 November 2020 (UTC)
GeneralNotability, Maybe have it only apply a tag (even if specified) if the block time is indef, since that seems to be the policy. If somebody really wants to apply the tag, they can always do so manually. -- RoySmith(talk)19:42, 4 November 2020 (UTC)
Sockpuppets should be blocked indefinitely. The only scenario I can see where a sockpuppet wouldn't be indefinitely blocked is if the user wants to switch from his old main account to his new account. However, this wouldn't be abusive sock puppetry; rather, it would just be a poorly-executed self-rename, so there wouldn't need to be any blocks in the first place. Reaper Eternal (talk) 17:01, 3 November 2020 (UTC)
This case is being reviewed by Cabayi as part of the clerk training process. Please allow Cabayi to process the entire case without interference, and pose any questions or concerns either on Cabayi's Talk page or on this page if more appropriate.
That's not only garishly distracting, it's untrue, as the case was closed (with a block).
AIUI this is from some standard boilerplate. Can that be updated? Perhaps as a the template, with a |closed= parameter which when set to "yes", updates the wording and removes the emphasis? And can a bot then update any simalr old cases to apply that?
I don't see why it's necessary - yes, we could have the archiving clerk remove the template or the trainee remove it when closing but there's not much need of retroactive cleanup, the statement was true at the time, and that's that. --qedk (t愛c)14:02, 5 November 2020 (UTC)
What you're looking at here is an archive page. It's function is to reflect the state of the case as it stood at time of archiving. In order for it to be investigatively useful clerks have to have confidence that it hasn't been altered after the fact. As such it should not be edited, barring cases of unnoticed vandalism and the like. Sir Sputnik (talk) 17:15, 5 November 2020 (UTC)
I agree with RoySmith and Sir Sputnik here. I don't see this as a big deal, and although Cabayi is not a trainee clerk today, they were one at the time that particular SPI case was archived, and Cabayi was the clerk that asked an administrator to impose the block (see [1]), so the claim that the template is "untrue" is itself untrue. Mz7 (talk) 18:18, 5 November 2020 (UTC)
What's untrue? At the time the request for a block was made, Cabayi was a clerk trainee. Maybe you find it distracting, but to almost everyone else seeing that template is no big deal. It isn't hurting anyone so were are going to keep it there. Sro23 (talk) 21:28, 5 November 2020 (UTC)
"although Cabayi is not a trainee clerk today, they were one at the time that particular SPI case was archived, and Cabayi was the clerk that asked an administrator to impose the block (see [1]), so the claim that the template is "untrue" is itself untrue." is untrue. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits12:01, 6 November 2020 (UTC)
The tag is sometimes removed when the case is reviewed before archiving (clerks aren't supposed to both close and archive the same case) but removing it is neither a requirement nor standard practice. Nevertheless, do not edit the archives. Ivanvector (Talk/Edits) 15:22, 6 November 2020 (UTC)
GeneralNotability promoted to full clerk
I am very pleased to announce that after a discussion among checkusers on the functionaries mailing list, GeneralNotability (talk·contribs) has been promoted to a full SPI clerk. On behalf of the checkuser team, I would like to thank him and all the clerks for their work in making the SPI process smoother. Mz7 (talk) 20:02, 15 November 2020 (UTC)
I've just released a new version of spi-tools (https://spi-tools.toolforge.org/). The major new feature is the consolidated timeline, on the interactions screen. This gives you a merged chronological listing of edits (including deleted if you have admin rights), blocks, and log entries for all the accounts you've selected.
This was the original feature I set out to write when I started this, but never quite got there in my initial efforts. I've got some more cleanup I want to do, but what I've got working now seemed useful enough to shove out the door. I particularly want to work on the sock selection screen, to make it easier to select useful subsets of socks, and also to be able to look at multiple cases combined.
Hello. We received a request to lock some Bhinegar socks, which included Special:CentralAuth/Naquℹs but that one is identified as a sockpuppet Xawq instead. If both cases are about the same puppetmaster, shall they be consolidated/merged so all info about them be in one place? Courtesy ping to ST47. Thanks. —MarcoAurelio (talk) 12:24, 17 December 2020 (UTC)
@MarcoAurelio: Yes, I believe the two cases are the same. It makes sense because many of the recent accounts are actually several years old. FWIW, these are all Confirmed:
I am very pleased to announce on behalf of the checkuser team that RoySmith (talk·contribs) has been promoted to a full SPI clerk after completing his training. Also on behalf of the checkuser team, I would like to extend our thanks to all of the clerks and patrolling admins for their work in making the SPI process smoother and more efficient. Callanecc (talk • contribs • logs) 00:23, 28 December 2020 (UTC)
Hello fellow clerks, checkusers, and other clerk-noticeboard-watchers! I speak from personal experience when I say that the clerk selection process seems pretty arbitrary to an outsider (I sat on the list for the better part of a year before I got picked up after my RfA) and it's not obvious how a prospective clerk can prove their competence other than filing SPIs. To that end, I've started an essay at Wikipedia:Advice for prospective SPI clerks (a lot of the basic inspiration for this came from comments at this noticeboard by Berean Hunter and Bbb23, I've got diffs for those comments on the talk page). I'm trying to strike the right balance between "here are things you can do to be useful and show us you're competent" and "please don't just edit a bunch of SPI cases to make useless comments to 'prove' your competence." It's currently a work in progress, but additions, complaints, and general feedback would be quite welcome. GeneralNotability (talk) 16:53, 6 January 2021 (UTC)
GeneralNotability, Thanks for taking this on. My own experience was similar to yours. I found the entire process baffling, and I'm sure I committed more than one faux pas as I felt my way around.
Perhaps the most surprising thing to me was the role of non-admin clerks. When I first started patrolling SPI, I had a few notes from such clerks instructing me to block somebody. I was like, "Um, no, I don't block people just because somebody tells me to". My apologies to anybody who was on the receiving end of one of those notes. I always assumed SPI clerks were basically the same as ArbCom clerks, who took care of clerical matters but didn't actually decide cases.
I also was totally taken aback when a CU told me that I shouldn't be working so many cases as a patrolling admin and that I should be leaving them for the clerks to handle.
I suspect there's a lot of stuff like that where the people involved are so tuned into the process they don't realize what's mysterious to outsiders. Maybe it would be good to post something to WP:VP soliciting feedback from a wider audience. -- RoySmith(talk)18:30, 6 January 2021 (UTC)
@RoySmith: I believe any non-admin clerk saying that is merely recommending a block, more than often any clerk would just let it be or else leave it open for another administrator to block - this means that saying that you won't block someone is completely cool, as long as any patrolling admins understand it's not a final decision for the case. --qedk (t愛c)19:37, 6 January 2021 (UTC)
100% what QEDK said. As one of the clerks who started as a non-admin, while I recommended blocks but absolutely expected that the admins looking at my block requests would exercise their own judgement on the case. We each take responsibility for our own actions. Cabayi (talk) 21:47, 6 January 2021 (UTC)
I've been keeping a google doc with notes about cases, mostly things I don't want to put in the SPI report per WP:BEANS. I imagine most other clerks do the same. I propose we start a new wiki for SPI team members to keep these sorts of notes in a central location where all the SPI team members can have access, in a beans-compliant way. -- RoySmith(talk)15:15, 6 February 2021 (UTC)
I’ll ping Risker on this, as when similar things have been suggested before she’s been pretty involved in the discussions. One thing to note is that while CUs do have access to CU wiki, English Wikipedia CheckUsers probably utilize it the least of any major project. I think we should utilize it more than we do, but as it stands we don’t do a great job of using this resource that we already have. I’ve seen LTAs speculate off-wiki on how we have these in-depth profiles on them, when the LTA speaking doesn’t even have an entry. It’s very haphazard what gets put there.I think there’d be a few concerns: first, another place to update stuff when we have problems keeping SPI and cuwuki current as it is. Second, a proliferation of private wikis isn’t necessarily something we want. They’re hard to track, and adding another place to put private evidence doesn’t add too much confidence to the system. Third, and this is one that will probably be more controversial, but right now the biggest thing contributing to the backlog is a lack of patrolling admins. We have plenty of clerks (and you all do a great job), but getting more people willing to block for socking without thinking it takes special permission or extra layers of bureaucracy is what we need. I’m not sure this moves us towards that. TonyBallioni (talk) 15:42, 6 February 2021 (UTC)
Thanks for the ping, TonyBallioni, and thanks for the question, RoySmith. We have these centralized pages already onwiki, at WP:LTA, and this is precisely where you and other clerks would best be gathering this information. The overwhelming majority of socks are not identified by SPI clerks; they're identified by other contributors, and are either blocked without an SPI, or identified as part of the SPI. There's almost nothing that you would likely have in your notes that can't be put onwiki, or hasn't already been put onwiki by someone else. The LTA pages are the centralized point for information gathering re repeat sockers. Nothing you are gathering should be covered under the privacy policy (e.g., checkuser data, certain aspects of the Arbwiki), or some need for confidentiality (e.g., Board or committee deliberations prior to publication of decisions), which would justify a private wiki; everything that clerks are gathering or analysing is based on public contributions. I think making better use of the onwiki LTA pages would encourage more users to identify socks and bring them to SPI. I'd discourage you and other clerks from keeping too many off-wiki notes, to be honest. Even the CU wiki entries are primarily made up of data covered under the privacy policy (i.e., CU results, location notes, sometimes personal info on the sockmaster), with generally only a few notations about behaviour patterns. With the exception of the CU results, a good LTA page will have at least as much info, if not more. Risker (talk) 05:25, 8 February 2021 (UTC)
Risker, I wasn't thinking of privacy issues. I was thinking of hints you don't want to give away to socks on how to modify their behavior to avoid detection in the future. -- RoySmith(talk)15:09, 8 February 2021 (UTC)
I concur with Roy, plus creating an LTA page is exactly the opposite of WP:DENY. In addition to LTAs, I'm very hesitant to publicly document details of suspected paid editing groups (not talking about incompetent Fiverr "Wikipedia experts" evading blocks and stacking AfD, I mean groups that appear to be sophisticated reputation management firms). Storytime: there's a paid editing farm out there which I've dealt with a few times at SPI, we know them as "Yoodaba". When a couple other editors and I first found them, we wrote up a detailed behavioral explanation at SPI of why we thought they were related, including some very distinctive quirks about their editing, and I handed out blocks for the lot. Later on, other suspected socks were brought to SPI, and we actually saw a behavior shift - there were accounts which had those same quirks in their editing history and then stopped having those quirks shortly after the SPI. And that right there is why I no longer go into any significant detail on-wiki when UPE firms are involved. GeneralNotability (talk) 02:32, 9 February 2021 (UTC)
Some version of this idea gets brought up every 18 months or so. 2017ish there was a suggestion for an LTA database, 2019ish we discussed the possibility of a clerks mailing list (definitely on list, think on-wiki too), and now this discussion. I’m sure there have been more before I got involved in SPI five years ago. It’s not a bad discussion to have, but I think the reasons we’ve shied away from similar ideas in the past are strong ones.I tend to think in terms of what something new would bring us. If we’re talking about LTAs, the immediately harmful ones get blocked without CU or SPI all the time, and for the ones that try to establish good hand accounts with thousands of edits, those calls are usually made after multiple CUs have examined the data and discussed whether to block.On UPE, we’re seeing behavioural shifts because the model has changed since 2017. We now primarily encounter white label stuff from freelancers. Commercial VPNs are not easily used on en.wiki compared to how they were pre-2018, when we started blocking them more systemically, and running up big sock farms of dozens of accounts operated out of the same ad agency is much less common than it used to be. I know because I’ve checked the accounts where I’ve seen the North American ad agency hired for an article based on complaints we’ve received, and then discovered the “socks” are different people thousands of miles apart from one another in South Asia. Yes, large farms still exist, but we have successfully turned the tide compared to where we were 5 years ago and SPI is becoming less effective at dealing with that problem in part because we’re mostly dealing with different people trying to make a buck. Doesn’t mean they shouldn’t be blocked, just that an SPI wiki wouldn’t contribute much there.Finally, the other main advantage of not having a private wiki is that it keeps on-wiki review of blocks the norm rather than ArbCom review. A private wiki would open up that door, which takes block review in non-CU cases away from the community. That’s not necessarily something you want in part because on-wiki review with CUs commenting as needed tends to be quicker, and in part because having it open to the entire community to review is by it’s very nature more reflective of the sense of the community on something than a private mailing list. That’s not saying there isn’t a place for committee review (CU/OS blocks, AE, etc.) but typically both the committee and the community have been moving towards more transparency and preference for community handling of appeals since the end of BASC, and I’m not sure adding a category for review would be a positive. TonyBallioni (talk) 03:28, 9 February 2021 (UTC)
Well said Tony. Whenever this proposal gets put forth, though I do understand where the need is coming from, I never want to support it, just because I'd rather keep as many notes and evidence on-wiki as possible. Also I just don't really like the idea of having to visit a private wiki/off-wiki database. It would mean more places to keep track of and update, and more work for us. All clerks, CUs, and nearly all active administrators have email enabled if you ever need to discuss a case where BEANS is an important factor. Sro23 (talk) 04:43, 9 February 2021 (UTC)
RoySmith, I like the idea. I messed with it a little, but it's just a minor punctuation fix, removal of the TW/RW banners (for now), and I expanded WP:ASSIGN to the full page title to make it easier for a newcomer to understand what's being linked to. I just tried substing it and it seems to work fine. Blablubbs|talk21:09, 7 March 2021 (UTC)
@Cabayi: According to {{SPIpriorcases}}, split case archives moved to a subpage (e.g. Wikipedia:Sockpuppet investigations/Blablubbs/1) will be automatically detected, so I think it should be enough to cut-and-paste a good chunk of the archive to Wikipedia:Sockpuppet investigations/Nipponese Dog Calvero/Archive/1. As far as I know, all the script does is append cases below <!--- All comments go ABOVE this line, please. -->, so as long as that stays intact, everything should continue to work. The transclusion limit seems to have been hit after 10 November 2016, so I'd recommend splitting from there at the latest. I'm happy to attempt a split like that – if it does go wrong, a couple reverts and a G6 should be able to unbreak things quite easily. --Blablubbs|talk13:08, 23 May 2021 (UTC)
I've just gone ahead and done this now; seems to work fine – the archivenotice points to the old archive and the script continues to archive correctly ([5]). --Blablubbs|talk14:02, 23 May 2021 (UTC)
@Cabayi: Since I'm the one who implemented automatic archive splitting in the SPI helper scripts, I can speak to this. The way it works is that when you archive a page it checks if the size of the current case page and the /Archive page combine to be big enough to break things, and if they are, it moves the /Archive page to /Archive/N for the smallest value of N that doesn't already exist. Then it performs the archive normally. Manually splitting should have no impact on automatic splitting. Sir Sputnik (talk) 15:24, 23 May 2021 (UTC)
Requested move 23 May 2021
The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review after discussing it on the closer's talk page. No further edits should be made to this discussion.
Strong Oppose an account could be created with the username "SPI clerks" and then said account proceeds to abuse multiple accounts. As such this may be the oldest account used by this sockmaster that could be found and so a SPI report would be desired at Wikipedia:Sockpuppet investigations/SPI clerks. The reason for the SPI subpage is so that it is clear that all first level subpages of Wikipedia:Sockpuppet investigations are places to place reports. Subpages of /SPI are then clearly not places to put reports. If an account is made with the username SPI on enwiki and it abuses multiple accounts a report can be filed at Wikipedia:Sockpuppet investigations/SPI and then archived without interfering with the pages as they are subpages of /SPI. Subpages of /SPI, such as the many listed here would also need moving if page is moved which would also cause conflicts in case of a sockmaster using the oldest account with a name such as "Guidance" or "Indicators" (or variations on that) which could lead to problems. Although I know of no tools which rely on active cases only being in first level subpages, having non-case pages as first level subpages may cause issues with tools. If this was moved to that location and then a case needs to be filed at this location, the SPI clerks page would then need to be moved to another new location and wikilinks repointed as the page wouldn't be a redirect. This may be creating extra work if someone decided to create an account with the username SPI clerks. Dreamy Jazztalk to me | my contributions10:02, 24 May 2021 (UTC)
Oppose on technical grounds per Dreamy Jazz. No good reason has been given why we should break so many things just so we don't have a floating subpage, which actually isn't a problem at all. Ivanvector (Talk/Edits) 14:04, 24 May 2021 (UTC)
Oppose If it ain't broke, don't fix it. Especially when the proposed fix would break things as explained above. The problems with SPI that need fixing are that we have too many socks and not enough people working SPI cases, not that we have an internal process page with an odd title. -- RoySmith(talk)15:09, 24 May 2021 (UTC)
Oppose per everyone above. While the page hierarchy here might be slightly unusual, I don't see how that's a problem, and I don't believe "fixing" it is worth potentially introducing new issues. --Blablubbs|talk11:04, 25 May 2021 (UTC)
The discussion above is closed. Please do not modify it. Subsequent comments should be made on the appropriate discussion page. No further edits should be made to this discussion.
I just released a new version of SPI Tools. The biggest change is I've added a "Pages" button which does something similar to the Editor Interaction Tool except that it provides a quick way to apply page protection to individual pages. I'll be working on the U/I (and adding more options) shortly, but for now at least it at least lets you protect frequently abused pages with a lot fewer clicks than before. Please try it out and let me know if you spot any problems or have ideas for future improvements. -- RoySmith(talk)00:21, 14 June 2021 (UTC)
I've received notes on my talk page about two different SPI archives that were incorrectly processed when being archived. The problem looks like when a large case archive is split, an incorrect redirect gets created. @GeneralNotability: I've opened a bug report about this. Everybody else, I suggest when archiving a case, you take a look at the archive to make sure it's OK. -- RoySmith(talk)15:04, 4 July 2021 (UTC)
PS, for other folks that are helping out at SPI, if you notice a broken archive and you're not a SPI clerk or CU, please don't try to fix it yourself. Instead, just ping any of the clerks to alert them to the problem. -- RoySmith(talk)15:07, 4 July 2021 (UTC)
There is a backlog of 92 closed SPI cases awaiting archival, the oldest of which is Wikipedia:Sockpuppet_investigations/Pziders from April 23. There are also 55 cases that have been CU'd and await administrative action, 44 open cases, 13 CU requests, 3 endorsed or relisted cases, 1 case awaiting other clerk action, 4 declined CU requests, and 1 case with missing information. –LaundryPizza03 (dc̄) 22:35, 13 July 2021 (UTC)
(For what it's worth, "closed SPI cases awaiting archival" is the least important backlog anywhere on Wikipedia.) KevinL (aka L235·t·c) 16:10, 17 July 2021 (UTC)
I am pleased to announce that Tamzin has been appointed as a trainee clerk. Blablubbs and I will be supervising their training, with GeneralNotability assisting as well. Welcome to the team, Tamzin! We're looking forward to working with you. KevinL (aka L235·t·c) 05:22, 28 August 2021 (UTC)
I'm excited to get to work! :D Thank you, Kevin, for putting me forward, and thank you to the rest of the SPI team for trusting me with this.-- Tamzin[cetacean needed] (she/they)15:59, 28 August 2021 (UTC)
Reworking CU instructions at Guide to filing cases
At L235's request, I've made fairly significant changes to Wikipedia:Sockpuppet_investigations/SPI/Guide_to_filing_cases#Whether_or_not_to_request_CheckUser_in_a_case since the list of "when/when not to request checkuser" doesn't really agree with current practice (when's the last time you heard a request for CU get declined because it involved a current arbitration case? Why is CU necessary for obvious 3RR violations?). I've removed the old table and listed out what I think are the two most common cases that the average editor should be requesting CU: sleepers and complex cases ("I'm pretty sure these people are related but there's just enough wiggle room that I want CU to verify"). I've also trimmed down the no-CU section to the most common declines. I'll work on the wording some more, and might add a bit talking about factors that clerks and CUs weigh (might be a bit subjective, though...still thinking about that), but wanted to bring this to everyone's attention. Comments, improvements, etc. are all welcome, and if anyone objects to this, go ahead and revert and we can talk this over some more. GeneralNotability (talk) 01:39, 14 September 2021 (UTC)
@GeneralNotability: good work, I am still confounded with CU accepts vs. declines. I do have an improvement suggestion, modify "Suspicion of sleeper accounts" to "Suspicion of additional accounts", a often the suspicion is for additional active accounts and not just sleepers. I would also add that if the presented evidence points out to a large number of active accounts (e.g. a report with 12 new very similar accounts, in which CU was not needed to confirm), then CU would be appropriate as additional accounts are likely. If the only causee for CU was a history sleeper accounts or complex cases, we would never have the initial CU in non-complex cases.--Eostrix (🦉 hoothoot🦉)05:52, 14 September 2021 (UTC)
I'll sometimes request (i.e. self-endorse) CU for tactical reasons. If there's a long chain of confirmed socks going back to the original master and the most recent one will go stale soon, it seems useful to get CU to extend the timeline another 90 days. But maybe that doesn't need to be called out in the public-facing instructions. -- RoySmith(talk)13:33, 14 September 2021 (UTC)
@GeneralNotability and RoySmith: for what it's worth, the CU log is forever. All we have to do is click the log entry on the account to see the context of the check (example) and you can figure out the ISP/range, and usually the exact IP (though not always.) Recently confirmed socks give us UA data, but most of us are halfway decent at being able to rule out someone based on the log. Can't really confirm based on it, but ruling out is half the battle. Anyway, thought that might be relevant. Still use your discretion if you think it is useful, but wanted to paint a fuller picture. TonyBallioni (talk) 03:57, 17 September 2021 (UTC)
Thanks, that's appreciated. I think there's still value to a clerk to be able to see an unbroken chain of confirmations. When I'm comparing behavior, I'll give more weight to a comparison against a confirmed sock than against a suspected one. Sometimes in cases with extensive archives, I'll pull out a subset of the old socks (i.e. just the confirmed ones) to study in more detail. -- RoySmith(talk)13:16, 17 September 2021 (UTC)
LGTM. I added this noting that just because CU isn't requested doesn't mean we won't run it. While not common, intentionally not requesting CU is a favourite tactic of people trying to railroad their wiki-enemies when that game is being played. TonyBallioni (talk) 00:47, 17 September 2021 (UTC)
Hi folks, DuncanHill voiced a concern on my talk page following my deletion of the IP-only SPI at Wikipedia:Sockpuppet investigations/82.22.42.5 (basically, that reporting IPs belonging to that sockmaster at AIV and linking to the relevant SPI was useful). My concern is that between us not tagging IPs and the general non-memorability of IP addresses, keeping a case named for an IP isn't going to help most folks except for a couple who happen to know the case name. Perhaps we could take a page from WP:LTA and nickname cases where an anonymous editor is repeatedly block-evading; I believe we have precedent in SPI cases named for BKFIP and...Tennessee country music vandal? something like that. That was just a spitballed idea from a minute's thought; if other folks have ideas or suggestions (or even think things are fine as-is), I'd really appreciate feedback. GeneralNotability (talk) 21:36, 9 September 2021 (UTC)
I'm kind of with Duncan on this. Even if they don't have memorable names, we've got some association that these IPs constitute a cluster, or at least somebody thought they did, which could be used by some sort of automated analysis tool. It seems dumb to discard that information. It doesn't actually save anything to delete it; it's still taking up the same amount of database space, and it's actually fewer clicks to archive it than to delete it. So what's the benefit of deleting it? On the other hand... -- RoySmith(talk)21:49, 9 September 2021 (UTC)
Thanks @GeneralNotability: for raising this here. This particular case is very long-running. The edits are very obvious when you are familiar with them, but of course not all admins at AIV will have that familiarity. Having some kind of repository for the history of it is useful. I don't really mind where such a record is kept, just so long as it is kept. DuncanHill (talk) 22:01, 9 September 2021 (UTC)
All right, two people objecting is enough for me - I've restored the page. Still would like to discuss whether there's a better way to approach these. GeneralNotability (talk) 22:05, 9 September 2021 (UTC)
I generally only delete or recommend deletion for IP only SPIs if the conclusion found was that it was likely caused by their IP address being dynamic and not through any deliberate block evasion or to evade scrutiny. Once it is likely that evasion has occurred I think deletion of the SPI may be less helpful as block logs or other pages are more likely to have referenced it for evidence when blocking, warning or doing some other action.
What else to call these IP only SPI cases is difficult. Some IP LTAs do have nicknames, but giving a nickname to most or all IP only SPIs will require a more structured system as it could easily get to a point where a proposed name is already in use for a different sockmaster with similar behavior. We shouldn't reassign the name to another IP only SPI, especially if this system provides a link to the SPI under the nickname, as this would break links and/or would lead older comments referencing the nickname.
I'm not sure of a system that would work, but it could be a template which maps nicknames to SPI case page names or vice versa. Alternatively it could be just redirects to the case page from the nickname (for example "IP on my network" could be a redirect to 192.168.1.1). Ideally the case page name should be the same as the IP/account listed in the {{SPIarchive notice}} template, but this nickname could also be included alongside it. This could be reversed too. However, because account names can be nearly anything (subject to limited technical restrictions) giving a nickname which then is a physical redirect to a case may become a problem if a sockmaster's oldest account is the nickname for the IP only SPI. As such, if the system of nicknaming IP SPI case pages becomes widespread, then to avoid this there can't be a redirect in a first level subpage. Dreamy Jazztalk to me | my contributions08:31, 10 September 2021 (UTC)
We ought to leave enough of a trail to be useful. The reports generally come from users with no sight of deleted revisions or CU evidence. (Try making sense of the CU block & Sock tags left by an IP on User talk:Kapamilyaasianseries03 for example) Deleting reports (other than the purely misguided, malicious or vindictive) is not conducive to open scrutiny or to allowing users to follow the trail of evidence.
It's also worth bearing in mind the WMF's stated intention to mask IPs in future. Either we find a way of linking old SPIs to new masked IPs (breaking the mask) or effectively give IP socks a clean slate to start over. Or go for the Portuguese solution & block all IPs from editing (require an account), which would save the WMF a load of work on their masking project. Cabayi (talk) 10:59, 10 September 2021 (UTC)
@Cabayi, RoySmith, and Dreamy Jazz:} not trying to be snarky here, but who cares if an IP vandalizing is the same person? The issue with sockpuppetry is that it is a named account trying to deceive the community: either through claiming they are not blocked when they are, doing good hand-bad hand work, abusively seeking positions of community trust, editing logged out to avoid associating and account with their name in one of the areas listed by the sock policy, or something of the like. In cases of named LTAs with accounts, it is useful because we can point to a reason not to have to assume good faith. For IPs? Just block them and not worry about who they are. In my opinion, trying to tie IPs together isn't an effective use of time, and I'm plan to continue to G6 IP-only SPIs when I come across as outside of our scope. TonyBallioni (talk) 01:56, 11 September 2021 (UTC)
I delete IP SPIs as a matter of course under G6. They're pretty useless and are also, by definition, not sockpuppetry since it is just an IP vandalizing by being an IP: dynamic and it changes as people move locations. Far outside of the scope of SPI. If someone wants to create a page on an IP-only LTA for documenting stuff, they can do so at WP:LTA. But don't waste our time at SPI with it. Those pages just add to the backlog and make it less managable. I strongly support GeneralNotability's initial deletion. TonyBallioni (talk) 01:44, 11 September 2021 (UTC)
(edit conflict)The history is weirdly long for a 0-transclusion template. But I can't imagine why anyone would have been subst'ing it... and if there was a consensus at some point to deprecate it and replace with something else, there'd be a backlink from the relevant discussion. Very strange. Must be something I'm missing.-- Tamzin[cetacean needed] (she/they)01:46, 8 September 2021 (UTC)
Quite a while ago; I don't really care what happens to my old stuff. I don't remember it well, and time changes our needs and use for old stuff. ~~Ebe123~~ → report02:14, 8 September 2021 (UTC)
Is there any reason to keep {{CUDone}}, an old RfCU-related template meant to be subst'd?
I've been going around tweaking some capitalizations of "checkuser"/"Checkuser"/"CheckUser", per recent discussion at WT:CHECK. The following maintenance cats all seem due for a rename:
(SPI cases declined for CheckUser, meanwhile, has the correct capitalization.)So, questions are about these: Is there any reason to keep them at their current titles? And if not, should I send them to CfD, or just speedily move them under WP:C2A and update the relevant templates?
I don't think so as it's empty and there are endorsed cases at the moment.
I don't think there is a reason to keep it around. We don't leave talk page notices to let filing users know a CU was run, so I don't see much use for this anymore.
Bots usually rely on the categories to determine what cases are open and also what state they are in. I know this is the case for the SPI table on WP:SPI. As such, unless they are using the pageid of the category, I think moving the category to a new name may cause bots to break. As such I recommend caution about moving categories without adequate notice to bot operators. Dreamy Jazztalk to me | my contributions12:51, 23 September 2021 (UTC)
Spicy appointed trainee clerk
Hi all, I am pleased to announce that Spicy has been appointed as a trainee clerk. Spicy will be trained in a cohort with Tamzin, with Blablubbs and I supervising his training, and GeneralNotability assisting as well. Welcome to the clerk team, Spicy! We're delighted to have you and excited to work with you. Best, KevinL (aka L235·t·c) 23:36, 16 September 2021 (UTC)
@Cabayi: What are Europeans meant to do with a page that says "The Wayback Machine has not archived that URL. This page is not available on the web because page does not exist"? DuncanHill (talk) 15:29, 29 September 2021 (UTC)
Apparently it doesn't play well with archive.org either. I have a WaPo subscription. Executive summary: The US Supreme Court runs on seniority and everybody hazes the new guy. -- RoySmith(talk)17:26, 29 September 2021 (UTC)
I have only used LTA pages in SPI once or twice, and mostly for looking to see what IPs are usually used by a sockmaster if an IP is reported (which is often listed in the infobox like thing on the right (not sure of the right name)). Dreamy Jazztalk to me | my contributions22:22, 13 October 2021 (UTC)
I invite everybody on Team SPI to read User talk:Mike Peel#Wikipedia:Sockpuppet investigations/Llywelyn2000/Archive. There are some legitimate points raised there. I would urge people not to read this in the context of any particular case, but rather to distill out the legitimate suggestions about how we could improve our processes. I realize we're all on the front lines of enforcement, and sometimes that gets ugly and repetitive and frustrating. It's good to take a step back and understand how things look to people who are not in the trenches with us.
The point which hit home with me is the issue of editing the SPI archives. I get why we don't want people to do that, but we don't communicate it well. What we do is chide people when they violate a rule that they're probably not even aware exists, and often their original question or complaint gets lost in the shadow of the "don't do that again" message. This doesn't jive well with WP:ADMINACCT.
Edit filter. Nice big warning for people who are not +CU or on a special list (the clerks list) that says "no, really, don't edit the archive". Seems straightforward enough to me. We do have a couple of filters that watch SPI archives, but those are mostly targeted at LTAs (I have 809 set to ping me on IRC whenever it hits, for example). GeneralNotability (talk) 20:15, 11 October 2021 (UTC)
I feel old and tired now - thanks all. IRC seems to be dead (mostly on telegram now?) - no idea how you still get pings there. Still not sure what SPI refers to. THanks. Mike Peel (talk) 20:20, 11 October 2021 (UTC)
My filter-fu is pretty weak. Is there a way to present the user with a useful message such as "Editing the archives is reserved for members of the SPI team, please address any questions to ..." -- RoySmith(talk)21:10, 11 October 2021 (UTC)
RoySmith, there definitely is...but I happen to know that you're hitting a different filter ;) Your testing alt hit 809, "possible SPI disruption," which covers a wide range of disruptive SPI behaviors performed by socks and LTAs, and I assume part of the problem is that the testing alt has a very low edit count and is editing the archives. 'zin is short for Tamzin did the same and got blocked for her troubles. 1170 currently doesn't have a message. GeneralNotability (talk) 21:43, 11 October 2021 (UTC)
Mike Peel, with due respect, what on earth are you talking about? The only person I see not being serious here is you: log-10 only of course. Wouldn't want to see you go exponential. Everything else has been testing a filter specifically to help address the confusion you encountered having about editing SPI archives. (And by my count, you have more edits than either Roy or I do, so I have no idea what the 300k edits crack was about). GeneralNotability (talk) 22:47, 11 October 2021 (UTC)
I took a look through the discussion, and personally, I feel like the confusion here is a consequence of the unusually complicated nature of the SPI case in question. Most SPI cases are relatively straightforward and easy-to-understand, even to outsiders: they usually consist of succinct evidence that someone is abusively using multiple accounts and a brief statement from an administrator blocking based on the evidence. However, this particular case was unusually drawn-out; it looks very similar to an WP:ANI discussion, so someone unfamiliar with SPI might've expected an ANI-like closure (i.e. with a closing statement clearly displayed within an {{archive top}} template) before the discussion is archived. In other words, I feel that the opacity of this particular case is a one-off thing and not reflective of how SPI usually operates. Mz7 (talk) 07:49, 12 October 2021 (UTC)
It's worth clarifying the most significant difference between SPI and forums like WP:AfD. It's more fundamental than the use or lack of template-formattted closing statements. In AfD, the people in the discussion are expressing their opinions. The job of the closer is to summarize the group's opinions and figure out if any one opinion gained consensus. The AfD closer should not be expressing an opinion of their own. That's why AfD (AN, DRV, XfD, RFC, etc) closes usually include phrases like most people felt ..., or a minority view was .... At AN, it's often more explicit: The community has imposed ....
SPI is fundamentally different. There's no discussion per-se (although sometimes it degenerates into that). All of the stuff prior to the close is people providing evidence. Sometimes that even includes a checkuser reporting on the result of their investigation. The job of the closer is to look at all the evidence and make a decision, which they then own. This distinction can be seen in the first-person wording typically used: "I am blocking and tagging ...", or "I am closing this with no action taken".
This distinction is noted in the instructions for the two processes. WP:AFD says: Articles for deletion ... proceeds based on community consensus. Contrast with WP:SPI: Investigations are conducted by a clerk or an administrator, who will ... determine whether they are probably connected. If we were looking for a (US-centric) judicial parallel, it's the difference between a judge presiding over a jury trial vs a judge acting as a tribunal.
I think part of the confusion is our odd use of the word "clerk" at SPI. In general, clerks don't make decisions; they help move processes along, with the actual decision-making authority resting elsewhere. At AfD, the role of the closer is basically clerical, and I've often used exactly that phrase when explaining an AfD close I did to somebody who asked about it. The role of an Arbcom clerk also fits this pattern; the arbitrators make the decisions, the clerks just manage the virtual paperwork to support the arbitrators. The introduction to WP:SPI/C makes it sound like SPI clerks are much the same (assist the CheckUsers and community by managing the day-to-day running and housekeeping tasks but in reality they are deciding cases. -- RoySmith(talk)15:22, 12 October 2021 (UTC)
Except we're not the only ones who can decide cases. The only things clerks can do that normal admins can't are inherently clerical, off the top of my head: complex case-management activities like merges, endorsing CU, archiving cases, and telling people that they are no longer welcome at SPI. "Decide cases" isn't on that list, and any admin can act on a case at their own discretion. GeneralNotability (talk) 15:28, 12 October 2021 (UTC)
True. Just making the point that the decision process isn't restricted to clerks, though I guess it's restricted to (admin || clerk). It's weird, I agree. GeneralNotability (talk) 16:28, 12 October 2021 (UTC)
(Non-admin clerk comment) There's a general principle on this project that people should not close discussions that they lack the technical ability to enforce. However, certain specialized venues have long made exceptions. TfD lets non-admins close as delete. RM lets non-admins/non-pagemovers make closes that will require G6s, suppressredirects, etc. CP clerks (who, unlike other clerk teams, are non-admins by definition) can request deletion with (as I understand it) a similar level of deference as non-admin SPI clerks. Non-admin ArbCom clerks can request blocks to enforce arbspace pagebans. I think the unifying thing we see across most of these venues is situations where adminship does not necessarily correlate with knowledge of a specialized policy area. The thing that I suppose makes SPI (and ArbCom) clerks unique is there being situations where a non-admin clerk's authority exceeds that of a non-clerk admin... but ultimately an endorse is more a convenience matter for the CU team than an administrative action. -- Tamzin[cetacean needed] (she/they)18:35, 12 October 2021 (UTC)
@GeneralNotability I've taken a first whack at an essay: User:RoySmith/essays/Please don't edit the archives. Folks should feel free to make improvements. Once your filter work is done, I think this should get moved into WP space somewhere, and your filter message should read something like, "Editing the SPI archives is restricted to whatever. See (link to the essay) for more details." I think that'll both head off people editing where they shouldn't, and provide a more user-friendly experience. -- RoySmith(talk)01:40, 14 October 2021 (UTC)
@RoySmith: I'd consider changing the title. Without context, I'd expect a page called "Please don't edit the archives" to be about why archived cases shouldn't be edited after the fact, not about the archiving process itself. I might suggest "Please leave archiving to the clerks", or something to that effect. Sir Sputnik (talk) 02:32, 14 October 2021 (UTC)
Hmmm, I need to think on that. The impetus for writing this was that people do indeed edit the archives after the fact. So, I think the title is right, but maybe the text needs to be updated to match the title? -- RoySmith(talk)02:45, 14 October 2021 (UTC)
@Sir Sputnik I've made some changes to make this all about why we want people not to edit, and got rid of most of the stuff about the archiving process. Good suggestion, thanks. -- RoySmith(talk)16:05, 14 October 2021 (UTC)