Open main menu

Wikidata:Administrators' noticeboard

Administrators' noticeboard
This is a noticeboard for matters requiring administrator attention. IRC channel: #wikidata connect
On this page, old discussions are archived. An overview of all archives can be found at this page's archive index. The current archive is located at 2019/10.

Requests for deletions

high

100 open requests for deletions.

Requests for unblock

low

1 open request for unblock.


I've been told to "go away"Edit

I recently merged Q67576167 into Robert M Graham (Q64684938). I then noticed a duplicate statement; as well as employer (P108) with a value of Victor Chang Cardiac Research Institute (Q7925795), cited to the subject's ORCID iD (P496) record, there was a second statement with same value, but uncited. Of course, I removed the later. I did not envisage for a moment that doing so would be challenged.

I was reverted by User:GerardM, with the edit summary "info orcid.org". I removed the duplicate statement a second time, reverting with an edit summary of "duplicate statement".

I have now been reverted a second time, which has restored the duplicate, uncited and redundant statement. This is bad enough, and the duplicate should be removed for a third and hopefully final time, but the edit summary in that - again, uncited - revert was "there is a source so go away".

I do not believe that addressing a fellow editor in this manner is acceptable; and the choice of words leads me to conclude that attempting further dialogue with GerardM (including notifying him of this discussion; therefore please will whoever responds as an admin do so) would be futile. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:25, 7 October 2019 (UTC)

  • @Pigsonthewing, GerardM: I removed the redundant claim that was missing qualifiers and references. GerardM, did you actually look at the item in the UI before reverting here? ArthurPSmith (talk) 18:32, 7 October 2019 (UTC)
    • Yes it says so on the ORCiD information of the person involved. GerardM (talk) 05:16, 8 October 2019 (UTC)
      • Nobody is disagreeing that ORCID provided the source of that information. The question is whether you noticed that the item had TWO entries for the same employer after you reverted, and the duplicate you were restoring was missing qualifier and reference information. ArthurPSmith (talk) 14:25, 8 October 2019 (UTC)
        • @GerardM: Do you have an answer to Arthur's question from 14:25 on the 8th of October? Mahir256 (talk) 15:14, 14 October 2019 (UTC)
          • No I did not. In my opinion, the way to handle such situations is to remove the extra entry, not to revert it. Not only is it more work I do consider it rude. PS I no longer follow chats as intensely as I used to do. I do not have the time nor the energy. GerardM (talk) 18:57, 14 October 2019 (UTC)
            • @GerardM: I am not following your argument. Andy's initial edit after his merge action, the edit which you reverted, was this one, which appears to be exactly what you are asking for, a removal of the extra entry, not reverting it. ???? ArthurPSmith (talk) 16:56, 15 October 2019 (UTC)

A week has passed; and the conclusion that can be drawn is apparently that telling another contributor to "go away" - and doing so when in the wrong - is unworthy of admin comment, let alone sanction. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:52, 14 October 2019 (UTC)

Persistent vandalism on China politics-related itemsEdit

During the past several months, multiple IP users: 220.169.0.0/16 (talkcontribslogs)、106.18.0.0/16 (talkcontribslogs)、124.228.0.0/16 (talkcontribslogs)、113.218.0.0/16 (talkcontribslogs)、61.187.0.0/16 (talkcontribslogs)、118.254.0.0/16 (talkcontribslogs) (banned for 2 weeks), which are all located in Hengyang, Hunan, China according to Geolocation, have been persistently vandalizing WikiData items related to China politics. These IP users share similar vandalism patterns. Typically they add non-neutral words usually expressing extreme political views (appearing to hold a nationalist stance like w:50 Cent Party or just seek attention) and leave insulting comments when reverting others' edits. In most cases, these edits are made through mobile-web. It is reasonable to suppose that all of these IP users are operated by a single person or a small group of persons. It is obvious that those (or that) user(s) is/are totally unreasonable and it is not possible to communicate with at all. For example:

In view of the fact that those IPs are not used by other kinds of edits for more than one year, I think it is fine to ban those IPs for at least several months to mitigate potential vandalism in the future. 虹易 (talk) 00:35, 9 October 2019 (UTC)

Possibly relevant: https://www.bbc.co.uk/news/technology-49921173 Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:08, 9 October 2019 (UTC)
@Pigsonthewing: Thanks for the information. But as far as I can see, no indication that there are coordinated disruptive edits backed by some state involving these articles/items so far. At least the above edits appear not to be so. Such kinds of disruptive edits have been lasting for years, dating back to early years of Chinese Wikipedia. It is possible that some people are encouraged or instigated by the recent political propaganda of the Beijing Government. But similar political propaganda has been also persistent for even decades, with peaks and troughs. Anyway, these problems are worth attention so that to keep Chinese Wikipedia (possibly as well as other WP in the future) neutral as it should be. (Off-Topic) Some recent news reports bring the risk of driving internal conflicts in the Wikipedia community. It is dangerous to raise dissatisfaction with other members of the entire community within a small group, which, unfortunately, has been normal in some regional groups (typically some Mainland China groups, not referring to that/those in the news). What's more, the benefits of "seeking solutions" from the public or politics outside the community are no more than bad, in my opinion. 虹易 (talk) 01:43, 10 October 2019 (UTC)

Request Semi-protection for Q5226214Edit

Merging this thread as it seems to be the same topic. Bovlb (talk) 23:53, 10 October 2019 (UTC)

Reason: IP user and new user vandalism. Thank you! --SCP-2000 (talk) 14:03, 10 October 2019 (UTC)

  Done Semi-protected for a month. I'll keep an eye on some of those IPs. Bovlb (talk) 14:08, 10 October 2019 (UTC)
I protected four additional items. There seems to be a lot of activity in Chinese labels, descriptions, and aliases on topics related to Hong Kong and Taiwan. Bovlb (talk) 16:49, 10 October 2019 (UTC)

In addition to the IP ranges listed above, I see multiple problematic edits from the following:

Bovlb (talk) 16:00, 11 October 2019 (UTC)

I listed those because it seems that MediaWiki only suppport /16. According to IP w:Whois, some are actually /12.虹易 (talk) 02:32, 12 October 2019 (UTC)

Here is another problematic IP 38.100.229.84 (talkcontribslogs). BTW I currently maintain a Telegram group to monitor Chinese-related label/description/aliases edits and anyone interested may pm me (@WhitePhosphorus) via Telegram (plz inform me your wikimedia username) I'll invite you to that group. I'll work on corresponding IRC channel as well if needed. --WhitePhosphorus (talk) 04:34, 18 October 2019 (UTC)

Blocked for a month, but I don't see anything related under the /16 there. Bovlb (talk) 13:58, 18 October 2019 (UTC)

QuickStatmentsBot and unattributed editsEdit

QuickStatementsBot (talkcontribslogs)

@Magnus_Manske

I'd like some assistance in dealing with a chronic issue with the QuickStatementsBot.

As I understand it, this is not a bot in the usual sense because it has no autonomous code to generate changes; instead, it merely provides a batch service to proxy edits for other users via various tools. This means that, unlike typical bots, we do not hold the bot operator responsible for the edits, but instead expect to find the responsible editor identified in the edit summary (see RFP).

Unfortunately, this bot has a long-standing problem whereby some edits have no editor indicated in the edit summary. I believe that this is a serious problem because, in the cases where an error is made, we cannot track down who made the error and, as a result, they may never find out. This deprives us of a crucial opportunity for process improvement. Some examples of unattributed bad edits: Special:Diff/884168986, Special:Diff/968618496, Special:Diff/972630488, Special:Diff/1011479498, Special:Diff/1011360190. For the avoidance of doubt, the vast majority of edits made by this bot (whether attributed or not) seem to be good edits.

This issue has been repeatedly raised with the bot maintainer by myself and others (Topic:V2fzk650ojg2n6l1, Topic:V85eg51k9dnxaqb1, Topic:V8th0j0qmza2yykx, Topic:Uwfg7rq8lq84kiv6, Topic:V64n8o927s1gmcr9) with limited response. I am at a loss on how to proceed here. I have attempted to work with the bot maintainer to resolve this, but what I have been doing has obviously not been working. I don't think it's appropriate to escalate this by blocking the bot. Does anyone have any suggestions?

Cheers, Bovlb (talk) 02:54, 10 October 2019 (UTC)

I believe these were done under the QuickStatementsBot account by the SourceMD service, also run by me. SourceMD (especially the batch mode that cause these edits) is currently inactive, as I don't have the bandwidth to fix it properly right now. --Magnus Manske (talk) 10:25, 10 October 2019 (UTC)
The most recent edit by QuickStatementsBot has the edit summary ‎Added qualifier: series ordinal (P1545): 2, #quickstatements; invoked by SourceMD:ORCIDator and is dated "2019-10-10T08:07:01". When did the "no user name in edit summary" code become inactive? Bovlb (talk) 14:11, 10 October 2019 (UTC)
@Bovlb: I agree this is an issue. I think QuickStatementsBot should not be used at all, it has been a long-standing loophole in the bot approval process that we should never have accepted in the first place. All edits should be made directly under the account of the user responsible for the action. − Pintoch (talk) 15:46, 10 October 2019 (UTC)
I once raised the issue in 2018 (this is really a security issue - anyone, includes anonymous and blocked users, can do disruptive edits without being discovered), but get no answers. I think we should take down new_resolve_authors tools in favor of author-disambiguator.--GZWDer (talk) 17:58, 10 October 2019 (UTC)
Where and what is "author-disambiguator"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:03, 10 October 2019 (UTC)
toollabs:author-disambiguator.--GZWDer (talk) 10:41, 11 October 2019 (UTC)
Great! If this alternative is available, should we simply block QuickStatementsBot then? − Pintoch (talk) 13:31, 11 October 2019 (UTC)
It's only an alternative for "new_resolve_authors" at the moment; and it actually uses Quickstatements for the edits (right now); batches may end up being run by Quickstatementsbot (but should be attributed to a user). ArthurPSmith (talk) 17:01, 11 October 2019 (UTC)
Looking at QuickStatementsBot's recent contributions, it doesn't look like any of them are attributed to users. This may be related to this change whereby the QuickStatements batch mode no longer uses QuickStatementsBot. I don't want to be legalistic about the RFP because circumstances change and remits expand, but is there a different RFP I should be looking at? Bovlb (talk) 18:19, 11 October 2019 (UTC)
@Magnus_Manske: It seems to me like the bot is currently operating outside of it's scope and that it might be the best choice to block it as long as it's in it's current state. ChristianKl❫ 08:18, 15 October 2019 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── There are plenty of bots which clearly operate outside of their approved scopes; many operators collect the botflag with an initial task, and later add completely different tasks without asking for approval. Not cool, but this is the reality. It would be odd if we now used the finding "edits outside of its scope" as a blocking reason when such behavior is pretty much the informal standard here.
I suggest to not block QuickStatementsBot as long as its edits are sane. When that is not the case any longer, the bot account would receive a block until User:Magnus Manske as its operator has cleaned-up. Magnus meanwhile knows that this situation is kind of problematic, and he has time now to implement a cleaner method to save the edits than using QSbot. —MisterSynergy (talk) 08:44, 15 October 2019 (UTC)

Agree with MisterSynergy that it may be excessive to hold a bot operator to the letter of the RFP, but the key point here is not about whether the edits are outside of scope or sane, but the fact that these edits are (apparently) proxied, without indicating the initiating user. No-one expects either a (conventional) bot or a user to have a 100% success rate, but it is important that people can learn from their goofs.
So, some specific questions for @Magnus Manske:
  1. When we see unattributed edits from QuickStatementsBot (e.g. Special:Diff/1032504936), are these edits made on behalf of a user, or are they the responsibility of the Botop?
  2. If they are made on behalf of a user, why is the user not indicated in the edit summary?
  3. If they are made on behalf of a user, do you have a plan for indicating the user in the edit summary?
Cheers, Bovlb (talk) 14:09, 15 October 2019 (UTC)
@MisterSynergy: Wikidata is about finding consensus. When bots are going a bit over their approved scope to make edits that are uncontroversial I don't see a reason to ban the bot about it. If it's however used to make the bot engage in edits that are controversial and for which the bot wouldn't get an approval when that's sought such as unattributed automated edits, I see that as a ground for stopping that behavior. ChristianKl❫ 19:05, 18 October 2019 (UTC)

Special:Contributions/Rii'jeg'fkep'cEdit

Vandalism. -- 05:35, 11 October 2019 (UTC)

Could you please give more details? I see a few items on which you have reverted their addition of Chinese labels, but I'm having trouble determining why you are describing these contributions with such a strong word. Bovlb (talk) 18:50, 11 October 2019 (UTC)
@予弦: Bovlb (talk) 19:24, 11 October 2019 (UTC)
@Bovlb: Lables and aliases vandalism.
  1. User:Rii'jeg'fkep'c (a native Chinese speaker) added "理勾" (transliteration of "LIGO") and "理勾引力波干涉天文台" (The correct Chinese translation is "激光干涉引力波天文台".) in Laser Interferometer Gravitational-Wave Observatory (Q255371), "伽林斯坦" (Transliteration of "Galinstan". The correct Chinese translation is "镓铟锡合金 / 鎵銦錫合金".) in Galinstan (Q675176), "阿美理卡" (transliteration of "America") in United States of America (Q30), "甫比阿" (transliteration of "phobia") in phobia (Q175854), "欧美伽" (transliteration of "omega") in Ω (Q9890), "美伽)" (transliteration of "mega") in mega (Q107205), "纳萨"(transliteration of "NASA") in National Aeronautics and Space Administration (Q23548), and "阿尼美" (Transliteration of "アニメ". The correct Chinese translation is "日本动画 / 日本動畫".) in anime (Q1107). All these transliterations are not used in Chinese.
  2. They also added "理[ㄍ区]" in Laser Interferometer Gravitational-Wave Observatory (Q255371), "伽林ㄙ坦" in Galinstan (Q675176), "赛[内区]甫比阿" in cynophobia (Q38579), "比[ㄍ矣]鲁" in beagle (Q21102), "尼侯尼吾ㄇ" in nihonium (Q1301), "雅伊ㄊ" in Yeot (Q1144106), "阿伊[不区], AIBO, aibo, アイボ, 机器犬阿伊[不区]" in AIBO (Q403438), "[ㄊ矣]ㄇ普拉" and "[ㄊ恩]普拉" in tempura (Q328709), and "戴森ス费ㄌ" (the correct translation is "戴森球".) in Dyson sphere (Q469). ㄍ, ㄙ and ㄌ are w:Bopomofo, which is used for phonetic transcription only. And ス is a Japanese w:kana. Some descriptions are also incorrect.
  3. They also added incorrect aliases in (Q40709), (Q40725) ,(Q40481), (Q40488), (Q40582) and (Q40454). -- 04:54, 12 October 2019 (UTC)
If I understand what you're saying correctly, you are raising three issues here:
  • Chinese-language aliases that are transliterations of English words (or initialisms) rather than being expressions used by Chinese speakers.
  • Using Roman characters in Chinese-language aliases for Japanese characters.
  • Some issue with descriptions that you have yet to explain.
From what I have heard so far, this seems like, at worst, a good faith editor with a misunderstanding of what makes a good alias. Also, I'm having trouble seeing where you raised this issue directly with them before bringing it to this noticeboard. Their talk page was a redlink until after your last post here, when you gave them a vague warning without drawing their attention to this thread. My recommendation is that you attempt to engage this editor directly, welcome them to the project, and give them an explanation of what you think they could be doing better and why. Coming here with claims of "vandalism" seems more likely to turn them against the project than to improve their editing. Cheers, Bovlb (talk) 17:21, 12 October 2019 (UTC)
@Bovlb: If such transliterations are really needed as "useful informations", then the correct way is to add properties transliteration (P2440) and their friends, mixing transliterations and main label names are just duck test (Q1324171)-based vandalism, so just block Rii'jeg'fkep'c, please. --Liuxinyu970226 (talk) 00:01, 15 October 2019 (UTC)
@Liuxinyu970226: I am not claiming that these transliteration make for good aliases. My point is that the contributions as described, while they may be misguided, are nevertheless likely good faith. My view is that the correct response would be to engage the editor, welcoming them to the project and explaining how their contributions might be improved. Coming to this board instead and describing their contributions as "vandalism" is poor behaviour because it fails to create a welcoming and collegial editorial environment. I admonish both 予弦 and Liuxinyu970226 not to describe edits as vandalism unless they are prepared to substantiate the claims. Cheers, Bovlb (talk) 03:39, 15 October 2019 (UTC)

Close as keep and block pleaseEdit

I have just dealt with an ugly argument between two users who both need a sanction of some sort I believe by hat boxing (I think that's the right term) the worst of it. The issue is here. The IP 2001 is not listening and should be soft range blocked for a period forcing him to make an account by other means. Quakewoody has been handling it very badly to the point that a block should be done there as well, although the IP is the real issue. The deletion request should be closed as keep. TLPG (talk) 03:24, 14 October 2019 (UTC)

Blocked the /64 due mainly to the Aussie's last message. It's disheartening to see few other admins having a look at the RfD backlog at the moment. Mahir256 (talk) 03:44, 14 October 2019 (UTC)
"forcing him to make an account" We have no policy of requiring editors to create accounts. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:30, 14 October 2019 (UTC)

2600:387:0:805:0:0:0:70Edit

Please block User:2600:387:0:805:0:0:0:70. See Special:Contributions/2600:387:0:805:0:0:0:70 - vandalism after warning. Thanks, --DannyS712 (talk) 18:25, 16 October 2019 (UTC)

  Done by Ymblanter. Esteban16 (talk) 23:33, 16 October 2019 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. DannyS712 (talk) 16:51, 17 October 2019 (UTC)

Mahendra BhimteEdit

Vandalism-only account. — Mike Novikoff 02:20, 17 October 2019 (UTC)

  Done--Ymblanter (talk) 19:04, 17 October 2019 (UTC)

Blockage of 24.37.129.34Edit

Each time 24.37.129.34 (talkcontribslogs) intervene, it is to vandalize an item. This IP has been warned before and should be blocked. Pierre cb (talk) 16:47, 17 October 2019 (UTC)

They have been already blocked three time, I have blocked now for 5 years--Ymblanter (talk) 19:06, 17 October 2019 (UTC)

LTA abuse filterEdit

Hi. At meta we have a reasonable abuse filter for managing this LTA, and we have made it quite specific to their edits. Happy for someone to contact me, or someone with rights to meta abuse filters to have a look.  — billinghurst sDrewth 23:13, 17 October 2019 (UTC)

Blockage or warning - Visite fortuitement prolongéeEdit

Hello. As you can see it here or here, user:Visite fortuitement prolongée harassed some users about fr-sysops actions. She's blocked indefinitely on wikipedia-fr. Sammyday (talk) 14:19, 18 October 2019 (UTC)

Block requestEdit

Hi, could you please provide some rest to 187.189.93.237 (talkcontribslogs) for his work on James Prescott Joule (Q8962)? Many thanks in advance, --—d—n—f (talk) 19:43, 18 October 2019 (UTC)