Wikidata:Project chat/Archive/2020/12

This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

MediaWiki:Villagepump-url/sv

Could someone recreate MediaWiki:Villagepump-url/sv with the text Wikidata:Bybrunnen and MediaWiki:Villagepump/sv with the text Bybrunnen as they have been re-established? --Sabelöga (talk) 09:53, 29 November 2020 (UTC)

• As far as I can see the page isn't protected. Is there anything stopping you from this? ChristianKl ❪✉❫ 13:16, 29 November 2020 (UTC)

@ChristianKl: That's strange, regardless, I can't create them. --Sabelöga (talk) 13:59, 29 November 2020 (UTC)

I have no undeleted it. Is that enough? ChristianKl ❪✉❫ 14:04, 29 November 2020 (UTC)

@ChristianKl:Thank you, it was also MediaWiki:Villagepump/sv as stated above. --Sabelöga (talk) 14:05, 29 November 2020 (UTC)

@ChristianKl: Any problems? --Sabelöga (talk) 14:25, 29 November 2020 (UTC)

I undeleted that as well. ChristianKl ❪✉❫ 21:45, 29 November 2020 (UTC)

Thank you very much! :D On another note that's really not that pressing, is that the page's content language could be changed on Wikidata:Bybrunnen to Swedish. But I'm not too sure what that does so no stress. --Sabelöga (talk) 04:05, 30 November 2020 (UTC)

I did change the language to Swedish. ChristianKl ❪✉❫ 12:02, 30 November 2020 (UTC)

Thanks alot! --Sabelöga (talk) 08:16, 1 December 2020 (UTC)

Notification of request for Oversight access

Hi. As instructed, I'm notifying the community of my request for Oversight access: Wikidata:Requests for permissions/Oversight/DannyS712. Thanks, --DannyS712 (talk) 03:19, 1 December 2020 (UTC)

Merge

Main article: https://en.wikipedia.org/wiki/180-degree_rule Shall be linked with it: https://de.wikipedia.org/wiki/Achsensprung_(Film)  – The preceding unsigned comment was added by 188.96.224.152 (talk • contribs) at 13:56, 10 November 2020‎ (UTC).

When do we use Start/End time qualifier and for which sources of information it is applicable?

Hi Community

Hope you are doing well!

I'm looking for your guidance regarding the start/end time qualifier that we have to include when an old website is switched to a new one, for example. However, I was wondering should we use this qualifier also for Name, Address, Phone number and etc? I would really appreciate it if you could explain more thoroughly when and where we should use Start/End time so I could follow strictly Wikidata's guidelines on this matter.

Thank you in advance!

--Valgetova (talk) 13:54, 30 November 2020 (UTC)

• It's applicable whenever information changes over time (was true at one point in time and not at another) and it's still the same item. ChristianKl ❪✉❫ 14:00, 30 November 2020 (UTC)

• Thank you for your kind response User:ChristianKl My last question is if this start/end time qualifier is mandatory to be added for every edit that we make in Wikidata? For example, if I change business name am I obliged to add this qualifier or it depends on my desire to include this information? Thank you once again! --Valgetova (talk) 06:42, 1 December 2020 (UTC)
  • It's against Wikidata's rules to delete valid data. If at some point in time a business was named X and has a official name (P1448) statement or a statement with another subproperty of name (P2561), that's name is not allowed to be removed. You can add a new name without adding qualifiers but then it's not clear to a user whether the old name still applies if you don't add qualifiers to the old name. ChristianKl ❪✉❫ 12:19, 1 December 2020 (UTC)

Translation of a title

Hello everyone,

Could someone translate the title of this page either in English (international language) or in Norwegian (language that is concerned). I can't do it by myself. I know there's only one article (from the Walloon Wikipedia), but I intend to translate this article in French and in Dutch.

Sincerely yours,

Èl-Gueuye-Noere

Model de-transition

I have created Keira Bell (Q103120340) but I dont know how to model a person who was born female, transitioned to male and then de-transitioned to female. My guess is that at least she should be recorded as female right now. Is there a consensus on how do this? Or should it even be recorded? Maybe significant event (P793) would be the right approach? I also could not find an example here on how to do this. --Hannes Röst (talk) 15:43, 1 December 2020 (UTC)

• I would use start time (P580)/end time (P582) to specify the time when the individual genders where true and mark the current gender as preferred value. ChristianKl ❪✉❫ 16:02, 1 December 2020 (UTC)
• end cause (P1534) may also be of use. --SilentSpike (talk) 11:20, 2 December 2020 (UTC)

Olympics

At Oscar Kreuzer (Q461349) we have "participant in=1912 Summer Olympics" and "participant in=tennis at the 1912 Summer Olympics – men's outdoor singles". Looking at various Olympic entries we have a mix "participant in=YYYY Olympics" and "participant in=YYYY event at the Olympics" with some containing one or the other, and some with both. What do we want as our standard? I was assuming once we had event, the "participant in=YYYY Olympics" would be redundant. But I think a bot may be involved adding "participant in=YYYY Olympics" based on a category at Wikipedia. --RAN (talk) 21:05, 2 December 2020 (UTC)

We have items for each and every event of the modern Olympics (example: tennis at the 1912 Summer Olympics – men's outdoor singles (Q2314432)) and they are all properly connected via has part(s) (P527)/part of (P361) with the Olympic discipline event (example: tennis at the 1912 Summer Olympics (Q855399)) and further via has part(s) (P527)/part of (P361) to the Olympic Games edition (example: 1912 Summer Olympics (Q8118)). I spent quite some time last winter to get this pretty complete when the Sports-Reference.com website as the most reliable source was about to be closed and it was unclear at that time whether there would be an appropriate successor. For participant in (P1344), the most specific item should be used, so please do "participant in=tennis at the 1912 Summer Olympics – men's outdoor singles"; it could be qualified with further related information, such as the final ranking in that event using ranking (P1352) and some other things. Mind that this is also pretty much the level of detail that one would add to an infobox in Wikipedia, which means that this data structure aims to be most useful for data use by Wikipedias. The "participant in=1912 Summer Olympics" claims have in most cases been added some years ago when event-level items were often missing, and they should probably be removed in order to avoid redundancy. —MisterSynergy (talk) 21:26, 2 December 2020 (UTC)

Kip Noll (Q390033)

El Gaafary has reinserted claims of birth date and death date, citing probably unreliable sources, like IMDB, and (possibly) using original research. The person's rumored death has been debated at Egyptian Arabian Wikipedia talk page and en:Talk:Kip Noll. At enwiki, the death rumor has been disregarded per core content policies like en:WP:NOR. Why couldn't the user apply the same principle to Wikidata and Egyptian Arabian Wikipedia? George Ho (talk) 09:13, 1 December 2020 (UTC)

• EnWiki has different policies then other Wikimedia projects. Given that the Libary of Congress believes those birth & death dates are real, I don't see a problem with them. ChristianKl ❪✉❫ 12:35, 1 December 2020 (UTC)

What about Wikidata:Living people? Are the standards just as high as enwiki? Furthermore, the LOC page cites Wikipedia, Wikidata, and IMDB as "sources". I don't see how it is reliable at this point. By the way, I don't know much about standards on pages about living persons at arzwiki. George Ho (talk) 18:47, 1 December 2020 (UTC)

The living people policy exists for living people and he seems dead. Even would apply it it's a different principle then the one you pointed to as being used on EnWiki. The libary of congress is an authority that does research. It's not perfect but no source is perfect. Better sources might his SALT LAKE TRIBUNE obituary from the May 24, 2001 but that unfortunately doesn't seem to be easily available online. ChristianKl ❪✉❫ 19:46, 1 December 2020 (UTC)

I was able to access the obituary via ProQuest (just go to Wikipedia Library). The obituary does not refer any alias of the deceased person "Thomas Earl Hagen", and it doesn't mention "Kip Noll" (or similar names) as well. To me, it doesn't verify the connection between the two names. George Ho (talk) 01:54, 2 December 2020 (UTC)

answers.com suggests "It appears that Kip Noll was actually Thomas Earl Hagen, born in Greenwich, CT, on August 7, 1957. The location and date correspond to information given by Kip Noll in a STALLION magazine ((June 1982) interview." is that article also available via Wikipedia Library or another way? ChristianKl ❪✉❫ 12:43, 2 December 2020 (UTC)

I don't know which specific policy prevents me from (re-)deleting unverifiable yet sourced birth and death dates and locations. Is it "use common sense" policy? However, that's not the only issue. As I have realized, Wikidata and other projects have become interdependent with each other, especially by sharing Wikidata into infoboxes. I notice other threads about. I fear that Wikidata's impact on projects would undermine the influence of English Wikipedia. Don't you think?

BTW, enwiki doesn't take Answers.com seriously; it's user-generated, like Wikidata. I tried finding the Noll interview on Google and Wikipedia Library; no such luck except data about the magazine issue itself. George Ho (talk) 18:06, 2 December 2020 (UTC)

@George Ho: In general sourced data should NEVER be deleted on wikidata; if you doubt its veracity you can adjust the ranking (if you have a reliable source for your doubts, then use "deprecated" rank and describe that with the reason for deprecation qualifier, otherwise add a new entry with "preferred" rank and the value you think is right, along with any source for that). Wikidata is designed to allow for multiple conflicting sources of information to be present on the same item. ArthurPSmith (talk) 18:21, 2 December 2020 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ I couldn't deprecate because customization is impossible. Instead, I must insert another item as "reason for deprecation". Say that I would like to insert either "English Wikipedia" or "chicken", but that's due to AutoCompletion feature. Also, connection Stallion interview and the obituary would be original research, which is against enwiki's policy (and other policies applied), isn't it? Somehow, I could not find Wikidata's policy about this.... unless Wikidata is allowing original research? Where's the RfC or past discussions on this? George Ho (talk) 18:51, 2 December 2020 (UTC)

• There's no Wikidata policy forbidden this. I'm not aware of any RfC that tried to implement such a policy. "reason for deprecation" indeed has to be an item. In particular it should be a subclass of Wikibase reason for deprecated rank (Q27949697). If none of the existing one's say what you need in a specific case you can create a new one. We do it that way, because Wikidata in multilingual and every language version can show the reason for deprecation this way by translating the label. ChristianKl ❪✉❫ 19:20, 2 December 2020 (UTC)

• Isn't creating another item less practical than what people have been optimistic about? Anyways, perhaps I should create an RfC without creating a draft yet... but I'll make the discussion about original research on living people, not about original research in general, which can be separate from my narrow(ed) topic. Any other ideas besides it? George Ho (talk) 20:37, 2 December 2020 (UTC)

• The idea is that there are a finite number of reasonable reasons for deprecating statements and in most cases we can reuse existing statements. Creating a new item is for those rare cases where there's a new valid reason for deprecating that doesn't already have an item. Wikidata is structured data. Structured data makes some things harder but has advantages like translation into other languages and it being easier for computers to interact with it. ChristianKl ❪✉❫ 23:20, 2 December 2020 (UTC)

• I created three more items: Q103598308, Q103598309, Q103598310. George Ho (talk) 00:14, 3 December 2020 (UTC)

• Regarding original research, I've been encouraged to publish original research on Wikiversity. I've accumulated a number of technical reports under "Wikiversity:Category:Freedom and abundance". There is now a w:Wikipedia:Wiki to journal publication procedure for submitting an article from Wikipedia to a refereed academic journal. I haven't studied that, but something similar could doubtless be done with a well written research report on Wikiversity. DavidMCEddy (talk) 20:07, 2 December 2020 (UTC)

Wikidata:Property proposal/cites work string

I made a proposal for a new property. Based on the feedback I have been adding "placeholder for <somevalue>" only to have it replaced by a bot. For me the whole understanding of the argument why the proposed property is controversial becomes lost again. My intention is not one of mix'n match, my intention is to have links to papers.

Having worked on adding citations for this paper, I find that adding a new item with only one maybe more properties would suit me fine. I would add a DOI and all the rest is extra. The benefits are: 1 it makes no material difference for my workflow, 2 it generates an identifiable paper that will show up properly as a referenced paper in Scholia, 3 it will enable links for citations from other papers enabling its growth. However such a paper is of the one but lowest level of quality; it only has a presence and one link. This would be no problem when there are processes that clean up and import the missing information..

Please comment. Thanks, GerardM (talk) 07:08, 1 December 2020 (UTC)

• The basic reason that it's opposed is that it solving the problem this way would mean creating a bunch of different XY string properties for different cases. Handling it with unknown value Help means that users can use that model with any property instead of us having to create new properties. ChristianKl ❪✉❫ 13:45, 1 December 2020 (UTC)
  • I understand. I hate it that a bot is replacing the placeholder with a text. I assume that you can query for "scholarly papers" with such a string. Thanks, GerardM (talk) 15:24, 1 December 2020 (UTC)
    • If you have a problem with a particular type of bot edits it would make sense to link to a diff and ping the bot. Otherwise it's a lot harder to engage with your complaint. ChristianKl ❪✉❫ 15:30, 1 December 2020 (UTC)
      • an example.. Thanks, GerardM (talk) 15:59, 1 December 2020 (UTC)
        • placeholder for "somevalue" (Q53569537) is a workaround for QuickStatements not supporting unknown_value being specified directly. The bot just changes it to the thing that's meant. ChristianKl ❪✉❫ 16:05, 1 December 2020 (UTC)
• what tool / workflow are you using to import these papers? is it entirely manual? is there a reason a tool cannot be used? in a perfect world the tool would make the items for the missing articles. in principle though i'm not against the proposal but I'd prefer we invest in tooling. BrokenSegue (talk) 17:13, 1 December 2020 (UTC)
  • What I do at this time is entirely manual not by choice but by lack of tools. I cannot trigger an update for a paper or for an author. I attended a presentation about this paper.. It introduced a methodology to consider ecological diversity. At this moment I am adding missing papers to Wikidata in order to complete the papers it used as a citation. There are no tools, the quality of what we have at Wikidata is poor as a consequence. So yes have tooling and have something that works for now. Thanks, GerardM (talk) 12:07, 3 December 2020 (UTC)

Tight crops only?

Will someone explain to me with reference to Joe Henry (Q1810016) why there is a need to crop the photo I took, which has been the image on this item since 17 February 2015? Was there some decision I missed that Wikidata wants tight crops only, or is someone being arbitrary? - Jmabel (talk) 05:16, 3 December 2020 (UTC)

Why are you asking here? An IP added a tag to a photo on Commons. It's a commons issue. (Your image is ... not very good - very poorly lit. I'll change the Joe Henry (Q1810016) image to Joe-Henry DSC00919.jpg, which whilst being b&w, seems about the best of those in the Joe Henry category.) --Tagishsimon (talk) 05:27, 3 December 2020 (UTC)

Q59318277 seems to have disappeared

How can the history of Q59318277 be researched?

I created a citation to it on 2018-12-04.

I believe it was a valid QID when I created it, though I can't prove that. However, it seems to have been valid on 2020-07-26, because it was NOT listed on Wikiversity:Category:Pages with empty citations on that date, even though three other Wikiversity pages were listed there then. However, early this morning, 2020-12-01, I was notified that it appeared then on Wikiversity:Category:Pages with empty citations.

I believe it used to describe A People's History of the American Revolution, which I just recreated this morning as Q102951790.

This says to me that something strange happened between 2020-07-26 and 2020-12-01. I see only two possible explanations for this:

1. Q59318277 was valid when I created it but disappeared from Wikidata sometime between 2020-07-26 and 2020-12-01.
2. Q59318277 was invalid when I created it but was only detected as such and added to Wikiversity:Category:Pages with empty citations until after 2020-07-26, over 1.5 years later, even though it had been invalid since I created it 2018-12-04.

??? Thanks, DavidMCEddy (talk) 19:35, 1 December 2020 (UTC)

On 24 April 2019 it was deleted by User:Jianhui67 following the request at Wikidata:Requests for deletions/Archive/2019/04/24#Q59318277. You can actually see the deletion log entry on Q59318277 and find the deletion request via Special:WhatLinksHere/Q59318277. Shall I undelete the item, so that you can merge the new one? —MisterSynergy (talk) 19:58, 1 December 2020 (UTC)

When it comes to creating entries for books and not having them deleted, it's useful to add an external ID like an ISBN number. ChristianKl ❪✉❫ 20:05, 1 December 2020 (UTC)

Please undelete it and point me to documentation on how to merge entries.

Also: @Randykitty: @Jianhui67:: You deleted Q59318277 as "Completely unreference" even though it was being used on Wikiversity, as noted above.

• How can I easily identify other Wikidata entries that I've created and that have been deleted on false claims that it was "Completely unreferenced"? This is frightening: I've created lots of Wikidata entries like that and used them extensively. Now at least one has been deleted as "Completely unreferenced", when it had been referenced from Wikiversity well over 4 months before it was deleted as "Completely unreferenced". Might there be others waiting to be detected? Fifteen months after this deletion, it was not listed as invalid on Wikiversity:Category:Pages with empty citations, as noted above. (An obituary for Mark Twain was reportedly published before he died. He replied, "the report of my death was an exaggeration.")
• What needs to happen to fix your algorithm that reported it as "Completely unreferenced", even though it had been referenced from Wikiversity well over 4 months before then? That algorithm should be able to identify references like [https://www.wikidata.org/wiki/Q102407368 Qui sont Julia et Agathe Cagé, mobilisées pour Benoît Hamon jusqu’au bout ? - Elle], which I plan to add to an article I'm planning to create on es.wikipedia.org. (I'm translating fr:Julia Cagé into Spanish. Will your algorithm find [[Bibliographie|Q102407368]] in the French language Wikipedia as a reference? Wikipedia:template:cite Q exists in French but not Spanish, so I'm using a naked URL. I've so far been unable to get a counterpart to Wikipedia:template:cite Q functioning in es.wikipedia.org.)
• Also, what needs to happen to make the defective algorithm that was used to search for references available to mere mortals, similar to the "File usage on other wikis" on Wikimedia Commons? "What links here" in Wikidata only identifies other links in Wikidata.
• What needs to happen so that "Requests for deletion" of QIDs trigger a notification to the creator of said Wikidata items? If your defective "Completely unreferenced" algorithm knows about uses in Wikipedia but not Wikiversity, in this case I could have replaced the citation to that book to the Wikipedia article on w:Ray Raphael with a reference to this Wikidata item, even if I might not have easily found where I used it in Wikiversity.
• Secondarily, how will adding an ISBN number prevent a book from being deleted as "completely unreferenced"? That does not sound like a "reference" to me.

Thanks very much for the reply. DavidMCEddy (talk) 21:10, 1 December 2020 (UTC)

• The item is restored. See Help:Merge how to proceed with the merge process.

@MisterSynergy: Help:Merge says, 'There are two ways to perform an automatic merge: the Merge gadget or the Special "Merge two items" page.' How can I find the '"Merge two items" page.'? I cannot find it, and I'd prefer not to install a gadget if I can avoid it. Thanks, DavidMCEddy (talk) 21:50, 2 December 2020 (UTC)

Special:MergeItems —MisterSynergy (talk) 22:29, 2 December 2020 (UTC)

• This tool shows which of the items that you have created were deleted.

—MisterSynergy (talk) 21:49, 1 December 2020 (UTC)

• Here completely unreferenced means that the item itself has no references for it's information. An ISBN number does provide the opportunity to check the information about a given book externally.

Usage of Wikipedia:template:cite Q is practically invisible from Wikidata. I recently opened a thread at https://www.wikidata.org/wiki/Wikidata:Contact_the_development_team#Template:Cite_Q_(Q22321052)_and_deletions about this being problematic and you might want to comment to it if you think this is important. ChristianKl ❪✉❫ 23:15, 1 December 2020 (UTC)

When Q59318277 was nominated for deletion, and when it was deleted, it had a statement author (P50)=Ray Raphael (Q7298002); it took only a second for a Google search to confirm it as a real entity. Furthermore, the author has a Wikipedia biography which discusses the work in detail. There seems to have been no attempt to notify its creator that the item was nominated for deletion, nor to ask them to provide further data. We need to stop valid, albeit incomplete, items from being deleted in this way. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:27, 1 December 2020 (UTC)

indeed the deletion policy is way too aggressive. I've complained about this many times but the admins insist that aggressive deleting with occasional undeleting is better than adding any process that would slow it down (notifying creators / simple googling /etc). BrokenSegue (talk) 02:32, 2 December 2020 (UTC)

• First, I have to say that I wouldn't have deleted the item in the state in which it was. Our line about what we delete and not delete comes from discussions on RfD. If you want to move it it makes sense to participate more on RfD.

The fact that we don't have notification of creators for RfD is not do to "admins" but due to no one writing the bot code for such notifications. Pasleim's DeltaBot would be in a good position to create pings but Pasleim isn't the only person who could do that work. ChristianKl ❪✉❫ 09:36, 2 December 2020 (UTC)

@ChristianKl: So unfortunately part of the problem is admins (or at least the policy as interpreted by some of them). Many many items are deleted without RfD review. Some admins assert that review would be pointless because deletions here are purely mechanical and thus there is no room for discussion (they say the RfD page is merely to notify admins not a place to reach consensus or discuss). Even if we did notify authors of deletion nominations there is no time limit there and admins can delete things whenever they want. For an example of my grievance see this discussion where multiple admins disagree with your assertion that RfD is for discussion. This isn't just a matter of solving a tooling issue or else I'd be glad to fill that gap. BrokenSegue (talk) 15:56, 2 December 2020 (UTC)

@BrokenSegue: In the linked discussion the only admin that replied is RfD is a mix of discussions and requests that are closed without discussion. It's my understanding that the current defacto deletion policy is essentially one of case law that comes from the discussions we have at RfD. It's my impression that deletions that don't go through RfD mostly reflect the RfD case law.

The issues involved in deletion are complex. Some of them are issues of tooling while others are issues of policy. The fact that articles can go to RfD without their creators knowing about it is an issue of tooling that could be easily fixed.

Creating a new process of speedy deletion that could be used in cases where admins currently delete items directly would also be good, but is complex mix of technology and policy. ChristianKl ❪✉❫ 16:23, 2 December 2020 (UTC)

@ChristianKl: Well one admin there did say "For good reasons RfD does not have a minimum discussion period defined or any other formal requirements that need to be respected while resolving cases" which seems kinda outrageous to me. What use is notifying users if there is no minimum waiting period (and what harm could such a waiting period even cause)? I think both issues are a mix of policy and technology. BrokenSegue (talk) 16:29, 2 December 2020 (UTC)

@BrokenSegue: the fact that there's no formal required minimum waiting period doesn't mean that most RfD discussions aren't open for many days. Especially deletion decisions that aren't clear cut are often open for longer periods of time.

The harm that a minimum waiting period causes is that it adds extra work to process RfD's. Keeping bureaucratcy down both for users who create items and for admins is valuable. ChristianKl ❪✉❫ 18:20, 2 December 2020 (UTC)

What should I do to argue more forcefully for the following:

1. Before any Wikidata item is deleted, any human who contributed to such item should be notified, explaining the time at which the item is scheduled for deletion and clear instructions on what needs to happen to cancel that deletion. I think the deletion time should be between 3 and 6 days (72 and 144 hours) from the time of the notice to allow user(s) time to schedule time to research the problem, to figure out what to do, and then do it. This deletion time might be 36 hours in the future IF you make it easy for users to see items to which they contributed after they've been deleted, as suggested below; I don't know how to do that right now.
2. Very compelling labeling of the Wikidata input screen for each new Wikidata item stating clearly and succinctly Wikidata policies for deletion. That labeling should disappear only after the criteria for retention have been met.
3. It should be easy for users to see the details of deleted items to which they contributed with appropriate banners explaining that it has been deleted and cannot be used, though they can copy the contents to new items.
4. It should be much easier than it is presently for people to get help understanding Wikidata policies and recommended practices. I've been a fan of Wikidata since 2017, but I failed to find adequate instructions on how to use it until I got individual help with it in Wikimania 2018 Cape Town. There have been several times since then -- like now -- that I've struggled with whether and how to do something in Wikidata. Some of my posts to "Wikidata:Project chat" have gone unanswered. It should be easy, for example, for a user to find how to delete a Wikidata item they created in error; I have not found it so.

Permit me to add to my comments above on why this concerns me:

I spent most of yesterday on tasks that should have been unnecessary if policies like what I recommend were in place.

First, I got a notice that Wikidata item Q59318277 appeared on "Wikiversity:Category:Pages with empty citations", even though it was cited on Wikiversity:The Great American Paradox.

If that {{cite Q|Q59318277}} had not had a page number following it, I would not likely have been able to overcome that immediate problem within any reasonable period of time. I found that I had added that reference to {{cite Q|Q59318277}} on 2018-12-04. Sadly, my research of my notes from that date failed to identify the document to which Q59318277 referred. Fortunately, the reference included a page number as, "{{cite Q|Q59318277}}, p. 89.". I then pulled a dozen books off my shelves and checked p. 89 in each until I found the one I needed. Then I recreated the Wikidata item for that book.

Then I began to worry that the replacement I created may itself soon disappear. I still don't know what I need to do to prevent that. I've heard that MAYBE if I add an ISBN number, it might reduce the chances that this replacement for Q59318277 that I created yesterday may itself disappear. However, the criteria for deletion are still not clear to me.

IF WIKIDATA ITEMS DISAPPEAR SEEMINGLY AT RANDOM WITHOUT EXPLANATION, IT SHOULD RAISE SERIOUS QUESTIONS ABOUT WHETHER WIKIDATA SHOULD BE USED FOR ANYTHING.

I've been a registered Wikidata user since 2012-12-24, but I made only 16 total edits in Wikidata prior to 2018 Wikimania Cape Town, as I mentioned above. I remember attending a presentation on Wikidata at 2017 Wikimania Montreal, but it was not feasible for me to actually use it until I got individual help a year later. It shouldn't be that hard.

Yesterday, I learned that 6 Wikidata items I had created had been deleted. Two have since been restored. I still have to create time to figure out what to do with those two. And it's not clear what I need to do to understand what happened to the other four. I may have requested the deletion of one or two of those four, but I doubt if I requested the deletion of all four.

I'm in the process of translating the Wikipedia article on w:Julia Cagé from French into Spanish. It's taking a long time, in part because I insist on creating Wikidata items for every reference I use, and I can't find some of their references ;-)

And I've encountered a certain level of mild hostility in my requests to have w:Template:Cite Q made available in the Spanish language Wikipedia. I plan to get around that using constructs like [URL comment], e.g., [https://www.wikidata.org/wiki/Q102407368 Qui sont Julia et Agathe Cagé, mobilisées pour Benoît Hamon jusqu’au bout ? - Elle]. If Wikidata items disappear seemingly at random, the mild hostility I've encountered from Spanish language Wikimedians is justified.

Thanks for all you to do help make the entire world better informed. DavidMCEddy (talk) 14:27, 2 December 2020 (UTC)

@DavidMCEddy:I went through your deleted one's yesterday. One was request by you to be deleted. The other three had no statements and were named Julia et Agathe Cagé, les sœurs douées/Essays in the Political Economy of Information and Taxation/news email. ChristianKl ❪✉❫ 18:11, 2 December 2020 (UTC)

@ChristianKl: It's difficult to see what the appropriate thing is to do with these.

• Q96951498 was created 2020-07-04 18:47 and deleted 2020-07-13. It was titled, "Julia et Agathe Cagé, les sœurs douées". This seems to be a duplicate of Q96951729, which was created 2020-07-04 18:52, five minutes later. Without being able to look at the edits on Q96951498, I cannot say more about this. If Q96951498 included more than one edit, I think it should be restored, and I should merge it with Q96951729. Wikidata item Q96951729 is currently cited in both the English and French language articles on Julia Cagé, and I plan to cite it as [https://www.wikidata.org/wiki/Q96951729 Julia y Agathe Cagé, las hermanas superdotadas (en francés).] in a Spanish language article on Cagé.
• Q96950823 was created 2020-07-04 18:37 and also deleted also 2020-07-13. It was titled, "Essays in the Political Economy of Information and Taxation". I assume it's the same as Q102421658, which I created on 2020-11-28 with English title, 'Certification of doctoral defense of "Essays on the political economy of information and taxation"'. The latter documents the defense of a French-language PhD dissertation by Julia Cagé. Could you please recreate Q96950823, so I can see if it's the same as the certification I have of thesis defense or it's the actual thesis? (It's in French, so the difference between "in" and "on" in the title is more a function of the translator than the content ;-)
• Might the "news email" document a comment on an email list that I might have cited somewhere? If yes, I think I should examine it as well.
• After Q59318277 was restored, User:Trade changed "instance of (P31)" from "book (Q571)" to "literary work (Q7725634)". Is "literary work (Q7725634)" preferable to "book (Q571)" for something like this? This is a carefully documented history of the American Revolution, for which "book" sounds more appropriate. More generally, how can I research when one should use Q571, when Q7725634, and when both? Also, I've looked at Help:Merge, and it's not clear to me where I should go from here. I suppose I should try the instructions for "Automatic merge" and see what happens. I see that, "Usually the more recent item should be marked" as "duplicated" and the other as the "main" item. Does that mean I should start at Q59318277 and merge Q102951790 into it?

Thanks again. DavidMCEddy (talk) 19:50, 2 December 2020 (UTC)

@DavidMCEddy: I undeleted Certification of doctoral defense of "Essays in the political economy of information and taxation" (Q96950823): Documentation of 2013 doctoral defense at EHESS and Q66424630: no description as well. Given that both lack statements they are not viable in their current form.

@ChristianKl: Thanks. I looked at Certification of doctoral defense of "Essays in the political economy of information and taxation" (Q96950823): Documentation of 2013 doctoral defense at EHESS and Q66424630: no description. You're right: They "both lack statements".

• I will try to merge Certification of doctoral defense of "Essays in the political economy of information and taxation" (Q96950823): Documentation of 2013 doctoral defense at EHESS with 'Certification of doctoral defense of "Essays on the political economy of information and taxation" (Q102421658)'.
• Since no other Wikidata item links to Q66424630: no description, I suggest you speedily delete it before someone else tries to use it ;-) DavidMCEddy (talk) 21:33, 2 December 2020 (UTC)

As far as the topic of books goes https://www.wikidata.org/wiki/Wikidata:WikiProject_Books#Bibliographic_properties explain our modelling. ChristianKl ❪✉❫ 20:59, 2 December 2020 (UTC)

@ChristianKl: Am I reading your https://www.wikidata.org/wiki/Wikidata:WikiProject_Books#Bibliographic_properties correctly to suggest that I should avoid using "book (Q571)", because it's too general, and instead use "literary work (Q7725634)", because it's more specific? And where I've assigned both, I should delete "book (Q571)"?

A related question is when to use "news (Q38926)" vs. "news article (Q5707594)"?

I've recently been assigning both. However, if "literary work (Q7725634)" is preferred to "book (Q571)" because it's more specific, then "news article (Q5707594)" should be preferred to "news (Q38926)" for a news report that I've found in a news medium? And, again, where I've assigned both, I should delete "news (Q38926)" in favor of "news article (Q5707594)"?

Thanks again, DavidMCEddy (talk) 21:20, 2 December 2020 (UTC)

Theoretically, the place that should document the answer to that question is https://www.wikidata.org/wiki/Wikidata:WikiProject_Periodicals . It currently doesn't speak about news aricles explictely but about article (Q191067). To me that suggests using news article (Q5707594) is the way to go here. At the same time it would likely be good if the documentation would explictely say so, so I raised the issue of being more specific on the discussion page of that Wikiproject. ChristianKl ❪✉❫ 22:49, 2 December 2020 (UTC)

As long as it's not visible from Wikidata whether or not items are used with w:Template:Cite Q and thus that has no inpact on deletion decisions, I do agree that skepticism towards it is warrented. Even if you do think we should generally delete fewer items, having access to that information seem valuable and that's why I created the topic over at Contact the development team. ChristianKl ❪✉❫ 18:11, 2 December 2020 (UTC)

Making policy more legible is not something you archieve by forcefully arguing for it but by actually doing the work of working on making policies more legible and doing tasks like answering questions on the project chat. ChristianKl ❪✉❫ 18:11, 2 December 2020 (UTC)

This is why we need to introduce a speedy deletion process (and change existing deletion to a 7-day process).--GZWDer (talk) 15:46, 2 December 2020 (UTC)

I strongly agree and proposed such a thing in this long discussion but got pushback (and others have made similar proposals [1] and [2]). It seems @MisterSynergy: opposes this kind of reform. BrokenSegue (talk) 16:09, 2 December 2020 (UTC)

@BrokenSegue: when it comes to create a proposal for such a system it's important to understand the technical features that our current system has. I'm generally willing to work to formulate policy. In the linked discussion there's an open question about whether abuse filters can be used to mark a property in a way that it can only be added by users with a specific right but removed by everyone. If that's possible we can go further and write a speedy deletion system based on that principle, but without knowing it there's no way to move forward with that specific way of implementing it. ChristianKl ❪✉❫ 18:15, 2 December 2020 (UTC)

Any speedy deletion process that does NOT honestly and effectively address the concerns I raised above is user hostile and drives volunteers away from the Wikimedia project. It is a threat to the sustainability of what we are trying to do here.

After I had translated the first 20% of the French language Wikipedia article on w:Julia Cagé into Spanish, someone speedily deleted what I had done. That already included 7 references documenting her status as an economist with PhDs from both Harvard and Science Po and her status as an expert in the economics of media and democracy. The person who speedily deleted what I wrote did not respond to my comments about why I thought it should not be deleted. I've gotten zero response to requests for information on what needs to happen to have that article restored.

You want to know why Wikimedia projects have such a hard time recruiting and keeping volunteers? This is one of the reasons.

I agree that there should be speedy deletion processes, but they should not be managed in such a user hostile fashion as I've experienced recently with es.wikipedia.org and Wikidata. DavidMCEddy (talk) 16:11, 2 December 2020 (UTC)

• Wikidata is very inclusive, at least compared to most Wikipedia projects. The biggest gap (as discussed above and elsewhere) seems to be whether the Wikidata item has incoming links from certain external projects (e.g. Wikipedia references, OpenStreetMap). Should being cited in a Wikipedia article be enough to make a Wikidata item notable? Should being in OpenStreetMap? Not absolutely, but it ought to move the needle towards inclusion. Unfortunately, other than the sitelinks, there is simply no mechanism for a deleting admin to detect such links and it is not explicitly part of the notability criteria. A possible mitigation would be to create some sort of "deletion review" tool that would report such external references.

Looking at the earlier discussions (and ignoring the specific implementation details), it seems like people are asking for: an easy way for any trusted user to propose deletion of an item; and a reasonable opportunity to object to proposed deletions, including notifications and delay. I would favour such a process. People are also asking for ways to view deleted content, which I sympathise with, but it would be hard to provide without raising concerns about hosting copyright- or BLP-violating content. Bovlb (talk) 19:01, 2 December 2020 (UTC)

@Bovlb: in addition to what you said, admins who delete a lot of items also don't want to create 1000s of RfD's per week but a way to delete items with less overhead. If we had a way where an admin could tag an item for deletion and if nobody objects it gets automatically deleted after X days and if someone objects it goes to RfD that would help with that concern. ChristianKl ❪✉❫ 19:25, 2 December 2020 (UTC)

In the English Wikipedia there is PROD (proposed deletion), which works exactly like this (but anybody can PROD articles, not only admins).--Ymblanter (talk) 20:09, 2 December 2020 (UTC)

we have the tools to implement this today. just make a new property called "wikidata review status" and have people set it to "proposed deletion". BrokenSegue (talk) 20:37, 2 December 2020 (UTC)

The difference between Wikidata and Wikipedia is that plenty of pages on Wikidata get a lot less attention and have fewer people who have it on their watchlist. I don't think that every user should be able to delete items that are on nobodies watchlist. On Wikipedia I would limit that to admins and either a new right or to people with the rollbacker right. Basically, only people who can be trusted to understand how our deletion policy works should be able to delete items without interaction from other people. ChristianKl ❪✉❫ 20:52, 2 December 2020 (UTC)

I thought more about the issue and wrote a draft that just needs a bot and no other technical features: https://www.wikidata.org/wiki/User:ChristianKl/Draft:ProposeDeletion#Detecting_manual_removal_of_propose_deletion_for_reason I'm happy for input. ChristianKl ❪✉❫ 11:12, 3 December 2020 (UTC)

@Ymblanter: I looked at the PROD policy and according to it there's admin review of items which means workload. Here we want a system that doesn't produce additional human workload and thus it makes sense to restrict the right to use the mechanism. ChristianKl ❪✉❫ 19:28, 3 December 2020 (UTC)

PROD indeed has admin review. If we have the system without additional admin review, the only way out is restrict the mechanism to admins, otherwise users without an admin flag get an opportunity to delete articles - which, I believe, is not something that the community generally supports.--Ymblanter (talk) 19:32, 3 December 2020 (UTC)

FWIW, the linked proposal restricts this to the rollbackers user group. BrokenSegue (talk) 20:09, 3 December 2020 (UTC)

• It is unfortunate, but the best way to prevent deletion of valid entries is to have as much cross linking as possible. The author should be linked, and at the author page notable_work=X. As mentioned the ISBN number and the Google Book ID as well as the Amazon ID should be included. The more you add the less likely the unfortunate deletion will occur. If the book has a specific topic, you can add described_by_source=X from that topic. When an item is linked from another item, autodeletion is prevented because the item has a structural need. --RAN (talk) 14:11, 4 December 2020 (UTC)

Wikidata for mythical names

Inviting participants of the WikiProject Ancient Greece to join the Linked Pasts session on Exploring Names in Wikidata, Mythical and Otherwise on Dec 8, 2020 UTC 13:00. Anyone with an interest to the topic is welcome! See the event doc for the Zoom link (Search for Exploring Names in Wikidata, Mythical and Otherwise) or get in touch with me or User:JBradyK. – Susanna Ånäs (Susannaanas) (talk) 06:55, 4 December 2020 (UTC)

Epìdosis B20180 Geraki Azertus Alexander Doria Shisma Sp!ros Xena the Rebel Girl Alexmar983 DerHexer Lykos EncycloPetey Jahl de Vautban JBradyK Mathieu Kappler Ahc84 Liber008 JASHough User:Tolanor User:Jonathan Groß

Notified participants of WikiProject Ancient Greece

deletion request

Please delete Karolina Protsenko (Q101071444). Thanks in advance!!! --2001:B07:6442:8903:4572:EA66:C788:E0C8 10:07, 4 December 2020 (UTC)

• It's unclear why. Isn't it the same person as Karolina Protsenko (Q91313507) and thus a merge candidate? ChristianKl ❪✉❫ 10:16, 4 December 2020 (UTC)

Anne

Why was Anne (Q564684) changed from instance of female given name to unisex given name, in 2019? @HarryNº2:? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:18, 1 December 2020 (UTC)

Does it even make sense to have male and female versions of identical names? E.g., Anne (Q47860848), is it really a different name? Ghouston (talk) 00:03, 2 December 2020 (UTC)

As long as Anne feminum and Anne masculinum has different orgins all the of them should exist. Then QXXXXXX will be the female name is a form of the Latin female given name Anna. This in turn is a representation of the Hebrew Hannah, which means 'favour' or 'grace and then Anne (Q47860848) is related to Germanic arn-names and means 'eagle. Also then Anne (Q564684) must exist do describe the unisex part. Also bear in mind interwikis. Some day maybe the property: name orgins from will be created as a complementary property (P8882) to named after (P138) Pmt (talk) 01:10, 2 December 2020 (UTC)

Wouldn't it be difficult to work out which "Anne" any particular person is supposed to be using? What about a male who's middle name is "Anne" to commemorate his grandmother, or some such? Ghouston (talk) 01:54, 2 December 2020 (UTC)

Because the given name Anne is given to women and men. --HarryNº2 (talk) 02:52, 2 December 2020 (UTC)

That is not in dispute. However, this item was at one point about the female form of the name. We have a separate item Q47860848, about the male form. I want to know why this item was repurposed, and what it should be classed as. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:05, 2 December 2020 (UTC)

Actually, it is a unisex given name (Q3409032) since 2014: https://www.wikidata.org/w/index.php?title=Q564684&type=revision&diff=153553200&oldid=152051515 Ayack (talk) 08:35, 2 December 2020 (UTC)

If one person is using both the feminum and masculinum the second name that in western (mostly) use will be a middle name (Q245025) and will have qualifier series ordinal (P1545) with 2 and then also has use (P366) with male given name (Q12308941) and so of (P642) as middle name (Q245025). See Q103420151. Pmt (talk) 15:25, 2 December 2020 (UTC)

And yet the item links to Wikipedia articles that are about the female form. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:05, 2 December 2020 (UTC)

@Pigsonthewing: Not in French: “Anne est un prénom épicène”. Ayack (talk) 18:24, 2 December 2020 (UTC)

Maybe we should get out of the business of assigning genders to first names altogether? We've settled on instance of (P31)human (Q5) for people, so why should it be different for their names? Between cultural differences and changes in law and society regarding both names and genders, the distinction is complicated at best and only getting worse. Since we link people with their first names, data on their use in the so-called real world is just a query away, and the distinction via instance of (P31) is both duplicative and less accurate. --Matthias Winkelmann (talk) 18:07, 2 December 2020 (UTC)

I've stopped assigning gender to humans unless their preferred pronouns are clearly documented somewhere.

A year or two ago I made a video for a trans friend of a memorial for murdered transgender people. One of the speakers insisted, "I am neither male nor female." DavidMCEddy (talk) 18:14, 2 December 2020 (UTC)

Different jurisdictions and cultures treat genders differently. At Wikidata we don't try to set standards but allow different jurisdictions and cultures to express themselves. In Germany it's legally relevant whether or not a name is a female or male name as parents are not allowed to name children with a name that goes against their gender. ChristianKl ❪✉❫ 10:36, 4 December 2020 (UTC)

Perhaps we could find a way to express that a particular name is considered gendered by a particular government body, but that still doesn't imply that we need more than one item for the name. Ghouston (talk) 01:16, 5 December 2020 (UTC)

Wikidata entries from a Wikipedia category?

I want to find all Wikidata entries from only articles in a Wikipedia category and in a specific language. How can I do this ? Thanks. Best regards, Pelanch3 (talk) 13:42, 4 December 2020 (UTC)

mw:Wikidata Query Service/User Manual/MWAPI#Find_category members, Wikidata:Request a query --Lockal (talk) 16:06, 4 December 2020 (UTC)

Hello @Pelanch3, Lockal: in addition to SPARQL also Petscan could be used:

• https://meta.wikimedia.org/wiki/PetScan
• https://petscan.wmflabs.org/

--M2k~dewiki (talk) 16:35, 4 December 2020 (UTC)

electric locomotives properties

Q841273 was bicourrant engine for 3 kV direct current and 25 kV alternate current. I try to use the P2436 property, but it does not accept Q1412243 as value. How can I do this?Smiley.toerist (talk) 14:49, 4 December 2020 (UTC)

As it turns out, this works: SNCB Class 15 (Q841273)type of electrification (P930)25 kV AC railway electrification (Q1412243). I have no idea what any of that means, so I'll count on you if anyone comes asking about these... trains(?) --Matthias Winkelmann (talk) 23:24, 4 December 2020 (UTC)

Newbie question: Instance of versus subclass of

Hi, I'm looking at Chess Openings as my first project on Wikidata. Starting with the top, we have:

King's Pawn Game (Q1631907) subclass of (P279) chess opening (Q103632).

Open Game (Q753060) subclass of (P279) King's Pawn Game (Q1631907).

When we get to King's Knight Opening (Q1237452), it is instance of (P31) chess opening (Q103632), and also is subclass of (P279) Open Game (Q753060). This makes sense because the King's Knight Opening is a chess opening and also it defines a class of chess openings which begin with that sequence, which is also true. What I don't understand is why Open Game (Q753060) isn't similarly instance of (P31) chess opening (Q103632)? It should be, right? Also, thinking about the "terminal nodes" of the graph defined by the subclass relation starting at chess opening, something like Semi-Slav Defense (Q1570575) seems like a terminal node (query). There are no subclasses of it, and no instances of it. By leaving the statement of Semi-Slav Defense (Q1570575) subclass of (P279) Queen's Gambit (Q107925), are we saying essentially that "There are a whole class of chess openings further branched from this one, we just don't have names for them?" Or is it just a relic of how the data was modeled? And wouldn't Semi-Slav Defense (Q1570575) instance of (P31) Queen's Gambit (Q107925) be more correct than chess opening (Q103632)? Given that, wouldn't the initial example, King's Knight Opening (Q1237452) be better to instance of (P31) Open Game (Q753060) instead of the generic chess opening (Q103632)?

Basically, I'm suggesting that chess openings are both an instance and a subclass of the chess opening they derive from. Does this make sense? And should the subclass relationship be pruned from openings where there are no named/known instances? Thanks!--Audiodude (talk) 06:25, 3 December 2020 (UTC)

• @Audiodude: Not sure I entirely follow, but it seems like these are modeled so that chess opening (Q103632) is a metaclass, and the classes of openings would have instances (if they existed) that were occasions when that opening was actually used in a game? Just because something is a class with no instances here in Wikidata doesn't mean that its "real world" (abstract) meaning wouldn't actually have instances of some sort. ArthurPSmith (talk) 18:47, 3 December 2020 (UTC)
  • Thanks for the response and the welcome message! I guess it makes sense to think about something analogous that's more basic, like colors. alice blue (Q372669) instance of (P31) color (Q1075), but also alice blue (Q372669) subclass of (P279) blue (Q1088). So following that example, it would make sense to model the chess openings the same way right? So King's Knight Opening should be "instance of" chess opening, but "subclass of" the specific chess opening that it derives from (in this case Open Game). The only thing about this modeling that strikes me odd is that you don't end up with any colors that are "instances of" blue. And is alice blue really a "class" of colors? Thanks! --Audiodude (talk) 20:54, 3 December 2020 (UTC)
    • @Audiodude: The difference between instance/subclass for non-physical things is really subtle sometimes and I think even experienced people get confused. Some of this is just trying to be consistent within the same realm. Does it make sense to say there is an "instance of" blue anywhere? Maybe a particular frequency? Really blue is a rough region of color space and other blue-ish colors are subregions of blue. I don't think adding chess opening (Q103632) to Open Game (Q753060) would hurt but it's a little redundant. Have you seen Help:Basic membership properties? BrokenSegue (talk) 21:10, 3 December 2020 (UTC)
      • Thanks for the reply! I can see how it's tricky, yes. And yes I was referring to that help document specifically when working on these issues, and the key that struck me was that for an entity to be a subclass of another entity, both of them must be classes. Looking at ArthurPSmith's answer above, it seems like it's okay to have a class "Semi-Slav Defense" of chess openings, even if there are no entities in Wikidata that are instances of that class. What concerns me more is setting "Semi-Slave Defence" to be instance of "chess opening" instead of instance of "Queen's Gambit" because if we follow that practice, the "class" Queen's Gambit is never going to have any instance members. Thanks! --Audiodude (talk) 21:17, 3 December 2020 (UTC)
        • Yeah I don't think it matters if a class never has an instance in wikidata. I think you could safely delete that "instance of" but optimally there would be a "Wikiproject Chess" that would come to an agreement on this. BrokenSegue (talk) 21:24, 3 December 2020 (UTC)
• @Audiodude: If they are all forms of chess opening, could it work to just make all the openings a tree of subclasses under chess opening (Q103632)? IMHO the instances arise when an opening is used in a notable game. Pauljmackay (talk) 15:16, 5 December 2020 (UTC)

Please merge

Westerplatte Monument (Q825098) and Westerplatte Monument (Q67012160). TIA. Also, how many more years before we get a nice merge button? I read the merging instructions and got bored after two minutes of trying to figure which version is superior (obsolete). If one is supposed to toss a coin, say so, if we prefer an earlier version, there should be an automated tool that does it after I list identical Qs and tell the system they need merging. --Piotrus (talk) 01:31, 5 December 2020 (UTC)

Piotrus Go to your Preferences (upper left) and Gadgets panel, and select the Merge gadget, the first in the list. Then the Merge Wizard ("Merge With") dialogue will appear on every item beneath the "More" tab. You can merge simply by pasting the Q code of the item you wish to merge, and the oldest item will be default unless you change it. -Animalparty (talk) 02:06, 5 December 2020 (UTC)

Animalparty Thanks, that wasn't as bad as the scary instruction page suggested it would be. --Piotrus (talk) 09:32, 5 December 2020 (UTC)

Vice counties

Some time ago I created an item for each instance of a vice-county (Q7925010) (see en:Vice-county for background).

I have just rescued one of them, Worcestershire (Q17581850), which was wrongly merged into Worcestershire (Q23135) some time ago. As you can see, that has a catalog code (P528) value of "37" in catalogue vice-county (Q7925010) (otherwise "VC37").

The equivalent item for the vice-county of Warwickshire, VC38, is missing, I can find no trace of it having been merged onto another item.

Warwickshire (Q67575123) claims to be both historic county of England (Q1138494) and historic county of England (Q1138494) (which is self-contrdictory); but the former is a more recent addition, not part of a merge, and no other items redirect to it.

Perhaps the original VC38 item was deleted? If so, how can it be traced, and restored? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:24, 5 December 2020 (UTC)

Looks like it was merged into Warwickshire (Q23140), not Warwickshire (Q67575123). Andrew Gray (talk) 21:26, 5 December 2020 (UTC)

@Andrew Gray: Thank you. I've restored it. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:47, 5 December 2020 (UTC)

• Nice concept, geographical areas which are independent of constantly-changing government administrative boundaries. Ghouston (talk) 01:00, 6 December 2020 (UTC)

Previous/next link in Special:WhatLinksHere pages

Tracked in Phabricator
Task T269830
Resolved

Currently, HTML tags of previous/next link in Special:WhatLinksHere pages seem to be sanitized and not work. --Okkn (talk) 07:00, 10 December 2020 (UTC)

Thanks, reported at T269830. (For future reference: WD:DEV would be a better place to report software issues like this.) --Lucas Werkmeister (WMDE) (talk) 09:30, 10 December 2020 (UTC)

Should be fixed now. --Lucas Werkmeister (WMDE) (talk) 12:32, 10 December 2020 (UTC)

Thank you for the speedy repair! --Okkn (talk) 13:53, 10 December 2020 (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. Lucas Werkmeister (WMDE) (talk) 09:51, 11 December 2020 (UTC)

Political convention delegates

Do we have a standardized way of showing someone was a delegate to one of the political conventions? --RAN (talk) 14:06, 4 December 2020 (UTC)

No one responded so I tried this: Richard Young Hahnen (Q70242725) --RAN (talk) 07:00, 6 December 2020 (UTC)

Separate actions for common concepts

Hi, I remember having discussed this earlier but I don't remember the outcome of the discussion. Is it acceptable with items like reverse engineer (Q103910331) that refers to the action of reverse engineering (Q269608). If yes, how do we best link them together. Do we need a new "action of" property?--So9q (talk) 13:46, 6 December 2020 (UTC)

• When creating items like this I think it's best to ask: "What value does the new item create?" reverse engineer (Q103910331) is an item with only one statement and no external references and thus of questionable use. I think the notability of the item is questionable as it stands currently. ChristianKl ❪✉❫ 14:28, 6 December 2020 (UTC)
  • Thanks for the input. Its very useful to map to the sense on a lexeme like demontera (L55465) and reverse engineer (L347677) which is a structural need because for tools like MachtSinn we connect the lexemes and the item by matching on the alias and the english verb "reverse engineer" does not match anything in reverse engineering (Q269608) right now and I'm hesitant of adding all possible verb forms of the actions related to the concepts on the item. Which is the less bad route to take? A new item that describes and action and can hold all the aliases and act as a link between the languages or items with a truckload of aliases for every conceivable verb related to the concept?--So9q (talk) 18:09, 6 December 2020 (UTC)
    • I haven't done much with lexemes, so I don't have a strong opinion about whether that's the best way to model this use-case but it seems to me reasonable to have the item for that purpose. In this case, it would help with discussing the item if you would actually link it the way you believe it's intended to be linked. It's better to discuss the problem then to discuss an item in isolation of it's usage. ChristianKl ❪✉❫ 18:28, 6 December 2020 (UTC)
      • I agree, sorry for not being clear, I have not linked it because it is tedious to do manually which is why I'm creating the items, waiting for MachtSinn to update and then use that tool to create the linking with little effort.--So9q (talk) 18:39, 6 December 2020 (UTC)
        • It's generally useful to start by mapping out a structure manually and then when the structure is clear automate the process. ChristianKl ❪✉❫ 20:13, 6 December 2020 (UTC)
• Both are fine, and reverse engineer (Q103910331) is the value to use with occupation (P106) (although there is no such use yet). I have connected both items with field of this occupation (P425) and practiced by (P3095). —MisterSynergy (talk) 14:46, 6 December 2020 (UTC)
  • While I agree that we have a structural need for that occupation, the act of reverse engineering called "reverse engineer" is not the same thing as the occupation of "reverse engineer". ChristianKl ❪✉❫ 14:57, 6 December 2020 (UTC)
    • I do not understand your concern. Can you please elaborate? Thanks, —MisterSynergy (talk) 15:00, 6 December 2020 (UTC)
      • @MisterSynergy: You did remove the existing instance of (P31) statement and turned the item into something of a different instance of (P31). ChristianKl ❪✉❫ 15:01, 6 December 2020 (UTC)
        • Yes, I normalized it according to common practices. But how does this relate to your comment above, and how is this an answer to my request for clarification? —MisterSynergy (talk) 15:04, 6 December 2020 (UTC)
          • So9q asked whether or not an item about the act was fine or not. Turning the item into something else then it was and then saying that the new thing is fine evades the question about whether or not having an item for the thing that he proposed is fine. ChristianKl ❪✉❫ 15:11, 6 December 2020 (UTC)
            • @MisterSynergy, ChristianKl: I agree with ChristianKl here, you hijacked my perfectly good item :). But doing that you reacted to my question by turning the item into a "more useful" normalized item. In its current form its not suitable to link to the verb "reverse engineer" because that's not what it means anymore. I guess we have to chisel out a solution that solve all our needs here. See above for my two suggestions, but there might be other ways to solve the issue of what the lexeme verbs should link to and how to do that linking effectively (with tools like MachtSinn for example).--So9q (talk) 18:09, 6 December 2020 (UTC)
              • I see. The lexeme/the verb reverse engineer (L347677) was not on the table here when I wrote my answer and modified reverse engineer (Q103910331), and there was no reason to assume that it plays a role here as we usually do not describe words/terms in the item namespace. Now I am not the most experienced editor regarding lexemes, but I think there is item for this sense (P5137) and maybe a few other properties to link from the lexeme to the activity reverse engineer (Q103910331) or the discipline reverse engineering (Q269608) (not sure which exactly fits here). See Template:Lexicographical properties for an overview of existing lexeme related properties. —MisterSynergy (talk) 18:55, 6 December 2020 (UTC)
                • Yeah I see now that I completely forgot to mention that in the beginning. Sorry about that. item for this sense (P5137) is very nice. We already have translation based on it via Ordia, see https://ordia.toolforge.org/L37303-S1 for an example.--So9q (talk) 19:01, 6 December 2020 (UTC)
          • Interesting turn, maybe "reverse engineer" (Q103910331) is probably an "occupation" as it's an activity, not profession. So it does cover the action of Q269608. --- Jura 18:14, 6 December 2020 (UTC)
            • Oh, I had not thought about that, maybe its both an occupation and an intentional action then? I leave that up to you and @MisterSynergy: to figure out. As long as it can meaningfully link the verbs for reverse engineer in different languages I'm happy :)--So9q (talk) 18:42, 6 December 2020 (UTC)

Advice on using Wikidata for vocabulary data

I'd like to add Items to Wikidata for foreign-language words, phrases, and sentences, with properties representing the information needed to create a flash card or vocabulary study list for a fluent English speaker studying that language. These would be used to support foreign language learning content on both Wikibooks and Wikiversity.

Right now, vocabulary is hand-written into lessons, the way it would be in a physical textbook. Sometimes a word will appear in multiple contexts (in more than one lesson, in flash cards and vocabulary lists, etc.) or more than one project (for example, in a Wikibooks language text and a Wikiversity language resource). This is less than ideal because it requires extra work and it invites inconsistency.

Having structured flash card data could also support other tools--for instance, customizable flash cards (e.g., showing things like parts of speech and gender in English or in the language itself, or with/without examples), and would make export to other tools like Anki easier.

I first tried to figure out a way to use Wiktionary data, but (1) Wiktionary only works for words, not most phrases or sentences, and (2) Wiktionary content is too detailed and too loosely structured for a flash card. I thought of creating a Lua-based lookup table, but for long vocabulary lists that gets resource-intensive very quickly.

So my thought was: for each vocabulary item, create a Wikidata Item, which in pseudocode might look something like:

Item: aisling
language: ga
vocabulary type: lemma
(link to Wiktionary entry)

definition 1:
 meaning 1: dream, vision
 meaning 2: a type of poem
 part of speech: verbal noun
 gender: feminine
 declension: 2

definition 2:
 meaning: dream
 part of speech: verb
 conjugation: 2

Wikibooks/Wikiversity editors could then choose to display these as:

• aisling: dream, vision
• aisling: dream, vision; a type of poem (n.)
• aisling:
  • drean, vision; a type of poem (n. f. 2nd)
  • dream (v. 2nd)

and so on. I'd be using this initially on my own projects (in the Irish language) but would try to design it in such a way that the general syntax/taxonomy could be used for any language.

I guess my question is: is this the sort of thing Wikidata is for? I'm experienced with other wiki projects and with structured data generally, but new to Wikidata. Should I just Be Bold and start creating content and see if anyone complains? Are there projects out there already doing vocabulary-type data? Would it be a problem if the only "authority" for a given phrase or sentence was that it's used in a book written for Wikibooks? Any feedback would be appreciated. --Chapka (talk) 12:26, 6 December 2020 (UTC)

• In Wikidata we have different datatypes. Items are not appropriate for vocabulary but lexemes are. On Wikidata you mostly can't just make up your own general syntax. General syntax gets created via property proposals which involve community discussions. Take a look at our existing lexemes and where their syntax allows for what you want.

Citing authorities is not necessary for creating new lexemes. ChristianKl ❪✉❫ 12:47, 6 December 2020 (UTC)

• Thanks. I will poke around at the L entries and see if there's something I can latch on to there--although there are some problems I can see right away (e.g. words with different meanings, genders, etc. but identical spelling might need to be on the same flash card, and you might not want to pull in all possible obscure meanings for one vocabulary list). But it's a lot closer to what I'm looking for. --Chapka (talk) 19:12, 6 December 2020 (UTC)

  You could also have a flashcard per sense (a meaning). ChristianKl ❪✉❫ 21:55, 6 December 2020 (UTC)

Descriptions in the Wikipedia Mobile App

Hello, in the Mobleversion under the title of the Article the description from the Wikidataitem is shown. It is from my point of view for a reader not clear where the description comes from. I havent found a information at the page. After the principle of Wikipedia is that every one can edit it, this should be easy to do it with all the content shown at a page. Does somebody know if and where there is the description where the second line of an article comes from so that readers can understand it. Is there a help page or something like that for this. --Hogü-456 (talk) 21:56, 6 December 2020 (UTC)

• It's worth noting that in English the description is frequently overwritten and doesn't come from Wikidata but Wikipedia. ChristianKl ❪✉❫ 22:13, 6 December 2020 (UTC)

Conflation

Can someone decide if Louis Desnoyers (Q98908679) can be teased apart into the two conflated people? --RAN (talk) 06:53, 6 December 2020 (UTC)

• @Marcok: maybe you could help. See Help:Conflation of two persons. --- Jura 17:25, 6 December 2020 (UTC)

@Richard Arthur Norton (1958- ), Jura1: Fixed, thank you. --Marcok (talk) 10:31, 7 December 2020 (UTC)

2020 Coolest Tool Award Ceremony on December 11th

Hello all,

The ceremony of the 2020 Wikimedia Coolest Tool Award will take place virtually on Friday, December 11th, at 17:00 GMT. This award is highlighting tools that have been nominated by contributors to the Wikimedia projects, and the ceremony will be a nice moment to show appreciation to the tools developers and maybe discover new tools!

You will find more information here about the livestream and the discussions channels. Thanks for your attention, Lea Lacroix (WMDE) 10:55, 7 December 2020 (UTC)

Member of

For a mason do we say member_of=masonic lodge (Q1454597) or say member_of=freemasonry (Q41726) or say member_of=freemason (Q23305046)? If the exact lodge has an entry I use that, as here: Special:WhatLinksHere/Q64966925. --RAN (talk) 18:23, 20 November 2020 (UTC)

• If the information of the exact lodge is available I would create an item for it and use it. member_of=freemasonry (Q41726) looks better to me then member_of=masonic lodge (Q1454597). ChristianKl ❪✉❫ 18:29, 20 November 2020 (UTC)
  • Shouldn't the value of member of (P463) be an actual organization? Not convinced that the changed from "instance of" to "instance or subclass of" for the type constraint was a good idea [3]. --- Jura 11:38, 22 November 2020 (UTC)
    • The exact lodge is the only one that fits exactly, do you think we need to create something new that fits "member_of=" better, or can we modify one to be a better match? We need something to let people know someone was a mason, all the other fraternal organizations have an exact match such as "member_of=Phi_Bet_Kappa" or "member_of=Schiners" or "member_of=International_Order_of_Odd_Fellows" or "member_of=Benevolent and Protective Order of Elks". Currently we are split among the various choices I mentioned above. --RAN (talk) 00:45, 24 November 2020 (UTC)
      • While Phi Beta Kappa or the IOOF are each a single organization, the Masons are not. Most lodges are affiliated and recognize one another, but there are definitely exceptions (e.g. Continental Freemasonry). - 01:55, 24 November 2020 (UTC)~
    • I was of the understanding that the lodges do form an organization together, if that's not how Masons work then I agree that individual lodges should only be used. ChristianKl ❪✉❫ 14:07, 24 November 2020 (UTC)
      • As I say, most do, some don't. I don't know my way around this in detail; someone else probably does and could weigh in with a better explanation. A good example is that the Order of the Eastern Star accepts women as equal members (and always has), which the F&AM still don't allow. - Jmabel (talk) 21:03, 24 November 2020 (UTC)
      • Modelling individual lodges (even if we can reliably work that out) sounds like it could get very complicated - there are (were?) thousands of them, and presumably a lot will only ever have one person linked in Wikidata. 21:49, 24 November 2020 (UTC)
        • I don't see any problem with thousands of item for Masonic lodges. Even when in the beginning there's only a link to one person, as Wikipedia grows the amount of links will go and the information that two people were in the same Masonic lodge is very interesting information that previously wasn't documented. ChristianKl ❪✉❫ 13:51, 1 December 2020 (UTC)

I think the problem is that an obituary may only mention that a person was a "mason" and we need some way to show that, so we can aggregate all the masons in a single search, perhaps someone may want to compare membership numbers in various fraternal organizations. Currently we have a mix of the examples I gave in my initial question. Even if only temporary, we need to harmonize on one. I agree "member of=individual lodge" is best for the future, if someone wants to tackle adding them all. --RAN (talk) 22:35, 2 December 2020 (UTC)

@Richard Arthur Norton (1958- ):Given that we now have P8929 (P8929) it might be a good for a Mason. ChristianKl ❪✉❫ 18:05, 7 December 2020 (UTC)

What's elytral declivity (Q21660529)?

I'm trying to understand what elytral declivity (Q21660529) happens to be to give it appropriate statements. Can someone who understands Russian tell me what the thing is called in English or what it means? ChristianKl ❪✉❫ 15:23, 6 December 2020 (UTC)

Gtranslate claims it is an anatomical feature, a depression in w:en:Elytron allowing gnawed wood to be passed backwards relative to a bark beetle's path through the wood they're eating. "Paving the way, the bark beetles push the gnawed sawdust ( drill meal ) to their rear end of the body. Here they hit the wheelbarrow. Moving back, the beetle pushes out a portion of the drill meal out of the way." --Tagishsimon (talk) 15:32, 6 December 2020 (UTC)

Unfortunately, Gtranslate doesn't help me to understand it well enough to give it proper statements or translate the name. ChristianKl ❪✉❫ 15:36, 6 December 2020 (UTC)

At w:en:Ips (beetle) it is translated as "Concave elytra and spines". At https://biologicalsurvey.ca/ejournal/dcgs_38/Ips_emarginatus.htm тачка is translated as "Elytral declivity". You won't find anything better, because this word is based on slavic word "тачать" (sharpen/roll) and not used in modern language. --Lockal (talk) 14:44, 7 December 2020 (UTC)

Thanks. I now booked it as an organ part. ChristianKl ❪✉❫ 20:07, 7 December 2020 (UTC)

Wikidata weekly summary #445

Here's your quick overview of what has been happening around Wikidata over the last week.

• Discussions
  • Comment on a proposal for proposed deletion system

• Events
  • Upcoming: Wikidata Lab XXVI: Structured citations module on Wikipedia, presented in English by Mike Peel (Wiki Movimento Brasil). The event is supported by a WikiCite e-scholarship. December 8th 17:00 UTC (14:00 BRT), via YouTube.
  • Upcoming: live SPARQL queries on Twitch and in French by Vigneron, December 8 at 18:00 CET
  • Upcoming: Wikimedia 2020 Coolest Tool Award Ceremony, December 11th at 17:00 GMT (details)
  • Upcoming: Online Wikidata meetup in Swedish #42, December 13

• Press, articles, blog posts, videos
  • A massive new effort to name millions sold into bondage during the transatlantic slave trade. The Washington Post writes about enslaved.org, a Wikibase instance run at Michigan State University, that collects and creates Linked Open Data about the transatlantic slave trade. (Source)
  • Video: Introduction to Wikidata by and with Alicia Fagerving from Wikipedia Day 2020
  • Video: How to use Wikidata Mix'n'match tool - YouTube
  • Video: Semantic Web in Libraries 2020 (replay)
    • Linking K10plus library union catalog with Wikidata
    • Building Wikidata One Branch of Knowledge at a Time

• Tool of the week
  • User:Zvpunry/EntitySchemaHighlighter.js user script highlights the entities in the content of a Schema and shows information when hovering the link. (2020-12-07)

• Other Noteworthy Stuff
  • We now have a central repository for wikibase-cli templates, with the aim to document best practices, offer general purpose request/create/edit templates, but also offer more niche templates as examples of what's possible: contributions and template requests welcome!

• Did you know?
  • Newest properties:
    • General datatypes: identity element, mathematical inverse, has operator, has marker, Hong Kong film rating, indexed in bibliographic review, number of taxpayers, ITRANS, complementary property, number of points/goals attempted, water area, MTRCB rating, power consumption index, heating energy consumption index, road name formatter, ELSPA rating, kunya, religion or world view
    • External identifiers: FINA Wiki ID, Foodlocate restaurant ID, Qobuz label ID, AAA campground ID, AAA hotel ID, AAA restaurant ID, Hyperion Records person ID, Southwest Harbor Public Library item ID, DriveThruComics numeric publisher identifier, Fellow of the Royal Society of Canada ID, HLTV player ID, Kinorium person ID, AstroGen ID, BBC sound effects asset ID, Namuwiki ID, CIP data sheet, AniDB episode ID, edition humboldt digital Flora ID, Sistema de Información Legislativa ID, BG EIK, All the Tropes identifier, Concise Literary Encyclopedia ID, Reliwiki page ID, Arken ID, HuijiWiki Wiki ID, Archives at Yale names ID, HAL journal id, Parler username, Museo del Prado ID, National Gallery of Ireland ID, Auckland Art Gallery artwork ID, CINE21 film ID, National Gallery of Australia ID, Fitzwilliam Museum ID, Sonneveld-index, Kicker.de player ID (actual scheme), NSSDCA ID, ZOBODAT person ID, ZOBODAT taxon ID, DIA ID, SKD ID, Max Movie film ID, Gab username, Movist film ID, KOBIS ID, Amsterdam Museum ID, SMB-digital ID, State Catalogue of the Museum Fund of Russia artwork ID, CQVIP article ID, DBLP conference ID, Filmstriben film-ID
  • New property proposals to review:
    • General datatypes: cites work string, artistic director, MDJF ID, Coordinates of depicted place, first family name in Portuguese name, propriétés à éviter, Commons category for the view from the item, Mailing List, curid, provides HTML microdata, provides JSON-LD data, archaeological site of, does not have cause, conductor for, reason for normal rank, value is one of, tribunsdelaplebe.fr ID, lineage marker, has surface, DBLP journal id
    • External identifiers: U.S. Masters Swimming ID, Artfacts ID, Order of Canada ID, AppGallery app ID, Czech Gymnastics athlete ID, Students of Prague Universities ID, theses.cz ID, Encyklopedie mostů v Čechách, na Moravě a ve Slezsku ID, CollEx-Persée ID, The Galileo Project ID, Dizionario degli artisti ID, Museo Galileo biography ID, Museo Galileo authority ID, Biblioteca Marciana owner ID, Chinese Painting Database ID, National Galleries of Scotland ID, Bavarian State Painting Collections ID, Chapels of southern Bohemia ID, Kunsthistorisches Museum ID, Max-Movie-Personen-ID, CINE21 person ID, Georgia Museum of Art ID, Basic Formal Ontology ID, Fichier des décès ID
  • Query examples:
    • Russian cuisine
    • Belgian artists that rise to the public domain in 2021 (Source)
    • Asian food made from noodles (Source)
    • Bubble diagram by wiki page numbers in each language on objects kept at the Bardo National Museum (note: Commons is counted as English) (Source)
    • Amusement parks around the world (Source)
  • Schema examples: Video game

• Development
  • Simple Query Builder development continues. We worked on combining different query conditions as well as internationalisation.
  • Making the precision translatable in geocoordinate and date inputs (phab:T269102)
  • Changes to descriptions in a specific language now show up on watchlist and recent changes on the Wikipedia and co for that language (phab:T191831)
  • Finished working on the problem of incorrectly-escaped label being included in a link to Special:NewItem if there are no search results (phab:T263258)
  • Investigating why we're currently skipping a significant number of Item IDs (phab:T268625)
  • Wikibase Release Strategy/Infrastructure: We have decided where to host future Docker and tarball releases of Wikibase. You can read about the details and the reasoning in the following ADRs: Docker Image Repository, Tarball Repository

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

• Monthly Tasks
  • Add labels, in your own language(s), for the new properties listed above.
  • Comment on property proposals: all open proposals
  • Contribute to a Showcase item.
  • Help translate or proofread the interface and documentation pages, in your own language!
  • Help merge identical items across Wikimedia projects.
  • Help untangle items that refer to different concepts.
  • Help write the next summary!

Read the full report · Unsubscribe · Mohammed Sadat (WMDE) 17:30, 7 December 2020 (UTC)

electoral district (P768) = Vermont Class 1 senate seat (Q101435082), etc. cleanup

It seems that we didn't actually reach a consensus for changing the "electoral district" from states (e.g. Vermont (Q16551) for the above), to Vermont Class 1 senate seat (Q101435082) and countless similar items. Accordingly, the original model should be restore, i.e. Vermont Class 1 senate seat (Q101435082) replaced with Vermont (Q16551), etc.

Also, I don't think any of the requested references have been provided for statements like "Vermont Class 1 senate seat (Q101435082)" subclass of "United States senate constituency", suggesting that it's a database artifact.

Discussion was at Wikidata:Project_chat/Archive/2020/11#Representing_United_States_Senate_classes. There was some further discussion at Wikidata:Project_chat#Cleaning_up_old_model_for_US_Senate_data (mainly about something else).

If further information is needed about the cleanup, please bring it up. --- Jura 18:03, 6 December 2020 (UTC)

There's little room for doubt that it is Senators, not their (in our terms) districts, that are 'divided as equally as may be into three Classes' (United States Constitution Article 1 Section 3 Clause 2) and that the constituency is the State (United States Constitution Article 1 Section 3 Clause 1}. Vermont Class 1 senate seat et al remain as objectionable today as they did when coined a few weeks ago. --Tagishsimon (talk) 18:14, 6 December 2020 (UTC)

Also not sure where we are with the invented end-dates for incumbents, as if we possess a functioning crystal ball, and against convention & good sense. Whole thing quite depressing. --Tagishsimon (talk) 18:19, 6 December 2020 (UTC)

@Tagishsimon: All your questions were answered in the thread cited. If you were truly concerned you should have followed up there. Gettinwikiwidit (talk) 23:23, 6 December 2020 (UTC)

The end date problem was fixed in the meantime (see discussion above). We just need to keep an eye on these things and watch out for artifacts. --- Jura 18:43, 6 December 2020 (UTC)

This is nothing more than harassment. Please stop. You had ample opportunity to contribute to the discussion you cited. I tried to have a non-public discussion with you about your behavior which you refused to engage in.

For those late to this discussion the major points are that Jura has been complaining about this model for over two years but has put in no effort into the underlying data itself. The full Senate record was spotty, inconsistent and occasionally wrong before this effort was taken up. The entire Senate history is now complete and consistent with enough info to reproduce pages such as this one. This would have been impossible with the model Jura proposed and certainly with the data as it existed. Tagishsimon's complaint is basically a bike shed argument as should be clear from reading the thread cited. Both of these individuals have insisted on making personal insults a feature of their criticism despite repeated efforts to get them to engage specifically about the data itself. To be fair, Tagishsimon's criticisms have been more on point whereas Jura's have never been. If this continues, I will follow up as appropriate. Regards, Gettinwikiwidit (talk) 23:23, 6 December 2020 (UTC)

It's also worth noting that Jura did not ping me on this proposal though I have put in the bulk of the work here. This is not an attempt to contribute to the community but rather an attempt to harass an individual. Regards, Gettinwikiwidit (talk) 23:33, 6 December 2020 (UTC)

• I'm not sure it counts as "cleaning up" when this wasn't very widely applied in the first place. Of the ~2000 old-style claims still present, only ~600 have any P768 qualifiers, so less than one in three items actually used it. This is the first time a data model has actually been rolled out consistently across all Senators and actually makes the data useful. Personally, I think what we have is fine. If it's desired to return the state rather than the class seat then it's really easy to write a query to do so (example with state, example with class).

In the earlier discussion I wrote up some lengthy notes on the options for modelling classes and states at Wikidata:Project_chat/Archive/2020/11#Options_for_modelling but it didn't get much traction. I think any of them are fine, really, but I'd much prefer it if we showed that one of the alternatives is clearly better than what we have, not just that it's what was sort-of there before. Andrew Gray (talk) 23:51, 7 December 2020 (UTC)

Q6263868

This item is incorrectly described as a human. Cleanup needed.--GZWDer (talk) 20:28, 7 December 2020 (UTC)

It should simply be merged with John Whitfield Bunn and Jacob Bunn (Q104008585), a sibling duo. VIAF will adjust in time, we needn't bend over backwards to treat every mistake on the internet like the infallible Word of God. -Animalparty (talk) 22:43, 7 December 2020 (UTC)

Help disentangle a conflation please!

Luigi Bassi (Q1751014) --RAN (talk) 19:59, 9 December 2020 (UTC)

•  Done  – The preceding unsigned comment was added by Richard Arthur Norton (1958- ) (talk • contribs).

I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Matěj Suchánek (talk) 11:29, 13 December 2020 (UTC)

MediaWiki:Queryservice

Could someone create MediaWiki:Queryservice/sv with the text Frågetjänst? --Sabelöga (talk) 01:32, 11 December 2020 (UTC)

Done. Stryn (talk) 09:24, 11 December 2020 (UTC)

Thanks! --Sabelöga (talk) 11:36, 11 December 2020 (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. Matěj Suchánek (talk) 11:29, 13 December 2020 (UTC)

Same or different?

Are Jane Loury (Q95699137) and Jane Loury (Q3161875) the same person. --RAN (talk) 06:28, 11 December 2020 (UTC)

Yes. Merged. --Tagishsimon (talk) 08:55, 11 December 2020 (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. Matěj Suchánek (talk) 11:29, 13 December 2020 (UTC)

Accidentally created a new item where there is already an entry - please delete

Hi, I just created counter-insurgency because I overlooked (spelling with "-" somewhat irritated me) that there is a respective item with the exact same title already, also describing the same subject. Hence, please delete my new one. Pittigrilli (talk) 12:09, 11 December 2020 (UTC)

• On Wikidata we don't delete items when there are doublicate items but merge them. You can do that yourself as explained on https://www.wikidata.org/wiki/Help:Merge ChristianKl ❪✉❫ 13:00, 11 December 2020 (UTC)

Success! Thank you, Pittigrilli (talk) 15:44, 11 December 2020 (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. Matěj Suchánek (talk) 11:29, 13 December 2020 (UTC)

Merge needs to be untangled

At Catherine Mann (Q75624136) three people were merged and one was incorrect, can someone braver than me figure out which one to undo. The incorrect person was "Harriett Jessop". --RAN (talk) 18:35, 11 December 2020 (UTC)

•  Done  – The preceding unsigned comment was added by Richard Arthur Norton (1958- ) (talk • contribs) at 12 December 2020‎ (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. Matěj Suchánek (talk) 11:29, 13 December 2020 (UTC)

Proposed deletion

Welcome to comment on a proposal for proposed deletion system in Wikidata.--GZWDer (talk) 17:46, 4 December 2020 (UTC)

• At the moment it's a draft of a proposal that's in motion and no proposal. ChristianKl ❪✉❫ 08:12, 8 December 2020 (UTC)

When should an item use position held (P39) and/or noble title (P97)?

It appears that Wikidata items inconsistently use position held (P39) and noble title (P97):

• Frederick III, Holy Roman Emperor (Q150966)
  • P39 only: Holy Roman Emperor (Q181765)
• Baldwin IV of Jerusalem (Q296850)
  • P97 only: King of Jerusalem (Q693614)
• Rollo (Q273773)
  • Both P39 and P97: Duke of Normandy (Q460244)
• Elizabeth II (Q9682)
  • P39 only: duchess (Q4593319)
  • P97 only: monarch of the Dominion of Ceylon (Q34518803)
  • Both P39 and P97: monarch of the United Kingdom (Q9134365)

Could I please have some guidance on when to use P39 only, when to use P97 only, and when to use the same value for both properties? -Thunderforge (talk) 02:13, 7 December 2020 (UTC)

@Thunderforge: I would use both, even though redundant in these cases. I would also make sure each noble_title is also listed as a position as I did at "King of Jerusalem". See: Talk:Q693614 where positions automatically create succession charts. For a completed chart see for example: Talk:Q11696. Almost all succession of positions are full of errors at Wikidata and need to be cleaned up. --RAN (talk) 21:15, 8 December 2020 (UTC)

Importing references from Wikipedia articles?

Is there any solution to automatically import works (to create Wikitada items from them) referenced on Wikipedia to Wikidata? --TRANSviada (talk) 19:23, 8 December 2020 (UTC)

Object with more than one role with conflicting properties

The question here is how to handle cases where one and the same thing has different roles, where for each role some of the properties would have different values.

My current example: there are protected areas (nature protection) which are under national protection (nature reserve (Q1627961)) and international protection (Special Area of Conservation (Q1191622)) at the same time. It's exactly or nearly the identical area, and of course it's covered by a single Wikipedia article and a single Commons category. However, the national and internation protection have different date of inception (P571), different WDPA ID (P809) und sometimes a slightly different area (P2046).

When I asked in the German-speaking project chat I was told to make two distinct Wikidata objects and connect them via territory overlaps (P3179). I tried that and naturally wanted to have a Wikidata Infobox for both Wikidata items in the commons Category, which yielded an error. Asking about that, I was told that the two items should be merged and one of the protection status should be included as has part(s) of the class (P2670), assigning all the properties of that protection status as qualifiers to that property. Even though I don't like the asymmetry of the two roles, and it seemed to me to contradict the intended semantics of has part(s) of the class (P2670), I tried it and it led to warnings about properties not being allowed as qualifiers.

So now I try to ask here, which I consider the "highest authority" for such questions. I really wonder what's the correct way of handling such a situation.

When I thought about it further, I came to the conclusion that a possible solution could be something similar to heritage designation (P1435), just for protection of nature rather than protection of cultural heritage. A single object could have more than one value for that property, and IDs like WDPA ID (P809), Natura 2000 site ID (P3425), and Common Database on Designated Areas ID (P4762) should be allowed as qualifiers. Probably we'd also want a qualifier like "affected area" or "protected area" for cases where the exact protection area differs slightly.

What do you think? --Reinhard Müller (talk) 12:11, 8 December 2020 (UTC)

They must be different items here, its not possible to put all the data into qualifiers of one statement. Take for example the area (P2046) - often the law which created the protected area states one roundabout area, then later the map is digitized and a slightly different value shows up, then eventually some parts are added to the protected area, and there's another value with a different start time (P580). If all is in one statement with lots of qualifiers, its impossible to show different values with different references or different times. Its never a good idea to model the data to fit with a technical limitation - maybe someone can create a commons infobox with the item ID as parameter, and then its easy to show both in one category. Ahoerstemeier (talk) 13:37, 8 December 2020 (UTC)

Ahoerstemeier is right. Conflating multiple different entities into one item is generally bad. ChristianKl ❪✉❫ 23:55, 8 December 2020 (UTC)

Property...MediaAppearance...

Hey folx - what is the good way to add a link to a web media appearance (online published TV report) into Wikidata profile of a person? Zblace (talk) 15:17, 8 December 2020 (UTC)

present in work (P1441) would fit the bill, as far as I can tell. (It's a bit of a mess, and I just tried to align the description(s) with the constraints)

You'd need some item for either the TV program (qualified with date and full work available at URL (P953)) or the specific episode. --Matthias Winkelmann (talk) 02:51, 9 December 2020 (UTC)

Linking a redirect?

I noticed that Four Seasons Total Landscaping (Q101248727) was successfully linked to Wikipedia, even though the page on Wikipedia is a redirect to Four Seasons Total Landscaping press conference (Q101424698). I'd like to do the same for some other pages, such as Ruth Chandler Williamson Gallery (Q18325652) and w:Ruth Chandler Williamson Gallery, but the documentation at w:Template:Wikidata redirect is lacking and I can't escape the error message it keeps giving me when I try to make the connection. Help? {{u|Sdkb}} ^talk 21:13, 29 November 2020 (UTC)

• At the moment work on the ticket that allows easy creation is stalled (@Mohammed_Sadat_(WMDE): might provide information on when this changes). In the meantime you have to deactivate the redirect on the Wikipedia page shortly then set the redirect in Wikidata (at a time where the Wikipedia page is not a redirect) afterwards you can reedit the Wikipedia page to turn it into a redirect. ChristianKl ❪✉❫ 21:49, 29 November 2020 (UTC)

@ChristianKl: There was an RfC to allow the addition of some redirects to items. However, we need to make sure that people are aware of the consequences that change will bring. Lydia wrote some more about that here. The next steps might not have been clear there, but in order to move forward with the ticket that you referred, we would like to have a look at the option to generate the links based on statements first. For that, we’ll need to collect a few specific cases where people want to link to redirects and then see which statements could provide a large number of them without too much overhead. Who can help collect such a list? -Mohammed Sadat (WMDE) (talk) 12:30, 5 December 2020 (UTC)

@Mohammed Sadat (WMDE): @Lydia_Pintscher_(WMDE): Whether or not sitelinks to redirects are allowed seems to me like a policy question and thus up to the community. Is it the WMDE position that it's not up to the community to decide on policy like this? ChristianKl ❪✉❫ 12:54, 5 December 2020 (UTC)

@ChristianKl: In what sense do you think the RfC that Mohammed Sadat linked to was not an expression of the will of the community? --Tagishsimon (talk) 14:41, 5 December 2020 (UTC)

@Tagishsimon: I think it was an expression of the will of the community to use sitelinks to redirects. Mohammed Sadat (WMDE) seems to argue that we should disregard it and instead see whether we can archieve the functionality of referercing redirects (or the pages they target) with mechanisms other then sitelinks. ChristianKl ❪✉❫ 14:45, 5 December 2020 (UTC)

Ah. You're relying on a very specific definition of sitelink, then. Don't get me wrong; if we are to have sitelinks to redirects, then having them in the same place as existing language wiki links, and marked (and hence machine-distinguishable) in much the way that we currently mark featured articles, would seem to be the way forward. I'm not sure I've seen any articulation of why that approach is technically problematic. Blessed if I know why Lydia launched into a riff on using statements to generate redirect sitelinks (adding redirects seems to be something users would do once the facility is provided) rather than addressing the question of storage arrangements for the data. --Tagishsimon (talk) 15:10, 5 December 2020 (UTC)

@Sdkb: If you would like this to be fixed, please vote for meta:Community Wishlist Survey 2021/Wikidata/Link Wikipedia redirects to Wikidata items. Kaldari (talk) 03:20, 9 December 2020 (UTC)

Fixing position held (P39) claims for historical US House of Representatives delegates

The following items are listed as having position held (P39) United States representative (Q13218630), but according to their entries in the Biographical Directory of the United States Congress they were merely delegates for territories not yet admitted to the Union. Should we have an explicit subclass of non-voting member of the U.S. House of Representatives (Q5253588) for such positions or simply use this entity? Current delegates have entries like this.

• Anthony Dimond (Q573153)
• Antonio Borja Won Pat (Q600957)
• Bob Bartlett (Q719965)
• Charles August Sulzer (Q1063605)
• Charles Debrille Poston (Q1064080)
• Coles Bashford (Q882895)
• Curtis Coe Bean (Q1145545)
• Daniel D. Tompkins (Q223545)
• Daniel Elliott Huger (Q1160550)
• Daniel Sutherland (Q1162848)
• Donna Christian-Christensen (Q461604)
• Edward Dexter Holbrook (Q1292048)
• Eleanor Holmes Norton (Q461649)
• Elizabeth P. Farrington (Q1331285)
• Eni Faleomavaega (Q729199)
• Fofó Iosefa Fiti Sunia (Q3880333)
• Francisco Perea (Q1384732)
• Frank Hinman Waskey (Q1443435)
• George Barnes Grigsby (Q1506947)
• George W. Jones (Q438644)
• Granville Henderson Oury (Q1543279)
• Henry Dodge (Q881703)
• Henry Hastings Sibley (Q880319)
• Hiram Sanford Stevens (Q1620134)
• James Wickersham (Q1681295)
• John A. Burns (Q880606)
• John Frank Wilson (Q1525522)
• John G. Campbell (Q1700145)
• Jonah Kūhiō Kalanianaʻole (Q1702876)
• Joseph Rider Farrington (Q1263994)
• Madeleine Bordallo (Q292988)
• Marcus A. Smith (Q1894073)
• Melvin H. Evans (Q2819651)
• Michael San Nicolas (Q16221559)
• Napoleon Bonaparte Giddings (Q1964936)
• Norton P. Chipman (Q2819666)
• Oakes Murphy (Q889353)
• Orrin Dubbs Bleakley (Q7104081)
• Ralph H. Cameron (Q976300)
• Robert A. Underwood (Q2156310)
• Robert D. Carey (Q436936)
• Robert William Wilcox (Q1133170)
• Ron de Lugo (Q2819582)
• Samuel Gordon Daily (Q2218316)
• Stacey Plaskett (Q18739104)
• Thomas Cale (Q1741474)
• Thomas D. Singleton (Q2423319)
• Vicente T. Blaz (Q2819706)
• Victor O. Frazer (Q2522517)
• Walter E. Fauntroy (Q2819659)
• William Woodbridge (Q883246)

Gettinwikiwidit (talk) 10:15, 6 December 2020 (UTC)

In the short term, I've made these all non-voting member of the U.S. House of Representatives (Q5253588). I personally would rather see all entries like Victor O. Frazer (Q2522517) use the same position held (P39) claim and use located in the administrative territorial entity (P131) to refer to which territory they represent, but my immediate concern is not to co-ingle these entries with United States representative (Q13218630) to make it easier to do a proper reconciliation of the latter. Gettinwikiwidit (talk) 07:37, 9 December 2020 (UTC)

I think there are three questions:

• (1) are they representatives? The answer is probably no, so non-voting member of the U.S. House of Representatives (Q5253588) shouldn't a subclass of United States representative (Q13218630).
• (2) is a specific position with just members from different territories? If the answer is yes, non-voting member of the U.S. House of Representatives (Q5253588) shouldn't have any subclasses. If they vary, we should probably have subclasses.
• (3) has the situation for (1) or (2) changed over time? If not, just one approach is sufficient.

--- Jura 17:24, 6 December 2020 (UTC)

WikiCite question

Can I add links as a QID for citations? For example, if I wanted to cite https://noticias.uol.com.br/ultimas-noticias/efe/2018/05/10/edificio-prestes-maia-o-maior-simbolo-das-ocupacoes-na-america-latina.htm , could I do it? Tetizeraz (talk) 01:18, 9 December 2020 (UTC)

Tetizeraz Yes, absolutely! For use here on Wikidata, it's more common not to use references to use reference URL (P854) in the reference section. But you can also create an item: see ‘Star Wars: The Force Awakens’ Leads Saturn Awards Nominees (Q24090324) for one of currently about 20,000 examples. Unfortunately, there isn't any automation that I'm aware of. You would need to identify it as appropriate, such as instance of (P31)article (Q103184) (or something more specific, such as letter to the editor (Q651270), news article (Q5707594) etc.) and provide the URL with full work available at URL (P953). I'm not familiar with how these are used on other wikis, but on Wikidata, you'd then link to it with stated in (P248).

In principle, it is fairly easy to fill in additional information such as authors, publication dates, and publication automatically, at least for professional publications that cointain such data in the source code for search engines, twitter, etc. (and title, as well, but you'd usually add that because the Item creation form asks for it). With that in mind, and for use within Wikidata, I'd tend to suggest to prioritize adding data that cannot be gathered automatically, such as quotation (P1683) before creating additional items for the publication, authors, etc. Opinions might differ on that, however. --Matthias Winkelmann (talk) 02:17, 9 December 2020 (UTC)

Matthias Winkelmann something like this? Prestes Maia Building, the greatest symbol of squatting in Latin America (Q104048129) Tetizeraz (talk) 06:23, 9 December 2020 (UTC)

Tetizeraz Yes, that is excellent! It's probably more detailed than the average item for heads of state. Since I'm a fan of such overkill, I've added a few things, such as creating an item for the author.

The one complaint the system had was the archive URL, which it expects as a qualifier on the main URL. But the archive URL can always be derived if needed, so it's not strictly neccessary. Most useful among the things I adde is probably main subject (P921), because these can't be derived automatically, and it's something that people are likely to search for. If you click on "What links here" on the building's item, it now links the article, as well. --Matthias Winkelmann (talk) 11:51, 9 December 2020 (UTC)

Finally fixing the Bonnie and Clyde problem

If you would like to see the Bonnie and Clyde problem finally fixed on Wikidata, please go vote for Certes' proposal at meta:Community Wishlist Survey 2021/Wikidata/Link Wikipedia redirects to Wikidata items. This would allow us to finally link to redirects without using an elaborate hack. (And yes, canvasing is allowed for the Community Wishlist Survey.) Kaldari (talk) 03:26, 9 December 2020 (UTC)

• I wrote a policy page that lays out how we will use the badges to tag redirect in Wikidata and clarified remaining ambiguities. This is likely the more direct way to get change. The RfC is at https://www.wikidata.org/wiki/Wikidata:Requests_for_comment/Adopt_Help:SitelinksToRedirects_as_policy . ChristianKl ❪✉❫ 11:41, 9 December 2020 (UTC)

How to make redirects in main namespace?

How do you make/edit redirects in main namespace? Take for examle, Q56167159, it just says "Redirect to: Q37487459." However, Special:EditPage/Q56167159 doesn't do anything, so how to make or edit redirects? --Rqkp (talk) 05:45, 9 December 2020 (UTC)

Redirects can be created using Special:RedirectEntity (I'm not sure whether it can also change the target, you can try it) or as a result of merging by a tool. --Matěj Suchánek (talk) 09:13, 9 December 2020 (UTC)

In general you don't create redirect manually. They are the result of merges and mostly done automatically by the merge gadget. ChristianKl ❪✉❫ 11:08, 9 December 2020 (UTC)

Wikidata:Requests for permissions/SPARQL endpoint

Somebody created this page, but I don't think this is a proper name - Wikidata:Requests for permissions is for requesting specific user rights. Please recommand a new name for the page.--GZWDer (talk) 16:16, 8 December 2020 (UTC)

I guess Wikidata:Contact the development team/SPARQL endpoint would be better? ArthurPSmith (talk) 18:41, 8 December 2020 (UTC)

@Zache: why are you duplicating Wikidata:SPARQL federation input? Multichill (talk) 19:31, 8 December 2020 (UTC)

@Multichill: I was thinking that the page is permanently defunctional and a replacement was needed. The last approval Wikidata:SPARQL federation input was from June 2019 and the only comment after that was in November 2019. Currently, page tagged with the text This page is currently inactive and is retained for historical reference and there is no activity and there is also phab:T265290. In the very minimal project chat discussion in October 2020 the only comment was by you with something like bot approvals would be ok as a process. So this was my take on for pushing things forward and I was going to take this to project chat too for the comments after I had done the pages first. --Zache (talk) 08:56, 9 December 2020 (UTC)

@ArthurPSmith: I dont have strong opinions on what the page names should be and any name would be ok for me. --Zache (talk) 08:56, 9 December 2020 (UTC) (edited own comment --Zache (talk) 10:44, 9 December 2020 (UTC))

@Zache: the problem laid out in the phab task is that we don't have a community norms about adding new endpoints. Those community norms don't get created by creating a new endpoint page. We have a policy for creating property proposals and we need a similar policy for adding WDQS endpoints.

To write a good policy one, we need knowledge about the costs about WDQS endpoints. The policy page on adding WDQS endpoints likely should explain that. After a policy is written it should be approved via RFC. ChristianKl ❪✉❫ 16:34, 9 December 2020 (UTC)

From the Wikidata:SPARQL federation input page the rules were:

The suggested SPARQL endpoints must satisfy the following conditions:

1. Complies with the SPARQL 1.1 protocol, "query operation" part, at least to the extent necessary to make federated SERVICE clause work (most SPARQL endpoints do).
2. Contains data that can be linked to Wikidata - i.e., either contains Wikidata IDs or can be queried by values contained in one of the Wikidata properties.
3. Has data freely available under license compatible with CC0 (preferred) or other free database license allowing unrestricted reuse. Attribution licenses like CC-BY are ok too. Currently, we do not accept endpoints with reuse restriction clauses like NC/ND.

For the rule #3 it is proposed that if Wikimedia Commons accepts the license then it would be ok. Least CC0, CC-BY, CC-BY-SA, and ODbL are used in currently allowed endpoints. Technically cost of adding on outside of the community discussion is creating the phabricator ticket phab:T200066 and updating the whitelist. There may be also technical limitations why Wikimedia tech wants to deny or postpone adding new endpoints but reviewing that is outside of the community discussion scope and this selection happens after community approval. In any case is there some reason why these rules aren't enough for RfC and if not what we need more? --Zache (talk) 19:24, 9 December 2020 (UTC)

Why can't I find lexemes in the statement value search box?

I'm having trouble linking lexemes to each other using statements.

Here's the most recent issue: I've created a lexeme Lexeme:L347938 for the Irish proverb "Georraíonn beirt bóthar." I want to link this lexeme to each of the three individual words. I create a "Has part" property and search for "beirt", an independent lexeme I know exists, and I can't find it. Searching by the lexeme number, same. Nor does it come up searching the other way round (trying to add a "part of" statement to Lexeme:L348027.

It looks to me like the search function in the statement interface isn't searching the lexeme namespace--is there a way around this? --Chapka (talk) 15:40, 9 December 2020 (UTC)

@Chapka: the value of has part(s) (P527) may only be items and this property is not for lexemes. combines lexemes (P5238) is the correct property to use in lexeme. Note there are no inverse property as a word (or affix) may be found in hundreds of phrases.--GZWDer (talk) 16:52, 9 December 2020 (UTC)

Dates related to appointments

Hello, there. We have the appointed by (P748) property, but not an "appointed on" leaving some ambiguity as to what start time (P580) is referring to. Take Kirsten Gillibrand (Q22222) for instance who was appointed on January 23, 2009 but the appointment "took effect" on January 26, 2009 and she took the oath of office on January 27, 2009. [1]. Or Rebecca Latimer Felton (Q271243), who was appointed on October 3, 1922 but took the oath of office on November 21, 1922. [2]. There are other cases such as Tom Stewart (Q2440140) who was elected on November 8, 1938 but did not take the seat until January 16, 1939. [3]. The same for Huey Long (Q314384). Is it worth having separate properties for "appointed on" and "sworn in on" and maybe even "elected on" to distinguish from "start time" (which I take it should mean "assumed the duties on" when used with position held (P39))?

I had the same question about popes, we have the start date for some as their election and some the start date as their consecration. --RAN (talk) 00:28, 10 December 2020 (UTC)

I'd assume that start time (P580) is supposed to be the date where an appointment takes effect. This also affects awards, where an award winner may be announced at one time, and the award handed out at a ceremony some time later. But that won't stop the "wrong" value being used in a lot of cases. E.g., the British New Year honours, where the date of "investiture" may not be readily available. Ghouston (talk) 02:21, 10 December 2020 (UTC)

I'd suggest using announcement date (P6949) for the other date. Ghouston (talk) 02:23, 10 December 2020 (UTC)

For awards (which is all a bit off topic), it could also be argued that the presentation is a mere formality, and the date of the award is when it's announced. I'm also puzzled by a couple of Victoria Cross plaques, such as c:File:John_Whittle_memorial_Cygnet_20201114-012.jpg, where the date 9-15th April 1917 refers to the heroic deed itself, not the subsequent date when the award was approved, nor the subsequent investiture. Ghouston (talk) 02:50, 10 December 2020 (UTC)

Protection on Q557839

There's been vandalism and other changes from IPs and new editors at Mirza Masroor Ahmad (Q557839). This could use another set of eyes. This item could use temporary protection. gobonobo ^{+ c} 21:51, 13 December 2020 (UTC)

It has been  Done for one month. --Epìdosis 17:52, 15 December 2020 (UTC)

This section was archived on a request by: --Epìdosis 17:52, 15 December 2020 (UTC)

List of all avaiable language codes

I remember there was all language codes but I forgot where it was. I will save it this time. Eurohunter (talk) 16:49, 9 December 2020 (UTC)

You may be looking for Help:Wikimedia_language_codes/lists/all. Bovlb (talk) 00:42, 10 December 2020 (UTC)

@Bovlb: Thanks. I think it was the other list but this one will be okey too. Eurohunter (talk) 12:06, 10 December 2020 (UTC)

@Bovlb: Do you have idea how to add/use en-us? Eurohunter (talk) 12:19, 10 December 2020 (UTC)

Langcom opposed adding en-us and thus it's not available. ChristianKl ❪✉❫ 13:12, 10 December 2020 (UTC)

/header's, /footer's, and the exclusion criteria

In Wikidata:Administrators' noticeboard#TemplateStyles pages, there is a disagreement about consensus over the inclusion (notability) of "/header" and "/footer" subpages of templates. Since 2013, Wikidata:Notability/Exclusion criteria explicitly listed any /header or /footer pages and there is still an abuse filter which actively prevents adding them as links. Peter James believes it was always wrong and only subpages in Project and Help namespaces should be excluded. Opinions? --Matěj Suchánek (talk) 13:47, 10 December 2020 (UTC)

What reference source is the most deprecated

I have been fixing errors and noticed that some sources have more typos than others. Is there a way to search for our most deprecated source ... typos and other errors that have to deprecated because they are never corrected at the source. See for instance Michele Cerimele (Q55836816) where I deprecated a date because of a typo in our source: "Archivio Storico Ricordi". I fixed a dozen or so typos like that today. --RAN (talk) 07:52, 10 December 2020 (UTC)

Thanks for this question. I wish I knew the answer. Jc3s5h (talk) 15:52, 10 December 2020 (UTC)

Help wanted for one single data structuring...

Hi, I have tagged the Commons file File:Suharto_at_funeral.jpg. I found out that more abstract values (historical) connected to the image can be added as "Significant Event (P793)". However, if I add the most important data in this manner, which is Q799299, I get an exclamation mark indicating:

Values of significant event statements should be instances or subclasses of one of the following classes (or of one of their subclasses), but Indonesian killings of 1965–1966 currently isn't:

   occurrence
   fictional event
   mythical event
   temporal entity
   accident
   event

It seems as if one of the above, probably event or temporal entity (happened during a period of 2 years), could be added to Q799299. I failed on that one, however, I do not know where to add it. If my basic concept is correct, please advise, else, other advice appreciated ;-) Pittigrilli (talk) 14:13, 10 December 2020 (UTC)

@Pittigrilli: So the issue was that Crushing The Rebellions of 1965–1966 (Q799299) wasn't categorized as an instance/subclass of anything. I added it as an instance of mass killing (Q56514238) which should be an event. Now it should be recognized as a proper significant event (P793). BrokenSegue (talk) 16:33, 10 December 2020 (UTC)

@BrokenSegue: Yes, now it is exactly as I wanted it to be. Thanks a lot. btw: How do I mark your action with a "Thank you" here? Did not find the option yet. Cheers, Pittigrilli (talk) 16:55, 10 December 2020 (UTC)

You would go to the history page of the item. So here. BrokenSegue (talk) 17:28, 10 December 2020 (UTC)

Tiktoker

Can I create item for a tiktoker? --2001:B07:6442:8903:DDE3:82F0:635A:782D 18:01, 10 December 2020 (UTC)

yes but they must meet our standards for notability (being a tiktoker isn't sufficient). Maybe read Wikidata:Notability. BrokenSegue (talk) 18:32, 10 December 2020 (UTC)

tiktoker is Alexuh1 --5.169.197.187 18:37, 10 December 2020 (UTC)

Are there other reliable sources on the internet that talk about him? If not, he isn't notable. ChristianKl ❪✉❫ 18:43, 10 December 2020 (UTC)

WikiProject Territorial Entities

The latest situation on West Germany (Q713750) where there's disagreement about what kind of items we should create for states suggests to me that we need common standards for that. Situations where there are cities and items for settlements in the same location similarly need common standards. I created WikiProject Territorial Entities as a place for having those discussions. ChristianKl ❪✉❫ 21:28, 10 December 2020 (UTC)

Might be also a good idea to ping WikiProject Former countries (Q15304996) Bouzinac ^{💬●✒️●💛} 21:43, 10 December 2020 (UTC)

@Bouzinac: As as I see there's currently no such Wikiproject on Wikidata. There are similar issues surrounding cities and towns where boundaries change and I see it best when those issues are addressed together in one Wikiproject. ChristianKl ❪✉❫ 21:47, 10 December 2020 (UTC)

Agriculture

We haven’t focused much energy on building out our coverage of agriculture. There are lots of opportunities for merges, multilingual labels, external ids, etc. in the areas of agricultural tools and processes. Is anyone interested in a formal WikiProject:Agriculture? If not, I’ll putter along. - PKM (talk) 22:01, 28 November 2020 (UTC)

@PKM: I looked at the profession tree from farmer (Q131512) last year - some discussion at Wikidata:Project_chat/Archive/2019/05#Farmers_and_agriculturers?. But I didn't manage to do much with the underlying field itself, and I found there was a lot of cross-language complexity to tease out (eg in some "farming" can primarily indicate arable, in others pastoral+arable). Definitely worth looking at. Andrew Gray (talk) 14:45, 29 November 2020 (UTC)

@Andrew Gray: Thanks for that link. We had a longish conversation on Commons Telegram recently about classifying types of hoes - this is an area with lots of overlapping sets across languages and cultures. (I notice Getty AAT has at least one term that is specifically Spanish-Chile.) I think we’d need a large team of participants to really make headway. - PKM (talk) 21:30, 29 November 2020 (UTC)

@PKM:: I would be interested in exploring this more. Just wondering, is there a case for starting a WikiProject and seeing who adds themselves? Pauljmackay (talk) 15:22, 5 December 2020 (UTC)

@PKM: Sorry for the delayed reply - if you do try and put together a project working on this, please do give me a shout. Andrew Gray (talk) 21:46, 10 December 2020 (UTC)

@Pauljmackay, Andrew Gray: Okay, I'll set up a starter site and see who wants to play! Watch this space. - PKM (talk) 21:55, 10 December 2020 (UTC)

@Pauljmackay, Andrew Gray: New project page at Wikidata:WikiProject Agriculture. - PKM (talk) 23:20, 10 December 2020 (UTC)

Clause or provision in a law

At failed election provision (Q104100163) I can't find a proper instance_of for a clause or a provision of a law, all the available ones involve contracts, and it seems silly to create a new version just for one instance of an Act of Congress. How is this handled in others, I can't seem to find other examples because I can't properly define what it is an instance_of, there must be a proper word, or should we expand clause and provision to be part of laws, as well as contracts. Can anyone help? --RAN (talk) 23:55, 10 December 2020 (UTC)

• The cost of creating a new item for this is low when you don't find an existing one. It's easy to merge in case there is another item. ChristianKl ❪✉❫ 01:17, 11 December 2020 (UTC)

Cleaning up old model for US Senate data

As suggested here by @MisterSynergy:, I plan on removing position held (P39) United States senator (Q4416090) which do not supply a term. The only information contained in such statements not contained in those which do supply a parliamentary term (P2937) was spotty and occasionally inaccurate. The statements which do supply a parliamentary term (P2937) contain the complete historical record (excluding the results of the most recent election). Still to do is to supply elected in (P2715) but that is pending an organization of election in the United States (Q279283) which should distinguish between scheduled general elections and unscheduled elections in a rational way and perhaps have a clear model to indicate midterm elections. I'll probably do this work at the end of next week if there are no objections. Regards, Gettinwikiwidit (talk) 03:21, 29 November 2020 (UTC) @Tagishsimon: @Oravrattas: @Andrew Gray:

So you want to delete the information about the current senators?

Can you provide us with a sample edit? How do you make sure that other data is complete?

As this provides us with a way of querying the current composition of the senate (e.g. position holder with a start date, but no end date), I don't think it's a good idea to delete these statements before you fixed the other statements (deleting fictive dates, as previously mentioned).

To sum it up: No data should be deleted before we can check the same information is available otherwise. --- Jura 05:29, 29 November 2020 (UTC)

current members. Gettinwikiwidit (talk) 10:45, 29 November 2020 (UTC)

@Gettinwikiwidit: Looks good, but I have also been wondering about end dates.

Sitting senators in the current term all have their end dates set to 3 January 2021 (except two no-P582, oddly), and presumably once we index the next term, they would get 2021-23 dates. However, there's a widespread convention to indicate positions which are currently held by setting start time (P580) but not end time (P582) qualifiers, even if a future end date can be predicted with some confidence. A lot of queries use this as shorthand ("presence of a P582 = not current") in much the same way that a birthdate but no deathdate is used to infer "person is alive". So I'd agree with dropping P582 qualifiers on the current members, which will make a traditional-assumptions query like this one predictably return sitting members with new-style data. (If we do have a need to use predicted end dates, we can pick that up easily enough by querying the end date of the term itself.)

Otherwise, definitely happy to remove all the "old-style" single statements. And then I think we'll have a pretty much consistent dataset, give or take a bit of error correction - congratulations! Andrew Gray (talk) 14:36, 29 November 2020 (UTC)

@Andrew Gray: I'm agnostic about end time (P582)s. I think it's always up to the person generating the query to be careful in interpreting the data since anyone can edit the data set. When the full data set was generated they all had an end time (P582). When I first started looking at the data in the old model it was also inconsistent and I had to figure out how to deal with future end dates. Deleting them now just means they will need to be added again in a couple of months. If someone leaves early the end time (P582) will have to be edited either way. As long as the future date can reasonably be assured I think supplying it is a defensible position. I think leaving them out is also defensible. Further, I don't own this data. I'm not sure who's job it will be to add the end date in the future. That person is free to remove them now to make work for themselves in the future. Gettinwikiwidit (talk) 21:56, 29 November 2020 (UTC)

It is yet more making it up as you go along, Gwwi. There are not distinct districts. 'end date' != 'anticipated end date'. Your caveat reporter point is absolutely the case if the content of statements bears little resemblance to the plain intent of the properties. --Tagishsimon (talk) 22:17, 29 November 2020 (UTC)

@Tagishsimon: It's not clear what value is added by turning this into a trading of insults. It's a fact that this is a volunteer effort. With that comes realities that need to dealt with. From a process stand point it's not clear that future dates will ever be added. Having them there when they're reasonably predictable means it's more likely to *remain* accurate in the future. My guess is that people jumping on all the "popular pages" will hand edit them to keep them up-to-date but the rest will be left to rot. No future dates are ever realized. If it were really a problem they could simply make it impossible to supply them. Gettinwikiwidit (talk) 22:27, 29 November 2020 (UTC)

Not for nothing, but harassing people contributing data makes it even less likely that they'll put in the effort to maintain it which further makes the case for having them there. Gettinwikiwidit (talk) 22:32, 29 November 2020 (UTC)

@Gettinwikiwidit: Personally, I would say the benefits of consistency here probably outweigh the minor hassle of having to update it in a few weeks (especially since we'll have do do a batch of edits then anyway to get all the new members in place, so closing off the older ones isn't much extra work). I'm happy to put a note in the calendar and take responsibility for sorting that out in January. Andrew Gray (talk) 23:05, 29 November 2020 (UTC)

@Andrew Gray: From a process standpoint this seems like unnecessary busy work for the reasons outlined above, but as I say, I'm agnostic on the topic. Feel free to remove them now and add them back in later. In any event this has no bearing on the topic at hand of removing the old, inconsistent and occasionally inaccurate claims. Gettinwikiwidit (talk) 23:32, 29 November 2020 (UTC)

Sure - I'll run this tomorrow if it won't cause any problems for you. (didn't want to blunder in and do it if it would complicate the ongoing work!) No objection to removing the older data, as noted. Andrew Gray (talk) 23:35, 29 November 2020 (UTC)

As suggested, I've now removed the end time (P582) values for sitting Senators which were set as 3 January 2021. This query confirms a simple question for "Senators with a known start time but no P582" returns 100 distinct people as expected - most are matching both the new and old statements, but it also works with only the new-style ones using parliamentary term (P2937). Andrew Gray (talk) 19:39, 30 November 2020 (UTC)

Thanks for your cleanup, Andrew. I think we really need to be careful to avoid speculative end dates. Senators are not immortal nor should we include user's predictions. If one wants to check when a term ends, that should be on the item for the term itself. I noticed that finally deceased senator(s) are no longer marked as being "in office" at Wikidata either.

As for the initial proposal, can we see a static list of claims to be deleted? Otherwise, once deleted, there is just no way of checking if what was done meets was proposed or agreed. --- Jura 17:43, 6 December 2020 (UTC)

Broadly speaking it will be these claims, which are all on items that have a more comprehensive set of statements. I am doing some manual checks at the moment to ensure that where the old data has date qualifiers, they match up with the new data - this will take a few more days. It won't be deleted until that's sorted. Andrew Gray (talk) 00:28, 7 December 2020 (UTC)

@Jura1: It's odd that you make the point that senators are not immortal since supplying and end date would seem to make exactly that case. In fact if no one maintains the data going forward you might be lead to believe that the senators term never ended. Gettinwikiwidit (talk) 02:00, 7 December 2020 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Quick update: manual checks complete, and these are now ready to be deleted (list, 1896 statements). I've run a report comparing all of the existing statements with start/end dates against the new-style ones, manually fixing one or the other where necessary, and they are now all look safe to move. The only remaining discrepancies are down to an ambiguity over whether terms ended on 3 March or 4 March in the early years and these will be fixed up when that's standardised in a seperate round of edits. Andrew Gray (talk) 22:43, 8 December 2020 (UTC)

Thanks very much. I'll delete the statements from this query tomorrow. Regards, Gettinwikiwidit (talk) 21:58, 9 December 2020 (UTC)

This is done. Thanks again to @Andrew Gray: for all his help cleaning up the data and cross-checking against the old model. Regards, Gettinwikiwidit (talk) 11:18, 11 December 2020 (UTC)

Encyclopedia article items

@GZWDer: is currently adding a gazillion items for encyclopedia articles (example: Furlanetto, Matteo (Q104083209)) without identifying the main subject or the encyclopedia in which the article is found.

I'm failing to see how these articles are notable. They do not have sitelinks. They are not notable in the sense that they can be described using serious and publicly available references (other, I suppose than each encyclopedia's index). They do not seem to fulfill a structural need - at least, one which could not be served by described at URL (P973) in the main subject's item.

My view is that Wikidata is not made better by the accretion of crud; and that these items are cruddy. Should we be adding such items, or deleting them?  – The preceding unsigned comment was added by Tagishsimon (talk • contribs).

large imports should be approved. I don't see that the User:LargeDatasetBot was approved to import this dataset. The policy suggests that each independent action be approved separately. Bot should be stopped until this is approved. BrokenSegue (talk) 04:24, 10 December 2020 (UTC)

GZWDer Do you any have plans to attempt to connect the encyclopedia articles to the respective subject (Benezit ID (P2843) would help)? Do you have plans to add essential missing structured data like published in (P1433) --> Benezit Dictionary of Artists (Q24255573) or useful descriptions like "article in Benezit Dictionary of Artists" (descriptions in any language would be an improvement from the ambiguous/no-context labels)? We've all seen your work before. -Animalparty (talk) 05:16, 10 December 2020 (UTC)

• I have brought this thread to GZWDer's attention. In general, it is preferable to raise issues like this directly with a user in the first instance.

Unless we're going to get a lot more information on these items, it seems to me that this sort of import would be better embodied in an identifier property. And yes, it would be good for the community to have advance notice of such large-scale imports. Bovlb (talk) 05:33, 10 December 2020 (UTC)

"connect the encyclopedia articles to the respective subject" - this will be done soon. "how these articles are notable" - an article is itself a topic, which can be cited in Wikipedia article. Also, this allows using Wikidata search to find relevant potential topic. The data is in Mix'n'Match, but Wikidata provides more information (e.g. author). Furthermore, currently some of these article are described as scientific articles, which is not correct; import them will prevent errorous imports in the future. If there are further significant oppose I will stop this task.--GZWDer (talk) 07:11, 10 December 2020 (UTC)

@GZWDer: I don't know enough about this import (e.g. where it's coming from, what further edits you will make, etc) to know if I would object. That's the point of seeking approval first. Yes, the approval process is kinda broken now (I waited over a month for my last BOT approval) but just breaking the rules because others are or its annoying isn't helpful. BrokenSegue (talk) 16:25, 10 December 2020 (UTC)

Many people import all sorts of things without discussion, sometimes causing controversies. However, there are few people active at the bot approval process. Some users raise their concerns, but do not comment further.--GZWDer (talk) 19:49, 10 December 2020 (UTC)

Yes and many people should not be importing lots of things without discussion. The bot approval process is poor but again that doesn't mean we shouldn't use it and try to improve it. BrokenSegue (talk) 20:46, 10 December 2020 (UTC)

@GZWDer: Good practices for imports (which you mention below) include: making them as complete as possible in the first attempt, so that humans and machines alike can efficiently identify, contextualize, and build upon them. SourceMD does an admirable job of scraping all essential bibliographic properties (most of the time) from a simple DOI. Your LargeDatasetBot is also often very good at being complete enough. What is less helpful is when thousands of incomplete items are generated that fail to import essential properties like source for the Benezit articles, or birth/death dates for Geni.com and genealogics.org imports, or descriptions in any language. -Animalparty (talk) 20:23, 10 December 2020 (UTC)

• @GZWDer: A better question of how these articles are notable, would be how the bot approval justifies their import. ChristianKl ❪✉❫ 13:24, 10 December 2020 (UTC)
  • Actually, many things are imported to Wikidata without any discussion at all. This is not very bad (many imports are valid), but not optimal either. We need to define good practices for imports. Users can raise concerns, but it is more important that how these concern should be resolved. Currently the resolution process is defective and may be stalled easily.--GZWDer (talk) 14:15, 10 December 2020 (UTC)
    • I agree a document explaining "good practices for imports" would be nice to have. Is there no such document already? If not we could draft it up. BrokenSegue (talk) 20:46, 10 December 2020 (UTC)
      • Not quite on point, but there's some interesting stuff at Wikidata:Dataset_Imports and the pages it links to. Also, issues related to import have come up many times in the past on this page and WD:AN. Bovlb (talk) 21:34, 10 December 2020 (UTC)
        • Some tasks are not about specific dataset to import, e.g. Wikidata:Requests for permissions/Bot/RegularBot 3 (please read in conjunction with my comment on 14:15 and 19:49).--GZWDer (talk) 23:06, 10 December 2020 (UTC)
        • The general problem with Wikidata:Dataset_Imports is that it creates a lot of pages about in process data imports that nobody reads and that aren't filled out https://www.wikidata.org/wiki/Category:In_Progress_datasets

They don't really help. ChristianKl ❪✉❫ 14:30, 11 December 2020 (UTC)

Community Wishlist Survey 2021

We invite all registered users to vote on the 2021 Community Wishlist Survey. You can vote from now until 21 December for as many different wishes as you want.

In the Survey, wishes for new and improved tools for experienced editors are collected. After the voting, we will do our best to grant your wishes. We will start with the most popular ones.

We, the Community Tech, are one of the Wikimedia Foundation teams. We create and improve editing and wiki moderation tools. What we work on is decided based on results of the Community Wishlist Survey. Once a year, you can submit wishes. After two weeks, you can vote on the ones that you're most interested in. Next, we choose wishes from the survey to work on. Some of the wishes may be granted by volunteer developers or other teams.

We are waiting for your votes. Thank you!

SGrabarczuk (WMF)

15:03, 11 December 2020 (UTC)

Newbie question

I joined the project two days ago. I worked, amongst other items, on counter-insurgency and would like to have a brief confirmation on what I did, just to be sure. I translate the Wikidata terms into everyday language to test if my understanding is right:

I added items to "subclass of" which are related to the subject in a manner such as "counter-insurgency can be abstractly described as:" a military strategy; a political strategy; etc.

I added items to "instance of" which are related to the statement "counter-insurgency can be/is": warfare; political repression; human rights violation; etc.

Coming to a specific question: "property of" would then imho be used for "who does/practices counter-insurgency", leading to items like: military; intelligence agency; special forces; counterinsurgency unit; etc.

I would appreciate a brief assessment, especially on my last question just above. Pittigrilli (talk) 16:17, 11 December 2020 (UTC)

I think typically "instance of" is reserved for statements which are more definitive. Counter-insurgency can be a human rights violation but it is not definitionally. You have the same issue with some of your subclasses. "counter-insurgency" isn't a kind of "military intelligence" (they are just related concepts). I think political strategy is arguable. "who does/practices counter-insurgency" is more for practiced by (P3095) not what you are saying. I made some edits to what I think is more appropriate. BrokenSegue (talk) 18:09, 11 December 2020 (UTC)

• Neither "subclass of" nor "instance of" are for adding all related terms.

Tagging an item both covert operation (Q1546073) and black operation (Q3394581) is wrong. All black operation (Q3394581) are covert operation (Q1546073) and thus you don't add additional information to an item that's already tagged as black operation (Q3394581) by adding covert operation (Q1546073).

There might be items that are both human rights violation (Q11376059) and counter-insurgency (Q760037) but given that not all counter-insurgency (Q760037) necessarily are human rights violation (Q11376059) it doesn't belong there. Items inherit all the subclasses of those classes that they reference via subclass of (P279).

When it comes to improving the quality of the items black operation (Q3394581) are covert operation (Q1546073) a key question given that covert operation (Q1546073) is more narrow is: What is true for all black operation (Q3394581) but not true for all covert operation (Q1546073). The description should answer that question so users know when to use black operation (Q3394581) and when to use covert operation (Q1546073). Wikidata is a lot about thinking how individual items relate to each other and not so much about focusing on one item like you would focus on one Wikipedia article that you want to bring to featured status.

In cases where a term as it's commonly used can mean multiple different things, the solution isn't to add different things into the same item but to create multiple items for the distinct meanings.

Currently, counter-insurgency aircraft (Q5176890) has use (P366) counter-insurgency (Q760037) is one way we say who uses counter insurgency. An item for "counterinsurgency unit" would use the same pattern. Items like "counterinsurgency unit" would usually be created for structural need. If we have an item about an individual counterinsurgency unit there we would need a instance of (P31).

I hope I could make things a bit more clear. If you have additional questions I'm happy to address them. ChristianKl ❪✉❫ 18:11, 11 December 2020 (UTC)

And maybe "counterinsurgency unit" practiced by (P3095) counter-insurgency (Q760037) would be more appropriate then has use (P366). ChristianKl ❪✉❫ 18:14, 11 December 2020 (UTC)

ChristianKl and BrokenSegue: Thank you very much. This clarified a lot. I have to change my approach coming from Wikipedia even more than I already thought... Pittigrilli (talk) 19:43, 11 December 2020 (UTC)

Archivio Storico Ricordi

Archivio Storico Ricordi (Q3621644) is being used as a reference by a bot, but I have corrected about a dozen typographical errors where dates have numbers transposed. Should we stop using it as a source adding in redundant information since it is so error prone, I can spot egregious errors with Wikidata:Database reports/items with P569 greater than P570 and Wikidata:Database reports/unmarked supercentenarians. It may be adding in more subtle ones. See for example: Johann Wilhelm von Mirbach-Harff (Q94785517) and Gabriel Astruc (Q3093605) and Paul Joseph Guillaume Hillemacher (Q2060201) and Giuseppe Taddei (Q471393). I will have to figure out a search to find the ones I corrected previously and did not track at Wikidata:WikiProject Authority control/Archivio Storico Ricordi errors. If we had a contact there, we could send them the errors to correct, like we used to be able to do at VIAF and LCCN. --RAN (talk) 18:18, 11 December 2020 (UTC)

Green Map icons

I am assembling a mapping between the terms used in the Green Map icons and Wikidata items. There may be 1-to-many mappings in some cases as some of the Green Map icons can represent quite general or multiple concepts. I am thinking to use described by source (P1343), where the source item would be Green Map Icons (Q103892210) and the reference URL (P854) would be the icon links found here. Ideally the icons would be uploaded as a collection to Commons but the license is not compatible. Does this sound like a suitable approach? Any alternative methods that could work better? Thanks! Pauljmackay (talk) 01:19, 12 December 2020 (UTC)

Political Graveyard

See Alfred Thomas Rogers (Q94293371) where the Political Graveyard ID gets an error message. We need to change the format to allow a number after the alphabetical surname, common surnames are broken up into multiple pages. --RAN (talk) 03:07, 12 December 2020 (UTC)

Onderzoek naar het Verlanglijstje van de Wikipedia-gemeenschap 2021

We nodigen alle geregistreerde gebruikers uit om te stemmen op de 2021 Community Wishlist enquête. Je kunt vanaf nu tot aan 21 december stemmen voor zoveel verschillende wensen als u wilt.

In de enquête worden wensen voor nieuwe en verbeterde tools voor ervaren gebruikers verzameld. Na de stemming doen wij ons best om aan jouw wensen tegemoet te komen. We beginnen met de meest populaire.

Wij, het Community Tech-team, zijn een van de teams van de Wikimedia Foundation. We maken en verbeteren bewerkings- en moderatie-instrumenten voor wiki's. Waar we aan werken wordt besloten op basis van de resultaten van de Community Wishlist Enquête. Eenmaal per jaar kun je wensen indienen. Na twee weken kun je stemmen op de wensen die je het meest interesseren. Vervolgens kiezen we wensen uit de enquête om aan te werken. Een deel van de wensen kan worden ingewilligd door vrijwillige ontwikkelaars of andere teams.

We wachten in spanning op je stem. Dank je!

SGrabarczuk (WMF)

15:05, 11 December 2020 (UTC)

Mendeley Profiles will be deprecated

See https://blog.mendeley.com/2020/11/02/weve-listened-to-our-users-and-are-refocusing-on-whats-important-to-them/. The property is Mendeley person ID (P3835).--GZWDer (talk) 15:59, 6 December 2020 (UTC)

we should archive them all and not delete the property. BrokenSegue (talk) 19:29, 6 December 2020 (UTC)

I archived about half of them on archive.org before they closed down the system in the last few days. I'm marking the property deprecated. BrokenSegue (talk) 04:06, 12 December 2020 (UTC)

User:Paptilian , Scope of contributions which need experience, data structure tutorials, global interaction

Itemized at User:Paptilian. Items that may need to be created.Paptilian (talk) 15:00, 12 December 2020 (UTC)

• What's your question? ChristianKl ❪✉❫ 15:00, 12 December 2020 (UTC)

Which Wikiprojects is responsible for GeoNames (Q830106) issues

We frequently have people that notice issues involving Ceb-Wiki and GeoNames (Q830106) sourced items. I think it would be great to have a page giving guidance for those issues and likely it could be located at a Wikiproject. Which is the best Wikiproject for it? ChristianKl ❪✉❫ 12:35, 10 December 2020 (UTC)

In my opinion the largest issue is 1. many pages in cebwiki are not linked to items and some people oppose further mass import for future duplicates; 2. Not many people care the unconnected pages in cebwiki; 3. People continue to import (eventual) duplicate from different source, and since Wikidata does not search unconnected pages, the number of duplicates will be higher and higher. If there is my choice, I will mass import all unconnected articles from cebwiki then discuss the next step. I don't like the current situation.--GZWDer (talk) 19:46, 10 December 2020 (UTC)

I think most people prefer not important more cebwiki articles. If the unconnected cebwiki articles are ignored that's great. The problem is with those that are connected to Wikidata. ChristianKl ❪✉❫ 21:21, 10 December 2020 (UTC)

"Ignored" does make more duplicates - Wikidata exists as a place that problems may be found and fixed. What I hope in long term is once Abstract Wikipedia is a thing most cebwiki articles will be migrated to that.--GZWDer (talk) 23:33, 10 December 2020 (UTC)

Do not mass-import cebwiki articles. >90 % of unconnected Cebwiki articles are duplicates or spurious. The data quality is atrocious. Existing entities in Wikidata that were created before someone had mercy and stopped these imports accounted for weeks of subjectively useless busywork I had to do this summer when I was trying to use Wikidata for actual work. That discussion you mention needs to happen before any import. If you feel any content on cebwiki represents actually existing entities, you're free to link them up with a reputable source. --Matthias Winkelmann (talk) 15:15, 11 December 2020 (UTC)

Many of data from Geonames actually comes from one of these: GNIS Feature ID (P590) or GNS Unique Feature ID (P2326).--GZWDer (talk) 18:27, 11 December 2020 (UTC)

That still leaves hundred thousand or millions of items that don't come from those. Important data from GNIS Feature ID (P590) which is public domain would be much better to import it from GeoNames which isn't.ChristianKl ❪✉❫ 15:06, 12 December 2020 (UTC)

• An awful situation where all the vague GeoName entries were imported, that we purposefully ignored. Many are synonyms for existing entries where there is not enough information to be either useful on their own, or have enough info to merge into the existing entry. At one time we debated importing all the VIAF entries, and we did not because of the very same reason, most that we do not have already are poorly defined and will never have enough information to be able to disambiguate properly or merge. If we reargue VIAF entries only import ones with a defined birth or death date, or both, so we avoid what happened at GeoNames. --RAN (talk) 00:05, 11 December 2020 (UTC)
  • At very least, they have a valid sitelink, so they fall within WD:N. However, it will be useful to document what problems Geonames have (e.g. where does data come from; are there always a primary source for information; etc.) For VIAF, I think we should create a secondary Wikibase instance that VIAF data may be imported to.--GZWDer (talk) 00:14, 11 December 2020 (UTC)
    • They might be notable but that doesn't mean that a bot job that adds a lot of doublicate items is reasonable. Anybody who wants to add items from CebWiki can do so if they check not to add doublicates. ChristianKl ❪✉❫ 14:26, 11 December 2020 (UTC)

Doubts about mobile phones

When reviewing the mobile phones data I have noticed some inconsistencies. With this query:

SELECT DISTINCT ?phone ?phoneLabel ?instanceOf ?created ?published
WHERE {
  {
    # Mobile phone
    ?phone wdt:P279 wd:Q17517.
  }
  UNION
  {
    # Smartphone
    ?phone wdt:P279 wd:Q22645.
  }
  UNION
  {
    # Phablet
    ?phone wdt:P279 wd:Q521097.
  }
  UNION
  {
    # Basic cell phone
    ?phone wdt:P279 wd:Q965424.
  }
  UNION
  {
    # Modular smartphone
    ?phone wdt:P279 wd:Q18611337.
  }
  UNION
  {
    # Dual SIM phone
    ?phone wdt:P279 wd:Q1262537.
  }
  UNION
  {
    # Dual SIM smartphone
    ?phone wdt:P279 wd:Q91702726.
  }
  ?phone wdt:P31 ?instanceOf.
  OPTIONAL { ?phone wdt:P577 ?published . }
  OPTIONAL { ?phone wdt:P571 ?created . }
  # Uncomment the following to show items that are not an instance of a manufactured object (or that are instances of other things)
  #FILTER(?instanceOf != wd:Q10929058).
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en" }
}

Try it!

Two things can be observed:

1. Most phones are an instance of product model (Q10929058) but more than 400 are instances of one of the following items:
   • mobile phone form factor (Q2636061)
   • system (Q58778)
   • cell phone model (Q19723444)
   • mobile phone series (Q20488450)
   • electronic device model (Q62008942) (removed)
   • model series (Q811701) (removed)
   • electronic machine (Q2858615) (removed)
   • brand (Q431289)
   • Wikimedia disambiguation page (Q4167410) (this can be omitted as it is a valid property)

What would be the correct value for instance of (P31)? I suppose product model (Q10929058) since it is in most of the phones but I don't know if there is a consensus about it.

2. Isn't inception (P571) and publication date (P577) somewhat redundant? Using the GSMArena phone ID (P4723) as a reference, I would think that the inception (P571) corresponds to the moment the phone was announced and publication date (P577) would be when it was released for sale, but from what I've been seeing it seems that both properties are used indistinctly to indicate the date when the phone became available to the public.--Swicher (talk) 01:55, 1 December 2020 (UTC)

Personally, I'd go with product model (Q10929058) for a single model and model series (Q811701) for a family or models (which will have subclasses which are product model (Q10929058).) For the dates, I think the most relevant are announcement date (P6949) and date of commercialization (P5204), since hardware isn't really "published" and inception (P571) presumably predates any public announcement, but is typically unknown. Ghouston (talk) 04:40, 1 December 2020 (UTC)

• Being both a Wikimedia disambiguation page (Q4167410) and another page is generally not valid. ChristianKl ❪✉❫ 12:38, 1 December 2020 (UTC)
• A case like this needs thinking about how mobile phones be modeled and that model written down in a Wikiproject where it can be discussed and referenced in the future. ChristianKl ❪✉❫ 12:38, 1 December 2020 (UTC)

The following answer goes to both Ghouston and ChristianKl but it is easier for me to make a single message instead of two separate ones for each one.

It seems to me that product model (Q10929058) is quite generic compared to more descriptive such as cell phone model (Q19723444) and smartphone model (Q19723451). In the case of Samsung Galaxy A51 (Q84849860) it has all three values together but I don't know if it should be kept like this or delete everything and leave only smartphone model (Q19723451) as the value of instance of (P31). One way these properties could be structured would be the following:

	Basic cellphone/feature phone/dumbphone	Smartphone	Unknown
instance of (P31)	cell phone model (Q19723444)	smartphone model (Q19723451)	cell phone model (Q19723444)
subclass of (P279)	feature phone (Q965424)	smartphone (Q22645)	mobile phone (Q17517)

The unknown column would cover the phones that are not well known what type they are (for example, when importing data) and the value product model (Q10929058) could be kept or removed/depreciated.

Other details that I noticed were the following:

• There are some phones that are instances of Android One (Q18063397) (see a list in Special:WhatLinksHere/Q18063397) but I think that it should be moved to a more appropriate property (class (P2308) or family (P53) perhaps?).
• Later I will start to change model series (Q811701) to mobile phone series (Q20488450) in some items since this value is more descriptive. I will do it manually since there seems to be normal phones mixed with the models.  Done
• With respect to the following items I have my doubts as to whether the subclass is correct but I'm not sure what to replace it with:
  • clamshell design (Q1532758)
  • DIGNO (Q11195846)
  • Cocoa Touch (Q2628895)
  • Personal Handy-phone System (Q1718639)

Regarding dates, I agree with Ghouston's idea, but with QuickStatements can you change inception (P571) to announcement date (P6949) and publication date (P577) to date of commercialization (P5204) in all items automatically? Because according to what I was seeing, this tool seems to only can be used to add or remove properties but not to "rename" them keeping the original values.

About how the phone data should be modeled, the wikiprojects that could perhaps cover this would be Electronic Components and Informatics but I don't think that cell phones are its scope.

I also updated the query to include smartphone (Q22645) and feature phone (Q965424) and a few days ago I was doing some cleaning on some items.--Swicher (talk) 13:38, 6 December 2020 (UTC)

---

I updated the query to include some additional elements and a few days ago I also changed model series (Q811701) to mobile phone series (Q20488450) (or smartphone model series (Q71266741)).

Regarding instance of (P31) and subclass of (P279) I am not going to make major modifications since there is no clear consensus on how this should be handled, the only thing I found on this subject was the following discussions:

• This talk in iPhone item
• Wikidata:Project chat/Archive/2019/09#Should iPhone 5s, model A1533 (Q66816688) be linked to smartphone (Q22645) via subclass of (P279) or instance of (P31)
• Wikidata:Project chat/Archive/2020/05#Android smartphone model

--Swicher (talk) 13:19, 13 December 2020 (UTC)

Recon service is down.

Hi the recon service at https://wikidata.reconci.link/ is down. Does anyone know if there is a backup we can use? Thanks --Nikola Tulechki (talk) 11:21, 10 December 2020 (UTC)

@Nikola Tulechki: I've been using this one for OpenRefine for some time now, is it down too? https://wdreconcile.toolforge.org/en/api Vojtěch Dostál (talk) 13:59, 13 December 2020 (UTC)

New property navbox

I havd created a new module, Module:Property navigation, which may generate a navbox from a list of properties stored in a structural way (such as Module:Property navigation/sports). This is only the first step, and this module does not support all types of property navboxes yet. The layout of navbox keeps the same and is not redesigned.

Soon I will introduce some check so that the information at property talk page, property page and navbox will be consistent (i.e. talk pages of all properties listed in the navbox should include the navbox, and vice versa; also, properties listed in the navbox should have specific metadata). Then, I will clean up properties which do not have a navbox.--GZWDer (talk) 23:57, 13 December 2020 (UTC)

A problem on Q1922067

E4024 has undone my edit here. Then I started discussing the issue with him/her. Finally he/she ended the discussion unilaterally. I think it isn't correct to delete an information with the source before it is discussed enough. Can you help on this situation? Uncitoyen (talk) 16:18, 11 December 2020 (UTC)

I do know that Germany would be an example where the head of government is also member of the cabinet. The term "prime minister" would, in itself, indicidate that these are similar. Turkey, of course, still has a prime minister as well as a president, IIRC, and it would make sense if only the former is a member of the cabinet.

An English-language source would probably help in sorting this out. Although we strife to be language-agnostic, it is, unfortunately, difficult.

As it currently stands, the claim remains and it's only the reference that was deleted? Let's loop E4024 in here to maybe shed some light. --Matthias Winkelmann (talk) 17:53, 13 December 2020 (UTC)

The president is not part of any cabinet and I have not seen any reference that says what this user claims. If I claim something I bring a ref, I cannot bring a ref for something that does not exist. Thanks. E4024 (talk) 18:08, 13 December 2020 (UTC)

[4], [5] As you can see in the English news, the cabinet now meets under the chairmanship of the President of Turkey. It managed by a presidential system in Turkey now. A government without a president of Turkey can't be formed. Naturally, the president is part of the cabinet. Previously presidents of Turkey can't establish a cabinet and couldn't administer over the cabinet. Q11696 It says the US President that he is part of the US Government. Uncitoyen (talk) 19:23, 14 December 2020 (UTC)

Hoping this is not out of line...

I'm hoping this post is not out of line, because I realize that this page is "not about me." I've been a moderately active participant here this last few years, but too much is happening in my life right now for me to continue to do so. I'll be pretty much exclusively on Commons, and maybe not even there as much as usual. If anyone actually needs to get hold of me here, please ping; I won't be reading this daily. - Jmabel (talk) 09:32, 13 December 2020 (UTC)

• @Jmabel: it would likely be more effective to post such a message on your user page because that's were people will look who might want to get a hold of you. ChristianKl ❪✉❫ 15:02, 13 December 2020 (UTC)
  • @ChristianKl: Very good thought. - Jmabel (talk) 16:58, 14 December 2020 (UTC)

Help with mapframe

Hello, I'm trying to make sure that every NJ municipality is tagged with wikidata information in OpenStreetMap and Wikidata. In order to do this I made a mapframe map to see them all. The query itself is working, but for some reason some entries are not reflecting while others are not. I made this page to debug what's happening w:User:Acebarry/sandbox/debug. The problems are in the southern area of the state. I have not started work on the northwest corner yet. I tried/checked the following:

• All wikidata entries have P402
• All OSM entries point to the proper wikidata entry
• The query returns the elements I expect
• The relations in OSM are valid (all parts of relation are "outer"/relation is fully connected)

As an example Q3014914 was entered around the same time as the bordering municipalities but it is not showing up. Any help or suggestions are appreciated! Acebarry (talk) 14:58, 13 December 2020 (UTC)

Acebarry, From personal experience, after adding OpenStreetMap relation ID (P402) it may upto a week to build map from it. ‐‐1997kB (talk) 15:19, 14 December 2020 (UTC)

Wikidata weekly summary #446

Here's your quick overview of what has been happening around Wikidata over the last week.

• Discussions
  • New request for comments: Adopt Help:SitelinksToRedirects as policy

• Events
  • Upcoming: live SPARQL queries on Twitch and in French by Vigneron, December 15 at 18:00 CET
  • Past: m:WikiConference North America/2020
  • Past: Localization and structuring of open data-Zhu Tinghong-Wikidata Community of Local Studies - MOPCON 2020 (replay), YouTube
  • Past: SMWCon 2020/Day 2 - "Wikidata Walkabout" by Yaron Koren (replay), YouTube
  • Past: Wikidata Lab XXVI: Structured citations module on Wikipedia (replay), YouTube

• Press, articles, blog posts, videos
  • Building the SWIB20 participants map - combining Wikidata with custom data points via a federated query (ZBW Labs).
  • Introducing the Mapping the Scottish Reformation Website - A new Wikidata-driven website to explore the Scottish Reformation.
  • Introduction to Querying Wikidata Knowledge Graph using SPARQL by John Samuel
  • Video: LIVE Wikidata editing #25 - climate change, YouTube
  • Video: How to take data from OpenRefine to Wikidata through QuickStatements, YouTube
  • Video: Editing Wikidata live - casual session by Jan Ainali, YouTube

• Tool of the week
  • datao.net is a visual tool for querying and exposing Wikidata easily.

• Other Noteworthy Stuff
  • The Coolest Tool Award 2020 featured some Wikidata tools: Lingua Libre won in the category “Diversity”; Listeria, Lexeme Forms and Entity Explosion got honorable mentions. Congratulations!
  • Complex constraint check for recency of value
  • The voting phase of the Community Wishlist Survey 2021 is open until December 21st. You can vote on as many wishes as you want. There's a separate category for Wikidata.
  • Template:Notebook Given name is a new template which gives for each given name a list of useful SPARQL queries. The template can be used on the talk page of given names. See Alexandre (Q16002466) for an example. Comments are Contributions are welcome.
  • The next Wikibase live session is 16:00 UTC on Thursday 17 December 2020 (Add to calendar). We will have a few people share out about their project in the first half of the call and then continue our conversation from the last session about what will be a better platform for community discussions. Everyone is welcome to attend!

• Did you know?
  • Newest properties:
    • General datatypes: category for the view from the item, mailing list archive, word lookup, inappropriate property for this type, compatible with
    • External identifiers: Church of Jesus Christ missionary ID, Encyklopedie mostů v Čechách, na Moravě a ve Slezsku ID, Czech Gymnastics athlete ID, Encyclopedia of the History of Ukraine ID, CollEx-Persée ID, theses.cz ID, Google Play developer slug, AppGallery app ID, The Galileo Project ID, Artfacts ID, Biblioteca Marciana owner ID, Museo Galileo biography ID, National Galleries of Scotland ID, Museo Galileo authority ID, Bavarian State Painting Collections artwork ID, Delft municipal monument ID, Dizionario degli artisti ID, Order of Canada recipient ID, Students of Prague Universities ID, vets.cz ID, Chapels of southern Bohemia ID, GS1 GPC brick code
  • New property proposals to review:
    • General datatypes: insufficiently precise value, OpenReview.net group ID, OpenReview.net submission ID, OpenReview.net profile ID, applies when property is used, River Code, Wikimedian in Residence, Commons compatible image available at URL (non-artwork), number of at bats, runs batted in, stolen bases, subdivision monétaire
    • External identifiers: PBA ID, Medieval Manuscripts in Oxford Libraries IDs, LFH staff ID, Cephalopod Ontology entity ID, Fungal gross anatomy entity ID, Plant Ontology entity ID, Hymenoptera Anatomy Ontology entity ID, DataTrek ID, DBLP publication ID, PeriodO period ID, KANTO, CITWF title ID, CITWF person ID, identifiant de la Cinémathèque québécoise pour une personne, Encyclopedia Titanica ID, DzygaMDB film, DzygaMDB person ID, XXXBios female performer ID, XXXBios transgender performer ID, AVN movie ID, Lichess Username, Evil Angel movie ID, Evil Angel video ID
  • Query examples:
    • All stumbling stones in the city of Hannover
    • (partial) list, timeline and gallery of successful ascents of Mount Everest (and few other summits)
    • Web & social media URLs for UK Civic Societies (Source)
    • Publication related to Research Institute of Text Analysis and Applications: no. of pages per author/year, authors network (Source)
    • Map of the ritual objects said to be used by accused witches in Scotland (colourcoded layer accessible on top right to see the list of objects) (Source)
    • Seats of the Parliament of England, and the first year we have a named person listed for them (Source)
    • Female race car drivers on Spanish Wikipedia (Source)
    • Places of Geological Interest in Alamedilla (Source)
    • Number of Nature articles that were published on a Sunday in a given year
  • Schema examples:
    • Schema for stumbling stones
  • Newest WikiProjects:
    • Agriculture

• Development
  • We spend the week doing some exploratory work for the next year. Among other things we looked into how to measure the number of constraint violations on an average Item and what we can learn from it.

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

• Monthly Tasks
  • Add labels, in your own language(s), for the new properties listed above.
  • Comment on property proposals: all open proposals
  • Contribute to a Showcase item.
  • Help translate or proofread the interface and documentation pages, in your own language!
  • Help merge identical items across Wikimedia projects.
  • Help write the next summary!

Read the full report · Unsubscribe · Mohammed Sadat (WMDE) 17:41, 14 December 2020 (UTC)

Alma Claude Burlton Cull

This is a man.

A. B. Cull (Q17084969) is the same person as Alma Claude Burlton Cull (Q21455905)

They need to be merged or whatever... -- Broichmore (talk) 18:07, 14 December 2020 (UTC)

Broichmore  Done. See Help:Merge. -Animalparty (talk) 18:23, 14 December 2020 (UTC)

Currencies : start date and end date

Hello, is there more appropriate properties for currencies, better than start time (P580) / end time (P582) ? I was thinking about inception (P571) / dissolved, abolished or demolished date (P576) but there are other dates properties whose definition are close : date of commercialization (P5204) discontinued date (P2669) ? Bouzinac ^{💬●✒️●💛} 09:35, 14 December 2020 (UTC)

I'm pretty sure inception (P571) / dissolved, abolished or demolished date (P576) are the right choice here. My understanding is: Deutsche Mark (Q16068)inception (P571)1948-1-1, Saxony (Q1202)currency (P38)Deutsche Mark (Q16068)start time (P580)1990-1-1 500 euro note (Q2503713)discontinued date (P2669)2014-1-1 --Matthias Winkelmann (talk). 02:09, 15 December 2020 (UTC)

Merge?

metastatic melanoma (Q55013644) and metastatic melanoma (Q55779849) and metastatic melanoma (Q18975855)? One says obsolete, but I am not sure why. --RAN (talk) 00:07, 15 December 2020 (UTC)

Yes. Merged. I think "Obsolete' = obsolete in that ontology. --Tagishsimon (talk) 00:21, 15 December 2020 (UTC)

How do I change the rank of a qualifier? How do I assign references to particular qualifiers?

For example, see Lucian Freud (Q154594) and Lucien Freud's marriage to Kathleen Godley. I would want to change the full date of marriage to be a preferred rank, and add the ODNB reference to it; add the ODNB reference to the number of children; add the Who's Who reference to the end date etc. Piecesofuk (talk) 09:09, 8 December 2020 (UTC)

@Piecesofuk: For the first part, it's unfortunately not possible - qualifiers don't have rank independent of the parent statement. In this case, I'd recommend only using the more precise qualifier and dropping the more vague one, since it's mostly redundant.

For the second, you could perhaps use applies to part (P518) on the reference? It's not very frequently used on references, though there are a thousand or so examples. I am sure I remember seeing a property proposal for some kind of "reference applies to specific qualifier(s)" property a year or so back, but I have completely failed to track down what happened to it. Andrew Gray (talk) 13:52, 8 December 2020 (UTC)

@Andrew Gray: Wikidata:Property proposal/Supports qualifier perhaps? A pity, because somebody was asking me this weekend how to deal with just this use-case. (I said you were a good person to discuss such things with). Jheald (talk) 17:20, 8 December 2020 (UTC)

@Jheald: That's the exact one, though for some reason I thought you were the original proposer :-). It looks like the proposal mostly failed on a "not quite clear what's being suggested" basis, so I wonder if it's worth resurrecting it with some more detailed practical examples (this one would be a good one, in fact - different sources for start, end, etc).

The item as currently shown has an end date of both 1957 and 1958, which points out the other problem with qualifiers - we can't chain qualifiers onto other qualifiers, so there's no way to eg say "circa" but only have it apply to end time, not start time. Andrew Gray (talk) 21:55, 8 December 2020 (UTC)

Maybe with earliest end date (P8554) and latest date (P1326)? Ghouston (talk) 03:04, 9 December 2020 (UTC)

• Been looking into this a bit more and there seem to be more cases of this than I thought - I make it 1813 items which have a "spouse" claim with two start dates, 1552 of which have both dates in the same year, and 1062 of which have the second year be "year" precision. (Edit: another 47 which have the second date at month precision, and both dates in the same month.) I think we could probably have a bot remove all those last thousand qualifiers (two claims, same year, one claim is year-precision only). Interestingly "spouse" seems to be a particular problem for this - the same query on P39 claims only finds 127 examples. Andrew Gray (talk) 08:37, 15 December 2020 (UTC)

Merge IDs Távora (Tabuaço)

Hello, I tried to merge the IDs Q33869513 and Q111920 as they both relate to the same parish (Távora in the municipality of Tabuaço) and need to be merged. But it seems that there is a problem as they just can´t be merged. I am not very familiar with merging and the indications didn´t help me here. Maybe someone more experienced here could fix this problem, that would be wonderful! Thank you very much! Best wishes,--Joehawkins (talk) 16:55, 15 December 2020 (UTC)

One is a former civil parish; the other a locality within the parish. They are distinct things, one nested within the other. They should not be merged. --Tagishsimon (talk) 16:59, 15 December 2020 (UTC)

Question from a new contributor

Hey, I'm fairly new to Wikidata. I've contributed pretty heavily to Wikipedia, moderately to Wiktionary, Wikispecies, and Wikivoyage, and a little bit to most of the other wikis.

I recently edited the entry for Q4917275 (Biscuits and gravy), and in addition to adding a Commons category I recently created for the subject, I also noticed there was no Wiktionary entry listed on its WD page. I tried to link the corresponding article from the English Wiktionary, but an error message came up saying something about interlanguage links.

I couldn't make heads or tails of it, so I clicked on WD:N in the message, and the policy reads: "On Wiktionary, items for citation pages are not allowed. Main namespace is also excluded because interlanguage links are automatically provided by Cognate." I went to a Wikidata page for a random noun (Q33487 (homophobia)) just to see if it had any Wiktionary links, and there weren't any there either. Can anyone help me make sense of this in a familiar-with-Wikimedia-but-not-with-Wikidata-specifically way?

Also, upon trying to publish the edit again to get the error message to paste here, the edit went through, since I guess Wikidata remembered that I'd seen the message. I can revert it if need be. TheTechnician27 (talk) 00:11, 16 December 2020 (UTC)

@TheTechnician27: Interwiki links between Wiktionary mainspace pages are handled by the Cognate extension, and adding a Wiktionary sitelink to an item here removes all sitelinks which the extension would otherwise have provided (consider a given word with definitions in multiple Wiktionaries). Also, even absent the extension, adding Wiktionary mainspace links to Wikidata items is not appropriate because such links potentially describe multiple lexemes in multiple languages (the page "mesa", for example, describes a landform in English, a table in Spanish, and so many other things). It is thus better in each case to revert your addition of the Wiktionary sitelinks. Mahir256 (talk) 00:21, 16 December 2020 (UTC)

Wikidata:Wiktionary/Sitelinks explains the current situation in regards to Wiktionary sitelinks in detail "The links to pages of the main namespace should not be stored in Wikidata. A link to a Wiktionary entry should not be linked on a Q-item on Wikidata, since Q-items are about concepts and not words. Therefore, the section "Wiktionary" on items pages should not contain links to a Wiktionary entry. " ChristianKl ❪✉❫ 01:46, 16 December 2020 (UTC)

Obalky knih.cz

Currently we have "described by source=Obalky knih.cz" for over 5,000 items, shouldn't this be made into an identifier instead? --RAN (talk) 22:00, 15 December 2020 (UTC)

It should. It makes about as much sense as described by Discogs (Q504063) or Virtual International Authority File (Q54919). However, described by source (P1343) is rather open-ended, rather prone to duplication with encyclopedic identifiers, (e.g. American National Biography ID (P4823)), and likely to attract as many sources as devoted users are willing to add. As is often the case on Wikidata, there are no rules, only endless, fruitless discussions, so people will just do whatever until someone disagrees. And in the absence of guidance, it's quite easy for someone to use a bot or script to make thousands of questionable decisions in a single swoop. -Animalparty (talk) 20:53, 16 December 2020 (UTC)

I will make it into a formal property proposal this weekend. Thanks for your input. --RAN (talk) 00:42, 17 December 2020 (UTC)

How to retreive data from external website geonames.org

Hello everybody, I realized that currently there're over 36.000 Italian human settlements in wikidata without located in the administrative territorial entity (P131) property but, at the same time, connected with the external website geonames.org which owns this information as 3rd level administrative entity. e.g. Q1942504 Is there any chance to import from this website data to wikidata? Thanks in advance.

• Items used: human settlement (Q486972)   , Italy (Q38)   

• Properties used: instance of (P31)   , country (P17)   , GeoNames ID (P1566)   , located in the administrative territorial entity (P131)   

SELECT ?item ?itemLabel ?itemDescription ?geonames WHERE {
?item wdt:P31 wd:Q486972.
  ?item wdt:P17 wd:Q38.
  ?item wdt:P1566 ?geonames
  MINUS {?item wdt:P131 []}
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}

Try it! Thank you in advance. Luckyz (talk) 10:40, 16 December 2020 (UTC)

• Geonames data is not CC0. In many cases there are also quality problems with geonames data that speak against just importing them all without being specific about what subset is to be imported and analysing the quality of that subset beforehand. ChristianKl ❪✉❫ 11:42, 16 December 2020 (UTC)

Thank you very much for your answer. Luckyz (talk) 07:01, 17 December 2020 (UTC)

Lookup by badge

Does anyone know whether PetScan supports filtering results by badge? If it can't be done via PetScan, is there another way to query this information? I am trying to find out what pages on English Wikisource have incorrect badges on Wikidata - for example, pages on English Wikisource in the category s:en:Category:Validated texts that are missing the validated (Q20748093) badge in the enws sitelink. Beleg Tâl (talk) 13:26, 16 December 2020 (UTC)

You can include some SPARQL in the Petscan (in the Other Sources) tab, to check for a badge. The SPARQL required might be

select ?item where {
?article schema:about ?item ;
          schema:isPartOf <https://en.wikisource.org/> .
filter not exists {?article wikibase:badge wd:Q20748093.}
}

Try it!

but I'm getting a connection error from Petscan right now, with and without this SPARQL, so I cannot test the above. --Tagishsimon (talk) 17:40, 16 December 2020 (UTC)

Petscan (or in fact toolforge) fixed. So. Petscan finding all articles with badge, all articles without badge (zero, as it happens) and, as a further proof of concept, articles that have P31=Q10872094, to demonstrate that the SPARQL works (given that all/zero results are always a little unsettling). Petscan settings for these are the obvious ones in the Category tab; some SPARQL in the Other sources tab; and Only pages with items checked in the Wikidata tab. hth --Tagishsimon (talk) 04:57, 17 December 2020 (UTC)

Removing "Other places to connect" from the header of the project chat

I believe having "Other places to connect" in the header of this page is undesireable. Consensus-finding is very important when it comes to Wikidata and questions of how certain issues should be solved. When questions about how to solve an issue are raised on telegram instead of on-wiki it results in the conversation not being able to be revisit when similar issues happen in the future. While I don't believe in ending other channels to discuss Wikidata, I see no reason to promote them on the top of this page. ChristianKl ❪✉❫ 16:11, 16 December 2020 (UTC)

It has always puzzled me that Telegram, which I've only ever heard of in the context of its use by nazis, is a tool of choice for wikimedians. --Tagishsimon (talk) 17:32, 16 December 2020 (UTC)

Telegram is a service with 400 million monthly active users. Just because someone finds it notable that some whie supremacists used it doesn't mean much for average usage. ChristianKl ❪✉❫ 21:30, 16 December 2020 (UTC)

And how does removing the link solve this? The Telegram group currently has over 600 members so it seems to serve a purpose. It's de facto the primary way to directly communicate to other community members. Removing the link and making it harder to find seems like a step in the wrong direction. Multichill (talk) 23:34, 16 December 2020 (UTC)

I do think that the Project Chat is the primary way to directly communicate to other community members and that people who visit this page should see it that way. ChristianKl ❪✉❫ 23:45, 16 December 2020 (UTC)

+1 to this. --Yair rand (talk) 07:18, 17 December 2020 (UTC)

+1 --Jneubert (talk) 13:15, 17 December 2020 (UTC)

Wikidata:List of 1000 articles every Wikipedia should have

I'm trying to fix the page Wikidata:List of 1000 articles every Wikipedia should have. I have changed the query and, when I try to regenerate it, Listeria fails with: "Killed by OS for overloading memory". The query runs with no problems in less than one second on the queryService [6]

Another question: I have noticed that the query returns more than 1000 items, because it includes items that have been removed from the list, for example Sorghum (Q12111) which has been replaced with Sorghum bicolor (Q332062) on 2020-09-03. Is it OK if, when I find a case like this, I add the end time (P582) qualifier? Tcp-ip (talk) 11:29, 17 December 2020 (UTC)

• @Tcp-ip: In general it's okay. However it's better to have the discussion about what to do with items when they are removed from a focus list should happen on the talk page of on focus list of Wikimedia project (P5008). That way in the future people can look at the discussion and if they want to know how we handle those items. ChristianKl ❪✉❫ 13:40, 17 December 2020 (UTC)

Quickstatements problem

Hi, can someone tell me why this quickstatements code is not working for this page?

CREATE

LAST P31 Q17928402

LAST Len "The story of mRNA: From a loose idea to a tool that may help curb Covid"

LAST Den "blog post published in 2020"

LAST P953 "https://www.statnews.com/2020/11/10/the-story-of-mrna-how-a-once-dismissed-idea-became-a-leading-technology-in-the-covid-vaccine-race/"

LAST P577 +2020-11-10T00:00:00Z/11

LAST P1476 en-us:"The story of mRNA: From a loose idea to a tool that may help curb Covid"

LAST P407 Q1860

Tetizeraz (talk) 15:40, 17 December 2020 (UTC)

I don't think wikidata does en-us. The story of mRNA: From a loose idea to a tool that may help curb Covid (Q104223724) --Tagishsimon (talk) 17:18, 17 December 2020 (UTC)

Yes, the issue has a ticket at https://phabricator.wikimedia.org/T154589 ChristianKl ❪✉❫ 17:47, 17 December 2020 (UTC)

Always better to haggle for 3+ years and get nothing done, than just create en-us. Sclerotic. --Tagishsimon (talk) 18:06, 17 December 2020 (UTC)

Thanks! Tetizeraz (talk) 00:48, 18 December 2020 (UTC)

Two different William Clark (artists)

We only have one painting by William Clark (Q21464475) who died in 1801

en:William Clark (artist) (Q21464479) has several paintings at commons:Category:William Clark (painter). All of the paintings there are assigned to creator: William Clark (painter). However they show in Commons images as the earlier William Clark.

Needs sorting out perhaps, either that, there is a very long time lapse...

File:William Clark (1803-1883) - The Steamship 'Earl Percy' Going into Dumbarton - BHC3303 - Royal Museums Greenwich.jpg Is the later William Clark persists to show the earlier man. All the images in the cat are the younger man.

Broichmore (talk) 17:03, 18 December 2020 (UTC)

@Broichmore: thanks for pointing this out. I untangled William Clark (Q21464475): British artist and soldier, died 1801 and William Clark (Q21464479): Scottish painter (1803-1883). Multichill (talk) 21:10, 18 December 2020 (UTC)

Tall skinny reference

Tracked in Phabricator
Task T259183

I added a reference to (Q34857540) and now see it in Opera under Windows as a column of single characters. Is that how it should be? It isn't easy to read. Jim.henderson (talk) 18:23, 18 December 2020 (UTC)

This looks like Phabricator T259183, which seems to have disappeared in Chrome in September unless you have a very narrow screen width. If you're using the default Vector skin, try collapsing the left "tools" column. - PKM (talk) 22:27, 18 December 2020 (UTC)

Help tease apart conflation of identifiers

At Giovanni Anastasi (Q5563656) --RAN (talk) 07:34, 20 December 2020 (UTC)

•  Done Bad merge reversed  – The preceding unsigned comment was added by Richard Arthur Norton (1958- ) (talk • contribs).

I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Matěj Suchánek (talk) 11:59, 24 December 2020 (UTC)

Merge

Hello, can we get Aqrah District (Q65401356) and Aqrah District (Q4782556) merged? Same subject. --Semsûrî (talk) 11:29, 23 December 2020 (UTC)

Done. See Help:Merge#Gadget. --Tagishsimon (talk) 11:38, 23 December 2020 (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. Matěj Suchánek (talk) 11:57, 24 December 2020 (UTC)

Help using Wikidata for US Public Laws?

Hi, Wikidata folks. I'd like some help figuring out how to use and leverage Wikidata for managing the use of US Public Laws on enwiki. For instance, see the thread I started at w:Talk:List of acts of the 116th United States Congress#How_can_we_use/levarage_wikidata?. There are a series of list articles on enwiki that could be generated from a database, and some (but not most) of the public law data is already in Wikidata. And it's also easy to imagine improvements to templates like {{uspl}} if they can look up the information in Wikidata. What's the right way to get started and get some help on this, to engineer the schema (or whatever) or get a sanity check on the basic idea, and learn about bulk import tools for Wikidata? Thanks for any and all assistance! Jhawkinson (talk) 02:34, 17 December 2020 (UTC)

@Jhawkinson: This certainly seems like something suitable for import into Wikidata. @SJK: did some work on this previously here. There was a relevant Wikiproject at Wikidata:WikiProject US legislation but it seems to have become inactive; maybe it could be revived? Wikiprojects are usually where the schema (list of properties and qualifiers etc. to use) is hashed out for a given topic area. Once you know what you want the relevant items to look like (and it seems there are at least a handful of existing ones you could copy) then the next step is to put them into a spreadsheet of some sort and set it up to import here via Quickstatements. ArthurPSmith (talk) 13:34, 17 December 2020 (UTC)

When it comes to reactivting Wikidata:WikiProject US legislation it's worth noting that if you want to put in the effort to create a schema, you can be bold and just do it. ChristianKl ❪✉❫ 14:17, 17 December 2020 (UTC)

I don't think I know enough about Wikidata to be bold and tackle that, at least not without some guidance or maybe even handholding. My goal is to have the schema such that no more is required to get this going than what is available as soon as a public law is assigned. That's this information, that comes in the announcement email:

December 15, 2020
The Office of the Federal Register has assigned a law number to the following bill approved by the President:
 
S. 4054 / Public Law 116-216
United States Grain Standards Reauthorization Act of 2020
(Dec. 11, 2020; 134 Stat. 1048; 4 pages)

Perhaps we are there already? But one of the problems is the existing keys wdt:P3825 and wdt:P3837 seem to be keyed off the title of a public law, as when it is a Wikipedia article. But often there's no short title, only things like "To designate the airport traffic control tower located at Piedmont Triad International Airport in Greensboro, North Carolina, as the "Senator Kay Hagan Airport Traffic Control Tower". (That's P.L. 116–193). So I'm not sure what to do about that.

And also how does one get the Wikipedia article to be generated as the result of the query? It looks like there are bots for that (ListeriaBot?) but I'm not sure if that's the right way to go? Thanks. Jhawkinson (talk) 21:22, 17 December 2020 (UTC)

Currently, templates by themselves can't make SPARQL-queries and only bots can do so. ListeriaBot reads the {{Wikidata list|}} template to create a list based on SPARQL.

Without having domain expertise in US law myself, title (P1476), short name (P1813) and subtitle (P1680) seem to me like suitable properties to store title information. Given that officially the long this is called title we might use short name (P1813) for the short title. Ideally you want to draw the information that you display in Wikipedia from those properties and not the label. If no short title is available it's okay to use the longer title as the label on Wikidata.

I filled out Government Employee Fair Treatment Act of 2019 (Q62774908) a bit. I would expect that the information about the sponsor of the law to be important. There could be list pages about the laws sponsored by senators/congressman from a given state. The fact that it's easy to generate additional lists might be valuable for Wikipedia.

If you know what information from your query should display https://www.wikidata.org/wiki/Wikidata:Request_a_query is a place to get help on Wikidata. ChristianKl ❪✉❫ 22:40, 17 December 2020 (UTC)

Thanks again, @ChristianKl: Is there some sort of guidance or understanding or rule about how much partial data is acceptable? And is there some place that the schema is formally laid out and defined for a thing (like a Public Law)? I agree a sponsor of a law is important to a lot of people, but it's not directly available from the source that new public law data is available, and I think it may be a complex multi-valued field. Is it a problem if new laws are added without filling that out?

I was under the impression I was pretty well set on the SPARQL query (there's one at the above-referenced talk page, w:Talk:List of acts of the 116th United States Congress), but I just added a call to Listeriabot told it to run and it came back with zero results, despite its ISPARQL button giving me a working query. Is this a Listeriabot bug or limitation and where do I look to solve that issue? Thanks. Sorry if I'm being too noisy here, I'm not familiar with this wiki's culture and norms. Jhawkinson (talk) 23:28, 17 December 2020 (UTC)

Partial data is completely fine for Wikidata. Having the ID of the law that uniquely identifies the law would be sufficient to justify an item for the law. As far as violating norms you are much more likely to get into conflicts with Wikipedians over using Wikidata over there. Having a way where you can show that using Wikidata provides benefits over the status quo might be important to win upcoming fights.

I played a bit around with it a bit and it seems like the "|links=local" parameter was wrong and blocked Listeria from accepting the template. Unfortunately, I don't have much experience with Listeria. Whenever I need something more complex with Listeria I usually go to the Request a Query page. ChristianKl ❪✉❫ 01:06, 18 December 2020 (UTC)

Thanks again, that was very helpful. Listeriabot seems pretty fragile, I guess I should report the "links=local" issue as a bug. Similarly, for the version of the query on the talk page now, it returns a different order than the ISPARQL link, and it formats the dates differently. Ugh. Also, I need to think some more about the schema, but I didn't see an answer to my question about whether there is someplace that it should be written down or documented? As you can see, there's some variation in how these laws are have been input for the 116th Congress. Jhawkinson (talk) 07:05, 18 December 2020 (UTC)

The main idea of how schema's should be written down is in a Wikiproject. In this case Wikidata:WikiProject US legislation. There are a few other was to also leave documentation to be more accessible. Act of Congress in the United States (Q476068) model item (P5869) Government Employee Fair Treatment Act of 2019 (Q62774908) for example helps people find the item as an example for the modeling of legislation. ChristianKl ❪✉❫ 13:15, 18 December 2020 (UTC)

I would expect that local behaves here like intended given that none of the lines have local links. That also seems like a key problem. The existing list has links both to Wikipedia articles describing the law and Wikisource pages giving the text of the law. From looking at the template myself it's not clear whether there's currently a way to get all the links. You likely need to talk to someone with more experience with the template then me to get forward on that list working.

https://github.com/magnusmanske/listeria_rs/issues is the issue tracker for Listeria bugs. ChristianKl ❪✉❫ 13:15, 18 December 2020 (UTC)

Schemas: Can you point me at a good example of a Wikiproject that documents a schema well so I can know what to look to? When I look at Act of Congress in the United States (Q476068) that seems helpful, although I'm surprised there's not a link to "Instances of Acts of Congress", i.e.

SELECT ?item ?itemLabel ?itemDescription
WHERE {
  ?item wdt:P31 wd:Q476068  .
      SERVICE wikibase:label { bd:serviceParam wikibase:language "en" }
} 
LIMIT 10000

Try it!

I mean obviously it's a simple query, but it would seem like that sort of thing would help new Wikidata users understand how things work? And is there somehow supposed to be a link from that qualifier to the relevant Wikiproject? Otherwise how are editors supposed to know which wikiproject to look at to understand schema?

Similarly, it's great that the Q476068 links to Government Employee Fair Treatment Act of 2019 (Q62774908) as a "model instance," but when you're viewing and editing Q62774908 there's no indication that it is designated as "model"? I would think we'd want that, so editors would be extra careful when they are editing a model item in recognition of its status?

Listeriabot: I see…I think I misunderstood what "links-local" really meant and assumed it meant to look for Wikipedia articles whose name matched the column output, but I guess actually there's a concept of a link in Wikidata that would be explicitly encoded, which is not the case here. (To me, this means Listeriabot could use a verbose or a debug mode or something, but that's more like a feature request than a bug.) Looks like {{Wikidata list}} has a "row_template" so we can use any sort of fancy template substitution, which may be the way to go. Although maybe there should still be links in the Wikidata? I'm too inexperienced to say. [help?]

Wikisource links: I'm not really sure what's up with the alleged Wikisource links. Since they're interwiki links, it's not so obvious when they are redlinked, but it appears that most of them don't exist (e.g. s:Public Law 116-2), and I don't know who is going through the effort to load the text of laws into Wikisource when there are other reliable sources of this information that are linked in the final column of the table (empirically, it seems like few people are). Obviously this is not the place to talk about them, though.

Overall: Thank you again for your responsiveness and assistance, and I hope I'm not taking up too many resources with both specific questions on how to accomplish my task and also more general questions about how Wikidata is structured and documented that I hope might benefit other new people the first time they come to Wikidata! Jhawkinson (talk) 14:52, 18 December 2020 (UTC)

I'm happy to answer productive questions that help people to use Wikidata and that's what this Project Chat is for.

I just saw that the Wikisource text only exists for the first item in the list and not for the following. Given that it might not be as important. In case someone does a bulk import of the texts on Wikisource that person should link those Wikisource documents with the Wikidata item. ChristianKl ❪✉❫ 15:38, 18 December 2020 (UTC)

We do have maintained by WikiProject (P6104) to tell users which Wikiproject has the documentation. https://www.wikidata.org/wiki/Wikidata:WikiProject_Books is a good example of a lot of schema work. ChristianKl ❪✉❫ 15:41, 18 December 2020 (UTC)

(I have no idea how far to indent this) The Stumbling Stone Entity Schema E206 might be a useful example. I seem to remember seeing documentation for the schema specifically fairly recently, but can't find it right now. The Structure section on the project page might be of interest, and the list of useful queries could also be worthy of imitation. (Other parts of that page are in German, because the project in question is, but they are mostly specific to that project) --Matthias Winkelmann (talk) 02:32, 20 December 2020 (UTC)

Adding a numeric ID as a qualifier on existing instances of a property

So, Internet Game Database game ID (P5794) uses slugs for its IDs, however this is a problem because games can change names in development (e.g. https://www.igdb.com/games/the-elder-scrolls-vi will likely change to be "the-elder-scrolls-vi-foobar" or whatever, once Bethesda actually names the game sometime in the next few years) or just have their slugs changed for other reasons.

Unfortunately, they don't support linking to game pages based on their numeric IDs. So the best solution I can come up with for maintaining data integrity is to write a bot that'll automatically add a 'numeric identifier' qualifier to every IGDB ID statement, based on the data I can get from their API. I swear I've seen this on some properties before but I can't remember where. Does anyone have an example of this in Wikidata somewhere, or other guidance for how it should be accomplished?

Thanks! Nicereddy (talk) 16:34, 19 December 2020 (UTC)

@Nicereddy: we do this for twitter. see X username (P2002) and X numeric user ID (P6552). Does IGDB at least maintain redirects so we can fix the slugs without using numeric IDs? BrokenSegue (talk) 15:05, 20 December 2020 (UTC)

@BrokenSegue: They don't maintain redirects when fixing duplicates, so I don't think they do it for renames either. Thanks for the link to how Twitter does it, I guess I'll have to propose a numeric ID for IGDB as well. Nicereddy (talk) 18:08, 20 December 2020 (UTC)

Deletion of Wikidata objects and effect on structured data on Commons (SDC)

Hello, if an Wikidata object is deleted, what is the effect on structured data on Commons (SDC)?

• Will any Wikidata object only be deleted if it is not referenced in structured data on Commons?
• Or will it be deleted anyway and the reference in Commons will become invalid?
• Or will the references in Commons be deleted as well? If so, by whom and when? (automatically by a bot, manually by the user, who deletes the object, ...)?

Thanks a lot! --M2k~dewiki (talk) 16:35, 20 December 2020 (UTC)

• If an item on Wikidata gets deleted statements that link to it whether they are on Wikidata or Commons don't get automatically deleted. Those statements will just link to the QXXXX number (and that gets shown instead of the previous label).

Currently, our rules don't give items automatic notability for being used in Commons. Otherwise you would have circular notability. A person could add an image of themselves and then use depicts to point to a Wikidata item and image (P18) from the Wikidata item to the Commons item and both would make each other notable.

Apart from that there's no way an admin who looks at an item will notice that it gets used in Commons. Neither the inverse statement gadget nor What links here give information about how a Wikidata item gets used on Commons. ChristianKl ❪✉❫ 18:19, 20 December 2020 (UTC)

Create an object for each article on arXiv?

There are many objects on Wikidata that correspond to scholarly articles, and they can be used for citing these articles in Wikipedia using w:Template:Cite_Q. It would be good if many more articles were available. Would it be possible to import all articles that are on w:arXiv? Or at least all arXiv entries that correspond to published articles? Sylvain Ribault (talk) 20:26, 19 December 2020 (UTC)

maybe it'd be better to have an "on demand" import function? import them just in time when a ref is wanted? BrokenSegue (talk) 22:21, 19 December 2020 (UTC)

I'm working on something that creates an item for a given URL, including author, title, publication, etc. I believe this is the single most important missing feature currently, at least if you consider missing references among the top issues from the data consumers' POV.

Importing all of arXiv does not strike me as the best use of someone's time and/or limited ressources. The academic papers we already have are rarely connected to items outside the narrow niche of academic publishing, i. e. author / journal / institution. I may very well be mistaken and there might be some subject matters where they are put to good use, molecular biology being my prime suspect here. But for history, politics, arts, culture, sports, the economy, and geograph, I would suspect (news) websites and books being the top two types of references, followed by a large gap.

Here's a query that almost, but not quite, speaks to that question: Query. If anyone knows a way to sample a random subset of items, please let me know. As it is, academic papers are entirely irrelevant as reference sources for the subset the query captures hefore timing out. You can vary the P31 line to explore a bit, but actually useful references are somewhat swamped by artifacts such as Facebook usernames being referenced to Facebook etc.

Books, (Website) Publishers, and journalists are far more promising if anyone wants to create items that may potentially be sources. I've sometimes created five items for a single reference, and have yet to find a therapist who could explain why I would do such a thing. --Matthias Winkelmann (talk) 01:32, 20 December 2020 (UTC)

Generally agree with everything you say. Is the tool you are working on open source? Can you link it? I use simple string filtering on QIDs to get "random" samples (first N QIDs ending in "001"). BrokenSegue (talk) 01:38, 20 December 2020 (UTC)

It will be open source. I initially started with books, where it's somewhta complicated to avoid duplication / find exisitng publishers and authors, and to deal with the variety of editions, translations, and their representation when they already have items.

Website should be easier, since almost everything has easily consumed metadata in the source code these days. This includes publication, author(s), title, date, language, summary, license, and sometimes subject, as well as a single canonical URL that can reasonably be assumed to be sufficient for avoiding duplicates. I'm fairly well along with that and the single largest todo is getting aquainted with the toolserver without shutting down some smaller country's internet. I should have something running there within the week and will post here.

An obvious next step after that would be a browser extension to do this, which would also allow creating the reference itself (first version will still require chosing "stated in" and the item that was created), as well as highlighting a sentence and having it automatically added as a quotation. Unfortunately, this would require either me overcoming my hatred of JavaScript, or JavaScript overcoming its nature of just being terrible, and neither are budging so far.--Matthias Winkelmann (talk) 02:18, 20 December 2020 (UTC)

@Sylvain Ribault: Check out Wikidata:WikiProject Source MetaData. -Animalparty (talk) 08:11, 20 December 2020 (UTC)

I have a plan to do so, but currently not a specific timeline.--GZWDer (talk) 08:30, 20 December 2020 (UTC)

Thanks to @Animalparty:'s link I found this chart which shows that arXiv is entirely irrelevant with regards to citations/references, at (eyeballing this) about 1 or 2 %. And that statistic does not even include weblinks. --Matthias Winkelmann (talk) 09:07, 20 December 2020 (UTC)

I think it would be a good idea to import them. I thought that someone was already doing it partially as we have 346.470 arXiv identifiers. arXiv is increasingly important in certain areas, e.g., machine learning. A common machine learning paper my site many arXiv (only) papers. One issue: Sometimes arXiv articles are associated with published articles and there might be a DOI listed. — Finn Årup Nielsen (fnielsen) (talk) 09:22, 20 December 2020 (UTC)

I have been an arXiv author for 25 years (first submission in 1996). In my field (physics), it is customary to publish a preprint there before submitting it to a peer-reviewed journal. These preprints are sometimes cited before publication, but they are largely forgotten after the journal publication (and definitely almost never updated in accordance with the published version). I do not see why we should import all of them here.--Ymblanter (talk) 12:45, 20 December 2020 (UTC)

@BrokenSegue: An on demand import function would indeed be almost as good, provided it checks that the import was not done before.

@Matthias Winkelmann Not so clear that arXiv is irrelevant, given that there is a large overlap between arXiv and DOI. A large fraction of arXiv entries are article that have DOIs, and the DOIs are part of arXiv metadata. (I suspect that articles that have both a DOI and an arXiv ID show up as DOI in your data.) By importing arXiv, you would get a lot of article with DOIs, and this might be an easier way to get them than via the journals. Sylvain Ribault (talk) 21:00, 20 December 2020 (UTC)

The question I have on this is the data model. Should preprints be a separate item from the published article, or united under a single item? I think I'd prefer a single item, but we seem to have both models present in Wikidata so far (I've seen up to 5 items for different preprint versions and the final published version of an article!) I don't think we need to go full FRBR on scholarly articles at least! ArthurPSmith (talk) 20:04, 21 December 2020 (UTC)

Very relevant question. I would prefer a single item too. An arXiv entry can already include several versions of an article. Adding a published version in the mix is no big deal. What matters is that the Wikidata item refers to all versions. In case there are several substantially different versions, and one wants to cite a specific version, it is always possible to add that information when citing. Sylvain Ribault (talk) 20:21, 21 December 2020 (UTC)

Localized links in user warning templates

I've asked a question here regarding using {{Q}}–links in the user warning templates. Since I doubt many people watch that page, I figured I would note it here. Perryprog (talk) 03:18, 21 December 2020 (UTC)

OmegaWiki

For about a week when I try to access the OmegaWiki site the only thing appearing is "500 Internal Server Error". Anyone knows what`s happening? I know it's an unrelated website, but since it is linked here from 2000+ pages, surely someone had noticed it.85.58.45.208 12:46, 21 December 2020 (UTC)

nope haven't noticed. BrokenSegue (talk) 18:36, 21 December 2020 (UTC)

Add Genastar IDs?

Should we be adding IDs from Geneastar? One big concern I have about this site is they offer little information on their home page about how they operate. For example, they don't mention if they have a goal of having a single world-wide family tree, where a given individual is mentioned only once, or if they host multiple independent family trees where a person will have as many IDs as there are people who have contributed a family tree that mentions the individual. Jc3s5h (talk) 13:08, 21 December 2020 (UTC)

We already had a catalog at [7]. GeneaStar is only a portal that provide information about people and links to Geneanet pages; it is itself not a family tree. Geneanet in turn is a collection of many user-submitted family tree with varying quality but not a reliable source as a whole (being user-generated); the most commonly used one (Tim Dowling's) have subpar quality (emphasis: this is about Geneanet itself, not the portal pages at GeneaStar).--GZWDer (talk) 15:33, 21 December 2020 (UTC)

Wikidata weekly summary #447

Here's your quick overview of what has been happening around Wikidata over the last week.

• Events
  • Past: OpenAlt 2020 Conference (replay)
    • Wikidata: how to get involved (in Czech)
    • Lexemes on Wikidata (in Czech)

• Press, articles, blog posts, videos
  • Video: HGAPS Speaker Series by Thomas Shafee
    • Wikidata Overview: how you can use it to improve the overall knowledge base of the Internet
    • Wikidata Walkthrough: hands-on editing experience
  • Wikidata for Bat Collectors, YouTube

• Tool of the week
  • Wikidata Live is a real-time visual dashboard for Wikidata edits; paper, interface (visualizes recent changes to Wikidata in quasi real time).

• Other Noteworthy Stuff
  • Complex constraint about labels' language

• Did you know?
  • Newest properties:
    • General datatypes: National Library at Kolkata romanization, URL match pattern, URL match replacement value
    • External identifiers: ClassInd audiovisual work ID, Georgia Museum of Art ID, Kunsthistorisches Museum Wien ID, tribunsdelaplebe.fr ID, PBA artwork ID, OpenReview.net profile ID, OpenReview.net group ID, OpenReview.net submission ID
  • New property proposals to review:
    • General datatypes: applicable "stated in" value, WMF hosted image that could be moved to Commons, Category for the view of the item, OSM object, property related to this topic, personality traits, graph girth, ALA-LC romanization
    • External identifiers: SAPA, Manus Online author ID, Mapy.cz ID, CollectieGelderland-creator-ID, SVTplay identifier, identifiant Mediafilm, IFLA form of music ID, IFLA key of musical work ID, IFLA mode of musical work ID, RDA support type ID, SLAM artwork ID, Kunstmuseum Basel ID, NPG ID, jeuxvideo.com ID, Dizionario bio-bibliografico dei bibliotecari italiani del XX secolo ID, Google Podcasts show ID, Royal Museums Greenwich ID, PIAC ID, Castbox show ID, Stitcher show ID, Pocket Casts show ID, Podcast Addict show ID, Radio.com Podcasts show ID
  • Deleted properties: street address (P969)
  • Query examples:
    • Tex-Mex singers (Source)
    • Map of images from Wicipics project (photography campaign in Wales) (Source)
    • Vito Russo Test in Wikidata
    • Distinct dialects of Wikidata lexemes (Source)
    • Patronage of rapid-transit transportation (Source)
    • People who played Julius Caesar, and what "version" of him they played
  • Newest WikiProjects: Media Representation
  • Newest database reports: Reports on Italian given names

• Development
  • Continuing to work on the query builder. It can now have more than one query conditions. One of the next steps is making it possible to query for Item values. You can follow along on the demo system.
  • Finished working on the problem of values of Statements that link to Forms and Senses not having language attributes associated with them in the HTML code (phab:T267023)
  • Fixed language selectors covering other input fields when using keyboard navigation for Special:NewItem and Special:NewLexeme (phab:T266638)
  • Fixed language fallback indicators sometimes still shown for variant fallbacks (phab:T267502)
  • Working on whitespace stripped while typing when editing lexemes (phab:T250550)

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

• Monthly Tasks
  • Help add property scope constraint (Q53869507) to properties that currently don't have it.
  • Add labels, in your own language(s), for the new properties listed above.
  • Comment on property proposals: all open proposals
  • Contribute to a Showcase item.
  • Help translate or proofread the interface and documentation pages, in your own language!
  • Help merge identical items across Wikimedia projects.
  • Help untangle items that refer to different concepts.
  • Help write the next summary!

Read the full report · Unsubscribe · Mohammed Sadat (WMDE) 16:49, 21 December 2020 (UTC)

Alemannic language problem for labels, descriptions and titles

Hi! I do not know where I can ask this question, so I'm writing it here.

I recently tried to use the user interface in Alemannic (Q131339) language. I speak Alsatian (Q8786) in its Low Alemannic German (Q503724) form, which is a dialect of the Alemannic group. Swiss German (Q387066) is another dialect of the Alemannic group.

As I created item Q104401427 using Special:NewItem, I chose Alemannisch (als) as language for label, description and alias. However, once the item was created, the label, description and alias appeared in the unnamed als language. There is another line for labels, descriptions and aliases in Swiss German (gsw).

The problem is: when I use the Alemannisch user interface, only Swiss German labels appear, so I cannot see the als-language labels. Furthermore, Alemannic is not available in the language selection of title (P1476), so I have to use Swiss German for titles that are in Alsatian, which is a problem.

I guess we have a big mess to solve here. I think all Alemannic dialects should be grouped as one Alemannisch language label once and for all, since there is a single global Alemannic Wikipedia (Q1211233). Any idea?

—Mathieu Kappler (talk) 08:06, 21 December 2020 (UTC)

See Wikidata:Project chat/Archive/2020/11#Some language names is not translated to other languages.--Snaevar (talk) 23:31, 21 December 2020 (UTC)

Canadian English labels

Why often Canadian English labels are capitalised (hundrests of cases) while English and British English are not capitalised? Example (in this case all should be capitalised). Eurohunter (talk) 16:23, 21 December 2020 (UTC)

@Eurohunter: why do you ask here instead of on the talk page of the the bot operator? Multichill (talk) 21:28, 21 December 2020 (UTC)

I think this is unrelated to the bot. ¿ The diff is about Portuguese ("pt") ? --- Jura 21:46, 21 December 2020 (UTC)

@Jura1: Link fixed. Eurohunter (talk) 23:08, 21 December 2020 (UTC)

Age at event

I have noticed that some death dates have a qualifier "age at event=" where the calculated age is added. Is there any plan to add this globally via bot? It would help when looking for errors at a quick glance, when matching against the age in an obituary with having to do the math in my head. We can restrict it to where we know the exact date of birth and the exact date of death. --RAN (talk) 07:45, 20 December 2020 (UTC)

I believe we're generally allergic to such data duplication. It needlessly consumes ressources and tends to go out of sync. It's the sort of flourish that UI clients should add, and somewhat unfortunate that the interface here has very little in that regard. This is a query that includes age, and you might be able to integrate it into your workflow. There is also a remarkably elegant solution that involves about two lines of JS in commons.js, and maybe someone feels up for the challenge of finding out what those two lines are. --Matthias Winkelmann (talk) 08:47, 20 December 2020 (UTC)

In cases where we know both the exact date of birth there's no reason to use age of subject at event (P3629). The property exists for those cases where we don't have the data to calculate the age but a source gives us the age. ChristianKl ❪✉❫ 14:01, 20 December 2020 (UTC)

If it's merely calculated from the known birth date, it should be removed. --Yair rand (talk) 04:27, 22 December 2020 (UTC)

Quotation marks surround objects from museum

See Portrait of Count Vladimir Bobrinsky, Lieutenant of the Life-Guards Hussar Regiment (Q21725664) as an example where the import of objects from a museum have quote marks, are they needed or should they be removed? --RAN (talk) 02:49, 22 December 2020 (UTC)

Remove. They serve no purpose except to annoy. --Tagishsimon (talk) 02:59, 22 December 2020 (UTC)

•  Done for that one example, but many more ... --RAN (talk) 03:12, 22 December 2020 (UTC)

Wikidata:Requests for permissions/CheckUser/1997kB

As required by policy, this is a notification regarding my request for CheckUser permission. ‐‐1997kB (talk) 06:47, 22 December 2020 (UTC)

Why aren't required qualifier suggestions visible at statement claim instantiation?

I recently added a Twitter handle to Wikidata for a Wikipedia article I was already editing, after noticing the hidden category that implied it was missing. I was surprised that the user interface has a concept of a "required qualifier constraint," but it only tells me about the required constraint after I publish the claim that requires it. This seems confusing both because it doesn't seem to actually be required (after all, it let me publish without it), and also because it would seem like you'd want editors to know about requirements while they were editing before they publish.

I just ran through this again with Bureau of Oceans and International Environmental and Scientific Affairs (Q4998406), which I picked out of a hat. Here's the initial publish screen:

And here's how it looks after I've published:

Is there a reason the warning doesn't come earlier, and am I misunderstanding what "required" means?

Thanks! Jhawkinson (talk) 18:55, 20 December 2020 (UTC)

Also if I'm right there is no instruction how to obtain Twitter numeric ID. Eurohunter (talk) 19:18, 20 December 2020 (UTC)

@Eurohunter: you start with "also," did you mean to say more? There are bad instructions on how to get the numeric ID in 3 places:

• Property_talk:P6552#Finding_the_user_ID_on_Twitter's_new_redesign
• Property_talk:P6552#Documentation_2
• Help:How_to_get_of_Twitter_user_numeric_ID

Is there a standardized place for that sort of thing? It sure would be nice if the Help popped up right when you're adding the qualifier. Jhawkinson (talk)

That is all what came into my mind when I noticed this thread. I just noticed I even have Help:How to get a Twitter user numeric ID in my watchlist but it would be better to see special property for instruction with link to Help:How to get a Twitter user numeric ID instaead Wikidata usage instructions (P2559) with "do not include the @-username". Eurohunter (talk) 19:33, 20 December 2020 (UTC)

Required qualifiers aren't actually required in a meaningful sense. Twitter numeric ID can be filled in later by a bot if you want to be lazy. Personally I wouldn't worry. BrokenSegue (talk) 20:22, 20 December 2020 (UTC)

This is a perfect opportunity to be "lazy". Let bots do the most tedious aspects of the grunt work. Allow humans more freedom to make more intelligent and abstract edits requiring human brains. -Animalparty (talk) 07:11, 21 December 2020 (UTC)

@BrokenSegue: @Animalparty: Could be there information that it will be updated by bot? Eurohunter (talk) 11:43, 21 December 2020 (UTC)

Coming back, the responses here don't really address the question, which was trying to be higher-level and using the above experience as an example. Is there a different place I should be raising questions about the Wikidata website's user interface? Let me enumerate:

1. It seems like the website should prompt for "required" qualifiers when a statement that requires them is instantiated. Why doesn't that happen?
2. If "required" qualifiers aren't actually required, why aren't they termed "recommended" or "suggested" or something like that?
3. Is there a way to make the instructions on how to ascertain a qualifier pop up (or be linked) at the point of need, i.e. when entering the qualifier? It seems like X numeric user ID (P6552) could really use that.
4. Is there a standardized location for help on a property? Is it just that P6552 isn't coded/linked properly to its help text?
5. Are these qualifiers in fact added by bots, as BrokenSegue and Animalparty suggest? (I'm not sure what "can be filled in later by a bot" really means) If so, shouldn't the documentation for the properties make that clear? Indeed, it seems like if we're going to prompt users about "required" properties (that are really "suggested"), there should be a similar classification for "required-but-a-bot-will-fill-it-in-anyhow" so the user knows how to distinguish the strength of the suggestion.

I guess I could source-dive the Wikidata mediawiki configuration and modules to get some of this information, but that seems like an awful lot of effort for what ultimately turn out to be policy questions rather than implementation ones? I think? Thank you! Jhawkinson (talk) 13:34, 22 December 2020 (UTC)

Some answers

1. because there is no 'should' about it. The UI could do something different. I've seen no demand for the difference you suggest in all the years I've been here. The devs have a huge backlog of other stuff to do.
2. one person's 'required' is another person's 'recommended' is another person's 'meh'. It's a highly subjective area: wikipedia does not break for lack of qualifiers, it's merely that some property statements are relatively suboptimal in the absence of some qualifiers. The 'Required qualifier' constraint is simply the constraint which nags you to add qualifiers. (There are many constraint types - Help:Property constraints portal). Having two such constraints, one for required, the other for recommended, would be feature creep and the locus of much wasted time deciding whether a qualifier is required or recommended. Sure, we could change the label, but, again, there has not been any demand that I've seen.
3. Could be done. Unlikely to be done b/c see #1.
4. No standard location. Experienced users will look at the property item, its talk page and perhaps the propoerty proposal. The talk page would be the most sensible place for usage information to be located. It's a volunteer effort. No-one has so far volunteered, that I know. It's possible to conceive of a gadget providing a link between the property selected for a statement and its help page, but no-one has written one, as far as I know.
5. It would be good to know if I can forget about a qualifier because a bot will take care of it. But a) that's quite a rare thing and b) bots are friable - they come and go as users or their interests come and go. Constraints advertising that a bot will handle something will sooner or later falsely advertise that a bot will do something.

Overall: few people actively care about the issues you're raising, and most people - devs and users - are busy doing other stuff. So whilst it is unimpeachably the case that you can suggest and believe that x & y would be improvements, you'll probably discover little appetite to discuss the issues and less appetite to make changes. I'd rate the likelihood that wikidata's technical owners, WMDE, will do anything in this space at about zero b/c busy. It really is the de facto policy that the UI does what it does / does as little as it needs to in order to be basically usable. The community builds on the infrastructure provided with 1,001 greater or smaller hacks. So it is more easy to conceive of improved help of the sort you think desirable via a gadget ( by which I mean https://www.wikidata.org/wiki/Special:Preferences#mw-prefsection-gadgets ) or via css or javascript hacks. And, on this note, you'd probably be surprised at how much, shall we say, after-market css / js / gadget / tool stuff exists to modify or replace the wikidata UI. Equally, you'd be unsurprised to learn that little of the css / js stuff is accessibly documented. iirc there is a bit more documentation for toolforge tools (or, at least, a registry), but it's all still pretty inaccessible for other than fairly experienced users. I'm guessing all of this is not the answer you want, but there we go. --Tagishsimon (talk) 15:10, 22 December 2020 (UTC)

• Whether or not a required qualifier is mandatory or suggested depends on constraint status (P2316): qualifier to define a property constraint in combination with P2302. Use values "mandatory constraint" or "suggestion constraint". In this particular case the constraint is tagged as suggestion constraint (Q62026391): status of a Wikidata property constraint: indicates that the specified constraint merely suggests additional improvements, and violations are not as severe as for regular or mandatory constraints.

The UI tries to tell the user that these are suggestions by writting Suggestions over them.

When it comes to standardized locations for help, there's the talk page of a property. A property itself can have Wikidata usage instructions (P2559): text describing how to use a property or item. Eventually, this statement is to replace usage instructions currently found in description. Until the corresponding feature is implemented, do not remove it from description while the constraints have constraint clarification (P6607): qualifier to provide details on the function or purpose of a property constraint on a given property attached to them.

I thought a bit about it now and maybe "desired qualifier constraint" would be an improvement over "required qualifier constraint. See https://www.wikidata.org/wiki/Talk:Q21510856#mandatory_%E2%86%92_required ChristianKl ❪✉❫ 15:33, 22 December 2020 (UTC)

Christmas in <insert country here> pages

Prompted by something that was shared at work, I looked at pages that link to Christmas tradition (Q717040). This includes a bunch of pages that are "Christmas in <country>". Most of these are marked as subclasses of Christmas tradition (Q717040), which seems wrong IMHO (such a page may list multiple traditions that are typical for that country or part of the world). Is there any convention for how to classify pages about a topic in a particular country? (this could apply to many other types of topic summaries) Pauljmackay (talk) 13:08, 21 December 2020 (UTC)

@Pauljmackay: You could use facet of (P1269) = Christmas tradition (Q717040), but then what would you make the item an instance of or subclass of (which is essential for its classification and discoverability)? subclass of (P279) is actually the correct relation for some more specific narrowing of an abstract concept. Jheald (talk) 18:27, 21 December 2020 (UTC)

@Jheald: actually I wonder if defining these pages as subclasses of Christmas (Q19809) and qualifying that with the country is best... Pauljmackay (talk) 11:42, 22 December 2020 (UTC)

Working with geospatial streetlight data as it might relate to economic, ecological, and public health data sets

https://www.wikidata.org/wiki/Q503958 is a wikidata node about streetlights and https://gis.data.vbgov.com/datasets/4469b7926ef94ec4bf2e176a2dffbe8e_2?geometry=-76.318%2C36.826%2C-75.824%2C36.922 is a geospatial database of streetlights in virginia beach…

What's the process of connecting those resources in wikidata - what belongs, what doesn't... in the bigger picture we're trying to understand the value of streetlights as they relate to public safety, ecosystem, and economic impact factors etc... and it would be helpful to maybe have some examples of other urban features that are similarly impactful and look at how they are handled in this system = pointers would be greatly appreciated. i guess another way of looking at this is whether it would alternately belong as part of the Virginia Beach https://www.wikidata.org/wiki/Q49259 data set...

Thank you @perryprog for the IRC conversation and also for pointing out this gem https://github.com/mfbx9da4/brightpath-backend Ansondparker (talk) 23:04, 22 December 2020 (UTC)ansondparker

• Whether or not it belongs is a community decision where different people are likely to have different opinions the way to get a decision would be to create a property proposal for an ID for the individual streetlights. ChristianKl ❪✉❫ 23:11, 22 December 2020 (UTC)

Change the Kazakh (kk) labels/descriptions/aliases to use Latin scripts or not?

FWIW, from next year, the Kazakhstan will officially start transition of official Kazakh scripts from Cyrillic to Latin, currently there are too much of "kk" labels, along with descriptions and aliases, on items and properties, maybe also Lexemes, are written in Cyrillic scripts, though ~60%(?) of them are also manually typed kk-cyrl/kk-latn/kk-arab, what about also starting transition for kk here? --Liuxinyu970226 (talk) 11:47, 17 December 2020 (UTC)

Ask kkwiki administrators on this question? @AlibekKS, Arystanbek, Ashina, Batyrbek.kz, Daniyar:@GanS NIS, Kaiyr, Kasymov, Qarakesek, Sagzhan:@Sibom, Ulan, Мықтыбек Оразтайұлы, Мұхамеджан Амангелді, Нұрлан Рахымжанов:@Салиха: --Liuxinyu970226 (talk) 11:51, 17 December 2020 (UTC)

•  Oppose for the time being, but may support later, right now, most Kazakh websites are still not yet started the transitions. What we need to do is only waiting, if one day most websites did such transitions, changing kk to use Latin can be a good idea even Kazakh Wikipedia doesn't do so. --117.136.1.181 00:12, 18 December 2020 (UTC)

..Good day! Change fonts to Latin early. The measures taken on the transition to the Latin alphabet, and in the present time are discussed in his introduction.Қатысушы талқылауы:Салиха--Салиха (talk) 05:31, 18 December 2020 (UTC)

•  Oppose --Kaiyr (talk) 09:34, 23 December 2020 (UTC)
• Let's wait atleast till transition will start. Right now we could standarise name of latin version code of Serbian (righ now it is sr-ec instead of sr-latn and sr-ec instead sr-cyrl). Eurohunter (talk) 11:18, 23 December 2020 (UTC)

Deprecating pronunciation audio on QIDs?

Please join to the discussion on the talk page. Right now nobody objects and multiple supports to deprecation.--So9q (talk) 07:04, 23 December 2020 (UTC)

How are these 2 items different - to be merged?

Both entities seem to refer to the same concept. Q157828 and Q97940890.

The first has data such as players who played for the team etc associated with it. The second which is quite new has team performance under P1344 etc. yet they seem to be about the same thing. Normally, I would simply merge them, but just want to make sure I am not missing something. CanadianCodhead (talk) 14:32, 14 December 2020 (UTC)

• A way to deal with a question like this is to ask the person who created the new item why they didn't reuse the existing one. @Balû:, do you think that Holstein Kiel (Q97940890) provides value being separate from Holstein Kiel (Q157828)? ChristianKl ❪✉❫ 15:13, 14 December 2020 (UTC)

Thanks @ChristianKl:, I had considered that, but based on their contribution log, which shows no activity since August, my assumption was they were no longer actively participating on the site, thus the broader outreach. CanadianCodhead (talk) 17:51, 14 December 2020 (UTC)

Unfortunately, it does not appear this is a one-off, the user appears to created duplicates of a significant number of German football clubs all of which will require cleaning https://www.wikidata.org/w/index.php?title=Special:Contributions&offset=&limit=500&target=Bal%C3%BB&namespace=all&tagfilter=&newOnly=1&start=&end= As noted, they no longer appear to be an active user on the site. CanadianCodhead (talk) 16:14, 18 December 2020 (UTC)

@CanadianCodhead:, @ChristianKl: sorry for late answer: the two items are different concepts: Holstein Kiel (Q157828) is the club/association with many different sport divisions amongst others football. Holstein Kiel (Q97940890) is only the first football team of men if this sports association. There are also a second team, a women team, youth teams and teams of other sports like handball. An association shouldn't have team performance P1344 - it isn't one team, there are many teams in one sports association. --Balû (talk) 18:36, 22 December 2020 (UTC)

@Balû: Thanks, I changed the instance of (P31) of Holstein Kiel (Q157828) to sports club (Q847017), does that seem right? ChristianKl ❪✉❫ 18:53, 22 December 2020 (UTC)

I don't think that is an accurate assessment The history of the item, from the associated statements, labels, the associated external links etc,even the Wikipedia link itself not to mention the hundreds of player pages that have the value listed under member of sports team clearly indicate this item was being interpreted as the first team football club, not the umbrella sports organization for the last 8 years. Is someone going to go redo all that data onto the 'new' entity now? CanadianCodhead (talk) 19:24, 22 December 2020 (UTC) @Balû: @ChristianKl:

• The Wikipedia link says "Other departments are team handball (Men and Women), Tennis, and Cheerleading. The women's handball team won the 1971 German handball championship. " That does indicate that the English Wikipedia article includes the parts that are not the football club. ChristianKl ❪✉❫ 19:35, 22 December 2020 (UTC)

So basically the feedback is to start correcting the thousands, more likely tens of thousands of items that now have inaccurate statements due to changing what the original item refers to and repurposing dozens of pages that unquestionably were being used to represent the football team to now be the umbrella organization?

I'm not qustioning the need to have the umbrella organizations as items, they should be in the database, but this does not seem the right way to do it. CanadianCodhead (talk) 19:22, 23 December 2020 (UTC)

A chance to redeem ourselves

As I've argued before, I think we failed quite badly at making ourselves the centralized repository of COVID-19 case/death/recovery statistics. Now that the vaccine has started rolling out, a user pointed out that there's a new type of data we're going to be seeing: vaccination counts. Do we even have a property for that? This is a chance for us to redeem ourselves, but we need to jump on it and set things up so that it's easy for e.g. Wikipedia infoboxes to display the counts stored here. It'll be an uphill battle even so, as we'll need to ensure that every place the vaccination counts appear on Wikimedia is sourced from us, but if we snooze and the data starts getting stored simultaneously in a gazillion different places, we're sure to end up with the same mess we got with the case/death/recovery stats. Any takers? {{u|Sdkb}} ^talk 06:36, 17 December 2020 (UTC)

@Sdkb: Have you analyzed why your previous suggestion didn't work? This doesn't seem like the sort of data Wikidata is intended for, since it changes on a daily basis. But if language Wikipedia editors are wasting their time duplicating effort they are welcome to propose a process to use Wikidata that would save them time. ArthurPSmith (talk) 13:38, 17 December 2020 (UTC)

I would not consider the (absense of complete) COVID-19 case/death/recovery/vaccination statistics a failure. There is a ton of such data pulished each day by various entities for various geographic areas, often inconsistent, often subject to changes at a later point. No chance to replicate this in Wikidata anywhere remotely close to "live". However, in five years from now most COVID-19 related data sources will probably not be available any longer. If we have something permanent by then, it’d be a success. Yet I am not sure whether Wikidata or Data at Commons are the best place to host such data. —MisterSynergy (talk) 13:55, 17 December 2020 (UTC)

• I agree with MisterSynergy and ArthurPSmith that it's not Wikidata core agenda to provide this kind of data. Wikidata can be used this way and if it's helpful for Wikimedia projects to use Wikidata here, I'm happy that the profit from Wikidata, but my efforts within Wikidata are directed elsewhere. While COVID-19 is a very important topic, I don't think whether or not we have vaccination data on Wikidata will do anything about the health outcomes.

That said if you want to put a lot of effort into bringing this kind of data to Wikidata in a quality that user Wikimedia projects want to use, go ahead. ChristianKl ❪✉❫ 14:11, 17 December 2020 (UTC)

I see it as part of Wikidata's agenda to serve as a centralized repository for factual information that is used on many different language Wikimedia projects, and this data is definitely in that category. {{u|Sdkb}} ^talk 19:57, 19 December 2020 (UTC)

It is not in that category. Wikipedia is a very poor venue for tabular data, because large-size items cause replication lag in the serialisation to RDF process; quite apart from the wikidata UI being rubbish for tabular data. That is why tabular data on commons has been developed. It is ideal for data like this to be served from a single repository, but a) talk of "redemption" is crass and b) language wikipedias will move at whatever speed they move. You coming here and suggesting inappropriate uses of wikidata does not help anyone. --Tagishsimon (talk) 20:05, 19 December 2020 (UTC)

Hello, all. Yep, the challenge is big, especially when we consider the imense variety of reliable sources. Historical case data (day to day) is really hard to keep track with multiprovenance. I have a suggestion for a solution for the most current data points. I will write a script to daily scrape the Wikipedia template which uses Johns Hopkins data (this one) and feed it as qualifiers to the parts in COVID-19 pandemic by country and territory (Q83741704). It is not perfect, but at least makes individual data points reachable via SPARQL and via Lua modules. I'll just "be bold" and do that, but any additional thoughts are super welcome. TiagoLubiana (talk) 17:46, 23 December 2020 (UTC)

Update: I'll actually do it for the item Template:COVID-19 pandemic death rates (Q102044164) , which seems more coherent. TiagoLubiana (talk) 19:04, 23 December 2020 (UTC)

Two ISBNs

Ethnic Identity and the State in Iran has two ISBNs (Print and Online), but two ISBNs result in error. How can it be fixed? Pirhayati (talk) 17:03, 23 December 2020 (UTC)

• Both versions should have their own item that are linked via edition or translation of (P629) has edition or translation (P747). ChristianKl ❪✉❫ 17:15, 23 December 2020 (UTC)
• @Pirhayati: If you look at ISBN-13 (P212) than you see the single-value constraint (Q19474404) has a list of separator (P4155) which includes distribution format (P437). You can add that as a qualifier. Multichill (talk) 17:19, 23 December 2020 (UTC)

Can we automate the deprecation of less precise birth and death dates

See for example at Alfred Thomas Rogers (Q94293371) where I manually deprecated the less precise value. I used to delete the less precise values but bots just keep re-adding them. I don't mind them there anymore, they help find errors when the years do not match. But if we deprecate them when a precise value is also included it will aid in some of the searches I do looking for people with less precise dates that I can upgrade to a precise value by looking for primary documentation. One search I do is to look for people that would appear in the US WWI and WWII draft registrations where I can get a precise birth date. --RAN (talk) 03:45, 12 December 2020 (UTC)

are the less precise values wrong? why deprecate them? one may have two sources which provide the date at different precisions. including both is valuable and neither is really deprecated. can we not just make it easier for you to identify/use the more precise value when multiple are present? BrokenSegue (talk) 04:05, 12 December 2020 (UTC)

• This is not an error or a wrong value, I thought I made that clear. In this case "reason_for_deprecation=item/value with less precision and/or accuracy". When they both have the same rank it implies that both are equally wrong or equally correct. Maybe someone was born on June 30, 1872 or maybe that is wrong and we only know that they were born some time in the year 1872. I assume that is the whole reason we introduced "reason_for_deprecation=item/value with less precision and/or accuracy". Also when a field has multiple values with the same rank, the infobox displays all the values for date_of_birth and date_of_death. See for instance: Commons:Category:Erika_Morini --RAN (talk) 05:17, 12 December 2020 (UTC)
  • You need to use "preferred rank" on the preferable (usually: most precise) claim, "normal rank" on other claims that are correct, and "deprecated rank" only for those which are incorrect, e.g. because they were for instance stated in a reliable source. Unfortunately, "deprecated rank" and the "reason_for_deprecation" qualifier are misused by many editors. —MisterSynergy (talk) 08:15, 12 December 2020 (UTC)

Q94293371 was done wrongly. The deprecated (though less precise statement) does have a source, the preferred one has no sources. This is a very bad precedent. --Matěj Suchánek (talk) 10:26, 12 December 2020 (UTC)

All those references are added by bots, they are not done by hand one-by-one. We have a bot that adds Findagrave and other references that use full dates. They are best handled by the bot, it doesn't create typos. I have a whole list of entries with transposed numbers for dates from a human transferring them from their source to Wikidata. --RAN (talk) 07:35, 13 December 2020 (UTC)

• item/value with less precision and/or accuracy (Q42727519) has no uses and it likely shouldn't exist in it's present form. What do we want to do with it? ChristianKl ❪✉❫ 12:29, 12 December 2020 (UTC)
  • Burn with fire. Multichill (talk) 16:02, 12 December 2020 (UTC)
    • While I understand the sentiment, maybe deprecating it's instance of (P31) Wikibase reason for deprecated rank (Q27949697) and giving it a description that makes it clear might be the better way to go about it. ChristianKl ❪✉❫ 16:46, 12 December 2020 (UTC)
• Ok, so the plan is to not deprecate year-only values, but to uprank the more precise value. That is fine with me, can we automate it? We do not need both values in the infoboxes. Perhaps we should create reason_for_preferred_rank=item/value with more precision and/or accuracy" as a way to explain the change in rank. What do you think? --RAN (talk) 07:30, 13 December 2020 (UTC)
  • Yeah, but this should only be done with more precise values with a reference to a respectable source, and only when the items do not already have some bestrank value statement (do not override). I am not sure if this can be easily automated.Vojtěch Dostál (talk) 13:52, 13 December 2020 (UTC)
  • Sources are more important for deciding what's the best rank. In the example of Alfred Thomas Rogers (Q94293371) the more precise claim has a source that's user generated content and thus wouldn't be welcome on EnWiki. While it's not a huge drama to have some items be like that, it's not desireable to automate the process in a way where every item should be set that way.

If you do a search for items that could be improved by adding good sources from primary documentation like US WWI and WWII registrations it's reasonable for this item to show up. ChristianKl ❪✉❫ 18:13, 13 December 2020 (UTC)

These were errors found in "Archivio Storico Ricordi" in just one quick pass: Johann Wilhelm von Mirbach-Harff (Q94785517) and Gabriel Astruc (Q3093605) and Giuseppe Taddei (Q471393). "Archivio Storico Ricordi" is one of our trusted sources being used to add in year-only dates. Also look at our pages of errors we are tracking in LCCN and VIAF. Please do not assume Findagrave has a higher error rate or a lower error rate than those sources until you have calculated that error rate. --RAN (talk) 03:27, 14 December 2020 (UTC)

I didn't speak about error rate but about it being user-generated content. Whether or not I want this to be significant it matters to Wikipedian's. ChristianKl ❪✉❫ 12:49, 14 December 2020 (UTC)

• There are over 100,000 entries with precise-dates and year-only dates displaying in infoboxes. I don't see that being corrected one-by-one by hand. If a precise value is wrong, then that should be deprecated by hand, and assumed to be correct until proven incorrect. --RAN (talk) 02:37, 14 December 2020 (UTC)
  • Wikipedian's don't want data in their infoboxes that's assumed to be correct until proven incorrect. Automating the process of putting data into infoboxes that Wikipedian's don't want there damages the acceptance of Wikidata by Wikipedian's and makes it harder to get infoboxes used more widely within Wikipedia. ChristianKl ❪✉❫ 12:49, 14 December 2020 (UTC)

You wrote: "Wikipedian's don't want data in their infoboxes that's assumed to be correct until proven incorrect" yet almost all data is unreferenced in Wikidata. That is why we are creating bots to add in references. There are about 15 fields in a standard human infobox and only 5 require a reference. I am pretty sure that English Wikipedia editors do not want to see two entries for date_of_death in their infoboxes, such as "December 25, 2020" and "2020", which is my whole point, to do something so the less precise one does not display, which is why I started this thread. --RAN (talk) 23:44, 14 December 2020 (UTC)

The EnWiki-RFC on Wikidata does indicate that they want references for every statement that comes from Wikidata even when it's allowed to make certain claims within Wikipedia without providing sources for it. The infobox should ask for a truthy value and thus not display both "December 25, 2020" and "2020" unless it desires to get people who see the infobox do the work of setting the ranks.

There's no problem if the infobox displays 2020 even so we have an statement that says "December 25, 2020". ChristianKl ❪✉❫ 17:36, 15 December 2020 (UTC)

• My 2c: my modus operandi when I find two dates, one more precise (day) and one less precise (year), provided that both have reliable sources, is: to order the most precise as first (using User:Tohaomg/rearrange values.js) and to give it preferred rank, while leaving other statements with normal rank. I consider wrong marking less precise statements with deprecated rank. --Epìdosis 17:48, 15 December 2020 (UTC) P.S. Moving this discussion in Help talk:Deprecation would maybe make it easier to find it in the future
• I agree the best strategy is to promote the more precise date, rather than demote the less precise date. We still need a way to automate it so that both dates stop appearing in the infoboxes. If you are worried about references there are half a dozen bots adding references for dates, most are adding references for the year-only ones, which is the source of the problem. One was using Findagrave for full dates, but I can't find the bot statistics. We need another full date bot using other sources like Familysearch. --RAN (talk) 22:10, 19 December 2020 (UTC)
  • @Richard Arthur Norton (1958- ): Which infobox currently shows both, can you give an example? Maybe the solution is to directly tell the person looking at the infobox to apply ranks. On Commons that might be a way to engage users to become editors. ChristianKl ❪✉❫ 13:32, 24 December 2020 (UTC)

Common base type for end cause (P1534) values?

I was browsing the possible values for end cause (P1534), and noticed there's probably some ambiguity about what's appropriate for this property. For instance, there are uses of Death (Q161936), legal death (Q6517655), death in battle (Q18663901), death (Q4) and death in office (Q5247364). I might argue that only the last is appropriate for this property, though I'll admit not carefully examining the use of the others. I did fix up a few uses of disqualification (Q1229261) which seemed clearly wrong as well as a few instances of health (Q12147). ( See my history from a few moments ago if you want to see them all. )

So my question is whether it makes sense to have a common base type for things like disqualification (Q40430404) to make it easier to put property constraints on end cause (P1534) to make this less likely. Or perhaps we prefer the extra flexibility even when it may lead occasional confusion when adding qualifiers?

This is not something that I've been focusing on, but I'll be happy to follow up as appropriate. Regards, Gettinwikiwidit (talk) 01:02, 23 December 2020 (UTC)

• I don't think ontology modeling that sees "death in office" as being in about Wikidata is a good idea (via Q66593009). The ontology should focused on what the items are actually about. What's an appropriate end cause (P1534) depends also a lot on the context. ChristianKl ❪✉❫ 01:50, 23 December 2020 (UTC)

[Moving this to the right section.] I'm not following. Honestly I'm having trouble parsing the grammar. Do you agree that there is ambiguity here and that some of these cases seem clearly wrong? Gettinwikiwidit (talk) 01:55, 23 December 2020 (UTC)

I edit for clarity. Values of end cause (P1534) shouldn't be Wikidata internal entity (Q21281405). ChristianKl ❪✉❫ 02:02, 23 December 2020 (UTC)

@ChristianKl: I'm still not following. Most of these end cause (P1534) values have associated Wikipedia articles including death in office (Q5247364). Mostly what I'm asking about is the clear misuse of some values for this qualifier. Do you not agree that some of these are clearly inappropriate regardless of whether or not there is an associated Wikipedia article? Death (Q161936) seems wildly off base to me, though it has an associated Wikipedia article. Or are you suggesting that adding a subclass of (Q21514624) qualifier somehow makes an entity a "wikipedia internal entity" data? If so, I'm not sure I understand. Again, my suggesting is to simply add some typing to help people avoid what seem like clear mistakes. Regards, Gettinwikiwidit (talk) 13:29, 23 December 2020 (UTC)

The fact that "death in office" has a Wikipedia article means that death in office (Q5247364) instance of (P31) Q66593009 is wrong. ChristianKl ❪✉❫ 17:07, 23 December 2020 (UTC)

@ChristianKl: Now I'm really confused as this sounds like the exact opposite of what you said above and doesn't address the questions I've asked. Would you mind being a little more verbose? My understanding is that being Wikidata internal entity (Q21281405) is exactly the same as *not* having a Wikipedia article. Is this correct? If so, then since death in office (Q5247364) has a Wikipedia article than it is not Wikidata internal entity (Q21281405). Above you said values of end cause (P1534) shouldn't be Wikidata internal entity (Q21281405). Also, I'm not sure how this all bears on my concern that people are clearly using the wrong values for end cause (P1534) regardless of whether they are Wikidata internal entity (Q21281405) or not. Do you disagree? If so would you mind explaining why? I'm really only trying to understand. At this point I have no tools for evaluating your concern. Regards, Gettinwikiwidit (talk) 21:41, 23 December 2020 (UTC)

@Gettinwikiwidit: We have Q66593009 as a class for entries like does not exactly match (Q42415624) which are Wikibase reason for deprecated rank (Q27949697). Those are items that are internal to Wikidata and that among others don't have Wikipedia pages. They are also often longer descriptions of a reason for deprecation.

When it comes to "end cause" we use items that are not internal to Wikidata as examples of causes. Thus the causes shouldn't be instance of (P31) Q66593009.

In the last months Q66593009 got created (and I now filled a Request for Deletion) which does have the problem of declaring a bunch of items that in the nature aren't about what happens in Wikidata (the reasons for deprecation in Wikidata are about what happens in Wikidata).

The answer to what kind of thing death in office (Q5247364) is, shouldn't be about end cause (P1534) but about what would be true about death in office (Q5247364) even if Wikidata wouldn't exist. To answer it well we should do research about the concept and how it's modeled elsewhere and not look inward. If you do that research you might find that there's an existing notion that's helpful for setting up type constraints. ChristianKl ❪✉❫ 22:59, 23 December 2020 (UTC)

@ChristianKl: Thanks for getting back. I think I understand what you're saying. That would suggest that end cause (P1534) shouldn't be used as a qualifier for position held (P39) unless it was describing a reason for removing such a claim. This is the vast majority of the cases in the link at the start of this topic. The description in end cause (P1534) isn't really helpful for understanding its intended usage. Do you see how it is currently being used in position held (P39) statements though? Would you object to having some qualifier to serve that function? If not, then why not re-purpose it for this function instead of what you describe as its intended function? If you do object would you mind explaining why? Is it detail that doesn't belong in Wikidata for some reason? @Andrew Gray: Regards, Gettinwikiwidit (talk) 23:14, 23 December 2020 (UTC)

Also, to be clear, using death in office (Q5247364) as a value of a qualifier isn't so much about answering what death in office (Q5247364) is as it is describing the subject of the qualifier. Gettinwikiwidit (talk) 23:21, 23 December 2020 (UTC)

I do believe that Q66593009 shouldn't exist. Given that it seems multiple people like the property I didn't delete it directly but raised a request for deletion. With that removed using end cause (P1534) as a qualifier for position held (P39) is completely fine. end cause (P1534) is supposed to be a qualifier for anything that has inception (P571) which are lots of different properties.

@ChristianKl: That's not what the description of end cause (P1534) says. It says it's to be used with end time (P582). Maybe it would help if you pointed to which of these values use it the way you're describing. Gettinwikiwidit (talk) 23:40, 23 December 2020 (UTC)

@Gettinwikiwidit: Sorry here, I meant end time (P582). Maybe, I shouldn't trust myself to remember those numbers. ChristianKl ❪✉❫ 23:57, 23 December 2020 (UTC)

When we use end cause (P1534) we want to express that in the world outside of Wikidata the relationship we are speaking about ended for that reason. When we use Wikibase reason for deprecated rank (Q27949697) we are speaking about why we do deprecate within Wikidata.

Using death in office (Q5247364) as a value is not saying anything about what death in office (Q5247364) is. However giving death in office (Q5247364) a "Common base type" is about saying something about what it is. ChristianKl ❪✉❫ 23:32, 23 December 2020 (UTC)

I agree with this, which is why I brought it up in the first place. We could have Wikidata internal values only for such values if we wanted the benefit of the extra type checking and not imply extra meaning to the base entities. Or we could just live with people using the wrong values. Do you agree that Death (Q161936) is being used improperly here? Gettinwikiwidit (talk) 23:46, 23 December 2020 (UTC)

I don't have a strong opinion about whether Death (Q161936) is used improperly here. Do believe that research into how other ontologies model causation in similar cases would be needed to come to a good resolution. ChristianKl ❪✉❫ 23:57, 23 December 2020 (UTC)

I just saw that you referred to Death (Q161936): personification of death and not death (Q4): permanent cessation of vital functions. I agree that the former is clearly wrong. ChristianKl ❪✉❫ 12:19, 24 December 2020 (UTC)

• killed in action (Q210392) is another one that I add for military people. --RAN (talk) 02:37, 23 December 2020 (UTC)

Is this related specifically to position held (P39) claims in the query linked above? There is only one instance there and I'm not sure how it applies specifically to that claim. Gettinwikiwidit (talk) 13:29, 23 December 2020 (UTC)

page banner (P948)

Please see Property_talk:P948#Single_value_constraint and share any better idea for the more appropriate qualifier or reference to be used in this case. Thanks, --Andyrom75 (talk) 12:24, 24 December 2020 (UTC)

Gregorio di Lorenzo and The Master of the Marble Madonnas

There seems to be modern research that establishes an identy between Gregorio di Lorenzo (Q3776551) and the Master of the Marble Madonnas (Q1792231) (Alfredo Bellandi: Gregorio di Lorenzo: Il Maestro delle Madonne di Marmo; Uno scultore e la sua bottega nel Rinascimento. Morbio Inferirore 2006; Alfredo Belandi: Master of the Marble Madonnas. In: A. Butterfield, A. Radcliffe (Hrsg.): Masterpieces of Renaissance Art: Eight Rediscoveries. New York 2001, S. 34–40). We have already cross linked the Commons Category. But I am not art historian enough to dare to make the final merge. We have articles as far as I see in Englisg, German and Italian Wikipedia pertaining to the same subject. --Wuselig (talk) 12:37, 24 December 2020 (UTC)

You've linked them with said to be the same as (P460). That may be the best we can do? --Tagishsimon (talk) 12:55, 24 December 2020 (UTC)

Individual books in the Yale Center for British Art library

The Mix'n'Match catalog import of Yale Center for British Art artworks includes a number of books in the Yale reference library (example: https://mix-n-match.toolforge.org/?#/entry/35571098 and https://collections.britishart.yale.edu/catalog/orbis:580157). These are records for individual books in the library, not for works or editions. Do we even want to import these into Wikidata, or should we mark them all N/A?

(Creating these items from Mix'n'Match currently makes them P31="artwork" which is obviously wrong, and changing the P31 causes 7 constraint violations for Yale Center for British Art artwork ID (P4738).) - PKM (talk) 22:59, 23 December 2020 (UTC)

• I think there are some individual books like the remaining copies of Gutenberg's bible that are individually notable and deserve their own Wikidata items. I'm unsure whether or not the books we are talking about here are special in a similar sense where individual copies are notable. How many books are in that collection? ChristianKl ❪✉❫ 23:59, 23 December 2020 (UTC)
• This topic has come up several times before, whether to merge existing entries on copies of the same book, or same engraving. There has never been a firm decision. But I agree we should not add more unless there are listed in the Wikipedia article like the extant copies of the Gutenberg Bible. --RAN (talk) 00:10, 24 December 2020 (UTC)
  • I'm not sure how many there are total; MnM search has a limited number it will return. Many of these are old exhibition catalogues, and some are art books published in the last few decades and readily available. The majority of the ones I looked at don't have matching work or edition items at all. - PKM (talk) 00:18, 24 December 2020 (UTC)
  • Update: I've requested a bulk "N/A" edit for these on Magnus's talk page. - PKM (talk) 20:36, 24 December 2020 (UTC)

Many items never have an English label

I recently encountered Thomas Hickling (Q10382186): this items had labels in six languages but not in English. An English label is vital in language fallback. There are multiple users adding labels in various languages to this item (Albanian, Asturian, Catalan, Slovene), yet nobody added an English one. Currently, there are more than 320000 items for people that have label in at least one major language in Latin script, but not in English.--GZWDer (talk) 20:17, 20 December 2020 (UTC)

what action are you suggesting here? we could have a bot populate it I guess but I believe search already will search across different languages. BrokenSegue (talk) 20:21, 20 December 2020 (UTC)

Wikidata interface displays label in user language, with fallback to English, but not the other direction (see phab:T89213). Some tools displays English labels only. So having an English label is important. The example I listed is even an American.--GZWDer (talk) 20:42, 20 December 2020 (UTC)

I would suggest requesting a bot job that can detect a Latin script label and run the default label set from nameGuzzler.js (aka VIP's Labels) on these items. - PKM (talk) 00:11, 21 December 2020 (UTC)

I can put together a bot script to do this fairly straightforwardly if you want. Pi bot already adds English labels from Commons categories, it's just a matter of changing the selection query and the source of the labels - I would need a list of languages that it is OK to copy from though. Unless @GZWDer: you or someone else is planning on doing this? Thanks. Mike Peel (talk) 06:35, 21 December 2020 (UTC)

I'm no bot wrangler, but a bot job seems most appropriate for human items, in which personal names are unlikely to vary significantly (or at all) between languages: a Jane Smith or Olaf Svenson are likely to be known as such in nearly every Latin-script language (Hungarian naming conventions, or names with "junior" or other suffixes are notable exceptions). given name (P735) and family name (P734) might be queried as well, if present. Items for non-human, non-proper noun entities will likely require more human oversight to properly translate the label into English. -Animalparty (talk) 04:35, 21 December 2020 (UTC)

• The main problem is that people with Cyrillic names can and sometimes are transcribed differently in some languages (e.g. the French label of an item about a Russian politician may not be suitable for its English label).
  There are also Latin script labels that are generally not suitable for English (e.g. Latvian, Hungarian, Latin, Esperanto),
  also, in some cases, Spanish, Catalan (Ancient Romans). There are also a few names that could be translated. --- Jura 10:12, 21 December 2020 (UTC)
• @GZWDer: The other issue I have noticed that is ENWP users often (it's like always) not linking their new articles, lists, templates, categhories etc. to Wikidata and they remain never linked to Wikidata till I discover them. Eurohunter (talk) 16:06, 21 December 2020 (UTC)
  • @Eurohunter: You are welcome to comment on Wikidata:Requests_for_permissions/Bot/RegularBot_2.--GZWDer (talk) 16:08, 21 December 2020 (UTC)
• I've coded this up and started a bot request at Wikidata:Requests for permissions/Bot/Pi bot 17. Thanks. Mike Peel (talk) 19:38, 21 December 2020 (UTC)
• Another low hanging fruit (but much more useful, than copy Latin to English) is to auto-transliterate Japanese names, which have name in kana (P1814) - like these 13342 items, and some more. I did this few years ago for Cyrillic (example) with custom Polivanov transliterator. It's basically just a question of "modified Hepburn" code. --Lockal (talk) 21:02, 25 December 2020 (UTC)

How to name obits

I notice we have a mix of entries for obituaries as "Label=John Smith (1800-1900) obituary" with "title=John Smith Dies, age 100" and others as the actual title of the obit like "Label=John Smith Dies, age 100". Do we have a preference for one format over the other? --RAN (talk) 06:06, 25 December 2020 (UTC)

Ideally we go with whatever the title of the obit is, presuming there is one, and use the description clarify that it's an obit if the label does not make that clear. No clue where there's not a title. --Tagishsimon (talk) 06:28, 25 December 2020 (UTC)

The best place for a discussion about names of obituary would be https://www.wikidata.org/wiki/Talk:Q309481 the project chat is not a place where any discussion that leads to a naming preference is going to be found again. ChristianKl ❪✉❫ 15:00, 25 December 2020 (UTC)

We can migrate the thread there after, I don't think many people will see it there. --RAN (talk) 19:45, 25 December 2020 (UTC)

redirect (Q45403344)

Not a identifier. What to replace the word "identifier"? Eurohunter (talk) 13:09, 25 December 2020 (UTC)

• The point of the item is to talk about identifiers that are redirected. No reason to replace anything. ChristianKl ❪✉❫ 14:57, 25 December 2020 (UTC)
  • @ChristianKl: But it really isn't used for identifiers. Eurohunter (talk) 15:31, 25 December 2020 (UTC)
    • It's a Wikidata reason for depreaction, it's not Wikimedia redirect (Q21528878) (which in turn is not to be used with instance of (P31) per constraint). It seems like the current usages are all not notable items or added to an item to which the didn't belong. I deleted the non-notable items. ChristianKl ❪✉❫ 15:51, 25 December 2020 (UTC)

Problem with Egyptian Arabic Wikipedia

User @HitomiAkane: from Egyptian Arabic Wikipedia is not allowing users to create redirects from original names in its original alphabets. He has noticeable requests about that at his talk page which he refuses. He refers to local policy page at arz:ويكيبيديا:طريقة الكتابه but this kind of redirects are present in other versions of Wikipedia and I will link for example redirect to en:Egyptian Arabic from en:اللغة المصرية العامية. He says that "creating redirects in another language means the we should create over one million redirects, and it'll be all useless" which is probably false because Egyptian Arabic Wikipedia has now over 1,180,000 articles and I don't guess they all are all releated to items which names are written in other alphabet than Egyptian Arabic in other words all these articles aren't for example about British cities so they would require redirects from original name in latin alphabet. So with this unique expection from Egyptian Arabic Wikipedia it will collide with Wikidata items. Summarizing you can't add redirect to original name in its original alphabet at Egyptian Arabic Wikipedia but you can do it at any other version of Wikipedia. That's atleast odd for me. Eurohunter (talk) 15:55, 25 December 2020 (UTC)

• Wikidata is not the place where the policies of the Egyptian Arabic Wikipedia get decided. Given that you don't read Arabic, what make you think that you know their redirect policy better then them? ChristianKl ❪✉❫ 16:11, 25 December 2020 (UTC)
  • @ChristianKl: What make them think they know better than other over 300 Wikipedia projects? Why they can have exception? Eurohunter (talk) 16:14, 25 December 2020 (UTC)
    • Every Wikipedia project has it's own policies about what's notable on their project. ChristianKl ❪✉❫ 16:19, 25 December 2020 (UTC)
      • Agree. Every Wikipedia has its own rules and I cannot see how this impacts Wikidata, so, it's also the wrong place to discuss. The right place would be the Egyptian Arabic Wikipedia. --Christian140 (talk) 16:42, 25 December 2020 (UTC)
        • @ChristianKl: Thanks for your reasonable comments, as I explained so many times to user @Eurohunter: we only uses Egyptian alphabet in Masri Wikipedia as articles names and redirects, if we to create redirects in Latin alphabet we'll need to create hundreds of thousands of redirects ( ex: lake category contains over 80K articles ) and after all it's useless to the Egyptians users and against the local writing style policy, all I can see here is a conflict of interest towards one article which is @Eurohunter: interested in. HitomiAkane (talk) 17:29, 25 December 2020 (UTC)
          • @HitomiAkane: So would you like to say that something is wrong with whole rest of the world who create hundrests of these redirects in diferent alphabets? Why is this possible? Please explain. Where is the logic? Eurohunter (talk) 17:56, 25 December 2020 (UTC)
            • It's worth noting here that the claim that all other Wikipedia's allow these kind of redirects is false as well. I remember the Chinese Wikipedia also not allowing them. ChristianKl ❪✉❫ 20:09, 25 December 2020 (UTC)
• what does any of this have to do with wikidata? BrokenSegue (talk) 18:06, 25 December 2020 (UTC)

Wikimedia category and other spaces

I think "Wikimedia category" should be separated from "Wikipedia:Categorization" (and other spaces such as portal, "Wikimedia portal" and "Wikipedia:Portal") because these are two different items. "Wikipedia:Categorization" is suppoed to be Wikipedia editing guideline page while "Wikimedia category" should be dedicated item for instance of (P31). Due to Wikipedia:Categorization is used on many pages as "Wikimedia category" for instance of (P31), I would move all interwiki to new item under "Wikipedia:Categorization", so previous "Wikipedia:Categorization" could be changed to "Wikimedia category" and could be used with instance of (P31). Eurohunter (talk) 17:01, 25 December 2020 (UTC)

• Can you point to the items that you are actually concerned about? ChristianKl ❪✉❫ 21:06, 25 December 2020 (UTC)
  • @ChristianKl: It would be any category for example Category:Formula One drivers (Q7217610) or Category:American film actors (Q5837414). Eurohunter (talk) 21:50, 25 December 2020 (UTC)
    • Why shouldn't both of those be instance of (P31) Wikimedia category (Q4167836)? ChristianKl ❪✉❫ 22:04, 25 December 2020 (UTC)
      • @ChristianKl: It depends what Wikimedia category (Q4167836) actually is now. Is it "Wikimedia category" or "Wikipedia:Categorization"? Somehow these two iems are messed. Probably someone added wrong sitelink "Wikipedia:Categorization" to "Wikimedia category" item or someone added inforrect label for "Wikipedia:Categorization" or someone merged these two items for no reason. Eurohunter (talk) 22:58, 25 December 2020 (UTC)
        • An item is not a string. There's no reason why an item has to have the exact same name as the sitelinks. ChristianKl ❪✉❫ 23:13, 25 December 2020 (UTC)
          • @ChristianKl: If it is named "Wikipedia:Categorization" then it isn't "Wikimedia category". It wouldn't make sense otherwise. Eurohunter (talk) 23:36, 25 December 2020 (UTC)
            • It makes sense to me. ChristianKl ❪✉❫ 00:13, 26 December 2020 (UTC)
              • Isn't it verb vs noun, like "category" (perhaps type (Q21146257) is the closest item) and categorization (Q912550) are different things? One is a (virtual) thing an other is an activity, or process. Ghouston (talk) 01:22, 26 December 2020 (UTC)
              • Although, Wikipedia articles like en:Wikipedia:Categorization deal with both together, and are conflations of the verb and noun. en:Wikipedia:Categories is a redirection to the same page. Ghouston (talk) 01:27, 26 December 2020 (UTC)

Political alignment (falsely) a subclass of literary work

From my understanding, "subclass of" is transitive, so the following chain

political alignment (Q28819924) is subclass of political position (Q20045456) is subclass of certain aspects of a person's life (Q20127274) is subclass of biography (Q36279) is subclass of biographical work (Q15706467) is subclass of non-fiction (Q213051) is subclass of literary work (Q7725634)

means that

political alignment (Q28819924) is subclass of literary work (Q7725634)

which seems wrong to me. I see "certain aspects of a person's life (Q20127274) is subclass of biography (Q36279)" as the culprit, but I'm not sure enough to just remove it. What is the recommended course of action in this case?

Karlb42 (talk) 17:35, 22 December 2020 (UTC)

Agree. We're probably missing an item called "life of a person". (Not very clear what it would be a subclass of.) Biography could link to "life of a person" via main subject (P921). But certainly the culprit needs knocking on the head, and the wording of certain aspects of a person's life (Q20127274) steered away from use of the term biography. --Tagishsimon (talk) 17:47, 22 December 2020 (UTC)

• When in doubt the Project Chat is a good place to bring cases like this.

When it comes to deleting subclass of (P279), the goal should be that every item still has a valid subclass of (P279). Given that certain aspects of a person's life (Q20127274) currently has only one subclass of (P279) the question is about how to replace it.

As far of "life of a person" goes, this seems to me a part of (P361) and not subclass of (P279).

From my perspective "aspects of an entity" would be a better upper class. ChristianKl ❪✉❫ 17:56, 22 December 2020 (UTC)

• Thanks for your feedback! After looking around a bit, biographic information (Q59157079) looks like a suitable superclass. I would add that and change the existing relation to certain aspects of a person's life (Q20127274) is part of (P361) biography (Q36279), if not one disagrees or makes a better suggestion Karlb42 (talk) 17:18, 23 December 2020 (UTC)

By the way, currently political alignment is an action, an event, an activity, etc. In my eyes only one branch is correct: Political alignment - ideology - morality - ethical concept - etc.

--Lockal (talk) 13:03, 24 December 2020 (UTC)

I support removing the other branches. ChristianKl ❪✉❫ 15:03, 25 December 2020 (UTC)

Done in [8][9][10][11][12]. Last diff is dubious (certain aspects of a person's life (Q20127274)subclass of (P279)information (Q11028)), but better than nothing. Feel free to add more specific parent class. --Lockal (talk) 19:24, 25 December 2020 (UTC)

Oops, missed one message above, fixed last statement to biographic information (Q59157079). --Lockal (talk) 21:15, 25 December 2020 (UTC)

• Great, thanks! This was my first discussion in the wikidata community and I'm very happy with the process and result! Karlb42 (talk) 13:29, 26 December 2020 (UTC)

Creating template that shows the small logo or icon (P8972) of instance of (P31) of an item

I want to create a template that shows inline for an item small logo or icon (P8972) of the item's instance of (P31). If the item's has instance of (P31) but those have no small logo or icon (P8972), no icon should be shown. If the item's doesn't have instance of (P31), File:OOjs_UI_icon_cancel.svg should be shown. The idea is that this would allow showing an icon for many items if items that are commonly used as targets for instance of (P31) get a small logo or icon (P8972) statement. What's the best way to create such a template? ChristianKl ❪✉❫ 17:42, 22 December 2020 (UTC)

• I created a template to do that with {{Wikidata}}. For example, Douglas Adams was born in Cambridge. --SixTwoEight (talk) 19:15, 23 December 2020 (UTC)
  • @SixTwoEight: Thanks, very much. Currently it seems like Cambridge displays the File:OOjs_UI_icon_cancel.svg despite having 4 instance_of claims (even so 1 of them even has small logo or icon (P8972)). ChristianKl ❪✉❫ 00:06, 24 December 2020 (UTC)
    • @ChristianKl: The problem is that the template is only looking at the "truthy" value (which is determined by ignoring deprecated items, looking at all values of the highest rank, then taking the one closest to the top). So it only looked at college town (Q1187811), then gave up after not finding a small icon there. In this specific case the solution was to set city (Q515) as preferred, although I don't know if it's possible to implement "use the first instance of (P31) value that has an icon" in the template. --SixTwoEight (talk) 00:26, 24 December 2020 (UTC)
      • @SixTwoEight: From my perspective it would be good if it would show multiple icons if the item has multiple instance of (P31) which have a small logo or icon (P8972). It might also be worth thinking about whether there's a way to improve on the links when clicking on the icon/the hover text. ChristianKl ❪✉❫ 02:06, 24 December 2020 (UTC)
        • I did manage to make it work . It makes sense to think a bit more about naming but I like how far I got. ChristianKl ❪✉❫ 13:23, 24 December 2020 (UTC)
        • Now I'm at Douglas Adams (Q42): English author and humourist (1952–2001) and Cambridge (Q350): city in Cambridgeshire, England. ChristianKl ❪✉❫ 11:58, 26 December 2020 (UTC)

It is possible?

How viable would it be to import information from pornographic databases such as Internet Adult Film Database (Q1052713)?

I ask this because I use IAFD frequently but it leaves a lot to be desired when doing advanced searches (for example, there is no way to search for more specific things like "blue-eyed actresses who have done double penetrations and with 20 or more scenes/movies"), something that in Wikidata could be solved by doing a query with SPARQL. I mention IAFD because it seems to be the most complete site of its kind, but there are also others that might be interested in donating their data, such as:

• Adult Film Database (Q732004)
• Freeones (Q64411346)
• JavLibrary¹
• Boobpedia (Q25457707) (according to this the text is under a Creative Commons Attribution-Noncommercial-Share Alike 3.0 Unported License)
• AdultDVDTalk¹
• DefineBabe
• PornTeenGirl
• Babepedia.com (apparently not only limited to porn)
• EuroPornstar
• BabesAndStars
• MyPornstarBook
• Define Fetish
• European Girls Adult Film Database (Q3734894)
• Adult Web Movie Database (Q54871283)

1. It seems to be more focused on the movies than the actresses.

--190.226.77.62 11:39, 26 December 2020 (UTC)

• It's viable to create bot requests to import data from individual sites when the license allows it. Creative Commons Attribution-Noncommercial-Share Alike 3.0 Unported License is not a license that allows it. ChristianKl ❪✉❫ 11:50, 26 December 2020 (UTC)

subclass of (P279)

Schouldn't we have inverse property "superclass of" (currentlty qualifier superclass of (Q66088480)) for subclass of (P279)? Eurohunter (talk) 12:58, 25 December 2020 (UTC)

• No. It would make everything more complicated. ChristianKl ❪✉❫ 14:58, 25 December 2020 (UTC)
• For small (closed!) sets of things there is has part(s) (P527). --SCIdude (talk) 15:23, 26 December 2020 (UTC)
  • @SCIdude: has part(s) (P527) is never a valid inverse of subclass of (P279). If it's valid then subclass of (P279) was used wrong. ChristianKl ❪✉❫ 16:59, 26 December 2020 (UTC)

Help page and wikiproject for templates

Like any other wiki, Wikidata has lots of templates. To my knowledge, there is no documentation or help page for templates inside Wikidata. What about creating Help:Templates or Wikidata:Templates similar to ?

Also it would be nice to have Wikidata:WikiProject Templates to create some cooperation between template developers. PAC2 (talk) 07:28, 27 December 2020 (UTC)

• There's a lot of things on Wikidata that someone could do. If you want to write a good Help:Templates or create a Wikiproject that would likely be valuable. ChristianKl ❪✉❫ 13:13, 27 December 2020 (UTC)

Thank you for not adding a new property

I asked for a property and it was denied. This worked out really well for two reasons. I was given the gift of the old SourceMD It made it easy to add items for scholarly papers. By now I have added many papers in the field of nature restoration, rewilding and ecology. What is good about it, it brings in Scholia in perspective all the cited papers a paper relies upon. So when a paper is about the controversies in Wales about rewilding, it becomes feasible to get a perspective of an issue I try to understand. In a reaction to a Wikipedian who famously said; "I read hundreds of papers to change a line in a Wikipedia article", I read hundreds of papers to better understand a subject and thanks to Wikidata/Scholia there is something to show for it. Thanks, GerardM (talk) 08:35, 27 December 2020 (UTC)

Is it possible to create items in userspace?

I'm just wondering... JJP...MASTER!_{[talk to] JJP... master?} 01:36, 28 December 2020 (UTC)

No, the system will always create items in the main namespace. Special:ChangeContentModel does not offer items either. --Matěj Suchánek (talk) 10:06, 28 December 2020 (UTC)

If you want to create test items without adding them to main Wikidata we have https://test.wikidata.org/ for that. ChristianKl ❪✉❫ 13:02, 28 December 2020 (UTC)

P39 property values

Hi there,

I want to get all possible values for the property P39 using SPARQL, e.g. politician, spokesperson, diplomat, etc. Here is my query SELECT DISTINCT ?item ?itemLabel WHERE

   {
       ?item wdt:P39 ?itemVar.
     ?itemVar rdfs:label ?itemLabel.
     FILTER((LANG(?itemLabel)) = 'en')
   }

ORDER BY ?item LIMIT 5

It works without 'ORDER BY' statement, but I need the latter to get all possible values utilizing OFFSET. Any suggestions will be appreciated? Thanks in advance!

Such questions are better asked on Wikidata:Request a query. However, try:

SELECT ?position ?positionLabel ?count WITH { 
  SELECT DISTINCT ?position (count(?item) as ?count)
  WHERE 
  {
    ?item wdt:P39 ?position.
  } GROUP BY ?position } AS %i
WHERE
{
  INCLUDE %i
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}

Try it!

--Tagishsimon (talk) 09:00, 28 December 2020 (UTC)

Note, too, that P39 is 'position held' (position held (P39)). Politician, spokesperson, diplomat will be found under occupation (P106). --Tagishsimon (talk) 09:02, 28 December 2020 (UTC)

Wikidata weekly summary #448

Here's your quick overview of what has been happening around Wikidata over the last week.

• Discussions
  • Open request for CheckUser: 997kB
  • There's a discussion about what the best label for required qualifier constraint happens to be.

• Events
  • Ongoing: rC3 (remote Chaos Communication Congress). Several Wikidata-related talks and meetups happening on December 27th, 28th and 29th: introduction to Wikidata, Wikidata for datajournalists, Wikidata meetups in German and English, and a Query Service workshop. (see Wikipaka schedule for more details)

• Tool of the week
  • Template:TP administrative area is a new template designed for all items which are instance of administrative territorial entity. It defines useful queries for those items. For example, {{QT|Q173695}} links to this page. Feedback and contributions are welcome.

• Other Noteworthy Stuff
  • Project Grants open call for proposals in 2021. Changes in the review process for Project Grants in 2021; the open call for community organizing proposals will be from January 11 to February 10 and the open call for software and research proposals will be from February 15 to March 16.

• Did you know?
  • Newest properties:
    • General datatypes: small logo or icon, lunar coordinates
    • External identifiers: CINE21 person ID, Chinese Painting Database ID, Cinémathèque québécoise person ID, Humanitarian OpenStreetMap Tasking Manager project ID, SAPA ID, Manus Online author ID, Lichess username, CITWF person ID, DBLP publication ID, DataTrek ID, KANTO ID
  • New property proposals to review:
    • General datatypes: abstract, thesis committee chair, character type, Calificación de videojuegos, Regente, Nação (candomblé), Liderança, Bharati Braille, number of walks, number of hits, doubles hit, triples hit, Through point
    • External identifiers: TuneIn Podcasts show ID, RadioPublic show ID, Player.fm show ID, World Rowing ID, MNBAQ ID, Libsyn show ID, PodBean show ID, BG School ID, JAD ID, RDA value vocabularies ID, BHCL UUID, Podtail show ID, Provenio UUID, AbandonSocios ID, Mitre ATT&CK ID, Lexikon zum Literatur- und Kulturbetrieb im Österreich der Zwischenkriegszeit ID, International Encyclopedia of the First World War ID, SFLI ID, Game UI Database ID, DiVo person ID, e-GEDSH ID, Projecto Vercial author ID, Calaméo ID, Lumières Lausanne ID, Swiss Unihockey Player ID, IFF player ID, Internet Game Database numeric game ID, AV Production person ID, Magyar Életrajzi Lexikon ID, Studium Parisiense ID, I professori dell'Università di Pavia (1859-1961) ID, A*dS Encyclopedia, Viceversa Letteratura author ID, Ancient History Encyclopedia ID, The Podcast App show ID, viaggiareinpuglia.it ID, Encyclopedia of Saami Culture ID, DiCamillo Database Country House ID, International Standard Bible Encyclopedia ID, Biographical Archive of Psychiatry ID, Play:Right video game ID, Spletni biografski leksikon znanih Primorcev in Primork ID, TheCocktailDB ID, Muck Rack media outlet ID, Biographical Dictionary of the Czech Lands
  • Query examples:
    • Check Corona vaccine progress using Wikidata SPARQL to get metadata about the countries see notebook - SPARQL query

• Development

Due to the winter holidays, the development team is taking a break and no deployment is happening for Wikidata at the moment :)

You can see all open tickets related to Wikidata here. If you want to help, you can also have a look at the tasks needing a volunteer.

• Monthly Tasks
  • Add labels, in your own language(s), for the new properties listed above.
  • Comment on property proposals: all open proposals
  • Suggested and open tasks!
  • Contribute to a Showcase item.
  • Help translate or proofread the interface and documentation pages, in your own language!
  • Help merge identical items across Wikimedia projects.
  • Help write the next summary!

Read the full report · Unsubscribe · Lea Lacroix (WMDE) 18:33, 28 December 2020 (UTC)

Wikinews month pages

What should items like Wikinews:2020/January (Q88774275) be instances of, and how do they relate to items like January 2020 (Q55018341)? They contain sitelinks in the Wikinews namespace. Ghouston (talk) 22:16, 28 December 2020 (UTC)

If the Wikinews category sitelinks were on category items, like Category:January 2020 events (Q30304775), then perhaps the Wikinews namespace items could go on January 2020 (Q55018341) etc.? Ghouston (talk) 22:20, 28 December 2020 (UTC)

US senators occupying an office until a successor is named

Hello,

There are a lot of senators such as George Walton (Q878512) whose entry in the Biographical Directory of the United States Congress (Q1150348) have statements such as "appointed to the United States Senate to fill the vacancy caused by the resignation of James Jackson and served from November 16, 1795, to April 12, 1796, when a duly elected successor presented his credentials and took his oath". Is not re-elected (Q22087114) the appropriate end cause (P1534) for this? Obviously the vast majority of senators were "not re-elected" at some point, but I'm guessing marking them all is overkill. My focus has been supplying an end cause (P1534) when a senator has not made it to the end of a term. Is it worth distinguishing someone who completed a term of office from someone finishing a "term of service" if you understand my meaning? Here are the current uses, FWIW. ( Pinging a few people who have used this value of end cause (P1534): @M2Ys4U:　@Oravrattas: @Jura1: @Shinnin: )

SELECT * WHERE {
    VALUES ?bioid {
        "W000059" "W000114" "N000142" "T000086" "J000220" "G000489" "B001051" "W000324" "P000322" "E000136"
        "H000226" "R000378" "B000058" "F000304" "M000889" "B000168" "W000646" "S000905" "F000174" "M000596"
        "M000650" "R000368" "C000656" "F000174" "W000138" "B000960" "F000106" "S000809" "W000609" "J000103"
        "W000629" "C000952" "E000185" "P000530" "A000279" "B000331" "P000557" "E000046" "H000428" "C000343"
        "M000165" "C000357" "J000063" "P000141" "T000324" "R000450" "G000112" "B000386" "S000917" "E000109"
        "P000570" "H000457" "S000038" "M000229" "J000196" "Y000046" "P000238" "J000030" "W000305" "D000495"
        "G000522" "B000360" "W000471" "C000657" "R000075" "F000069" "A000028" "B001196" "W000499" "C001028"
        "G000510" "B000053" "A000126" "M000392" "S001060" "M000993" "W000069" "P000091" "C000597" "E000202"
        "T000225" "L000400" "B000389" "H000612" "H000646" "R000099" "M000787" "B000417" "S000478" "B001061"
        "B000099" "R000443" "W000096" "J000026" "W000248" "H000913" "S001117" "S000792" "B001069" "D000522"
        "L000173" "D000048" "P000575" "S000517" "B000966" "M000236" "F000438" "C001099"
    }
    ?sen p:P39 ?ps;
         wdt:P1157 ?bioid.
    ?ps ps:P39 wd:Q4416090;
        pq:P2937 ?term;
        pq:P582 ?end;
        pq:P580 ?start.
    ?term wdt:P580 ?termStart;
          wdt:P582 ?termEnd;
          (p:P31/pq:P1545) ?termNum.
      BIND(URI(CONCAT("https://bioguideretro.congress.gov/Home/MemberDetails?memIndex=", ?bioid)) AS ?bioURL)
      SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
ORDER BY (xsd:integer(?termNum))

Try it!

Regards, Gettinwikiwidit (talk) 01:25, 23 December 2020 (UTC)

From my POV it doesn't make sense to use not re-elected (Q22087114) unless the subject served until the end of the term, ran for re-election and failed to be re-elected. Are there items for their replacements' elections? If so I would use those as the end cause (P1534). That, I think, maps more cleanly on to what actually happened: their time in office came to an end as a result of the election of their replacement. --M2Ys4U (talk) 01:17, 24 December 2020 (UTC)

@M2Ys4U: Thanks for replying. So would you think that all politicians who serve out their final term should have an end cause (P1534) qualifier? Those who don't serve out their term would seem to more require clarification. I'm happy to use something else, but nothing here seemed appropriate. Is it worth making a new element like "interim period ended"? Gettinwikiwidit (talk) 12:06, 24 December 2020 (UTC)

For fixed-term temporary positions, such as these elected positions, "expiry of term" would also be a perfectly good end cause I think. Ghouston (talk) 23:10, 24 December 2020 (UTC)

Ideally, yes, every entry should have an end cause where known. end of legislative term (Q63323711) is fairly generic. I'd use that when a person serves out their full term, and it would do in a pinch for people appointed for the interim who are then replaced, but I'd defer to other more knowledgable people (say, in Wikiproject Every Politician) on that one. --M2Ys4U (talk) 21:58, 25 December 2020 (UTC)

@M2Ys4U: @Ghouston: My problem with these suggestions is that these are all examples of people being replaced mid-term. Per the Seventeenth Amendment, "In addition, it allows the governor or executive authority of each state, if authorized by that state's legislature, to appoint a senator in the event of a vacancy, until a general election occurs." Also, ideally this field would serve as a distinction between people serving out an elected term, though I suppose the indication of being appointed would make this clear as well. Would anyone object to making an "end of appointment" entity? Gettinwikiwidit (talk) 07:09, 29 December 2020 (UTC)

Sitelinks to closed Wikimedia projects

Turkish Wikinews (tr.wikinews) has been closed earlier this year, but we still have around 8000 sitelinks to this project (query). How do we handle such situations? Do we keep these sitelinks, or shall they be removed? How did we act in similar situations in the past? —MisterSynergy (talk) 19:22, 28 December 2020 (UTC)

• Not sure. I vaguely recall we once deleted some.
  Neither keep them as "sitelinks" nor deleting them seems an optimal solution.
  Maybe we should convert them to values for some url-datatype property? --- Jura 19:31, 28 December 2020 (UTC)

Is the content being retained in situ? The sitelinks still work and it seems unecessary to remove them at the moment. But, converting to a url-datatype property would be a good option for longterm preservation. Simon Cobb (User:Sic19 ; talk page) 21:38, 28 December 2020 (UTC)

In this particular case, the content is apparently not being deleted, but this *can* happen as well per meta:Closing projects policy.
I stumbled upon this via a German Wikipedia bot job that required to load Wikidata items by the bot; the bot was not possible to do certain dewiki edits because is not registered in the closed project and could thus not load the item. While this is strictly speaking not a problem of Wikidata (but rather a pywikibot problem), it indicates which sort of problems may arise when we keep such sitelinks. —MisterSynergy (talk) 21:48, 28 December 2020 (UTC)

So, conversion to a url property does appear to be the right solution and it will also give us a consistent way of handling sitelinks if/when this scenario occurs again. There is a broader issue of perpetual access to the Turkish Wikinews articles to consider - if the content is not available, I would question whether there is any reason to keep the Wikidata items with no other sitelinks? Simon Cobb (User:Sic19 ; talk page) 23:23, 28 December 2020 (UTC)

See Wikidata:Project_chat/Archive/2014/03#Sitelinks_to_closed_wikis.--GZWDer (talk) 10:31, 29 December 2020 (UTC)

Is it preferable that the image of a QID be CC0, or it's just a recommendation?

I noticed that the pictures in house cat (Q146) aren't CC0, and I recall that data in Wikidata is CC0. What about images? Tetizeraz (talk) 10:22, 29 December 2020 (UTC)

Image are linked from Wikidata; they are not part of Wikidata data proper.--GZWDer (talk) 10:27, 29 December 2020 (UTC)

I think GZWDer intended to say that images are linked from Wikimedia Commons. Bovlb (talk) 19:12, 29 December 2020 (UTC)

The main point is that Wikidata doesn't store images but only links to images. The links are structured data and CC0. The image in turn are not hosted on Wikidata and can be of any license that's allowed on WikiCommons. ChristianKl ❪✉❫ 01:14, 30 December 2020 (UTC)

Property proposals for authority control page is full

Wikidata:Property proposal/Authority control is once again too long to display all of its entries. Can we split it in two some way, or otherwise reduce its size? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:38, 29 December 2020 (UTC)

Archiving the proposals that have been open for months without attracting any support would be a good way to reduce the size. For example, Qobuz album ID was proposed on 13 June 2020 and has received no support whatsoever. Simon Cobb (User:Sic19 ; talk page) 21:03, 29 December 2020 (UTC)

Subject has role

I am trying to parse out the roles of people in an accident. I have role=victim for the people that died, but is there a word for someone that causes an accident, the person declared "at fault" by the inquiry? --RAN (talk) 19:43, 25 December 2020 (UTC)

• I don't think this has something today with our "subject has role" property. The nearest thing we have might be perpetrator (P8031) but that might need an expansion in scope. ChristianKl ❪✉❫ 20:11, 25 December 2020 (UTC)

Thanks! That will suffice for now, we can always search for them later and replace them if we create something more appropriate. --RAN (talk) 21:30, 25 December 2020 (UTC)

Using perpetrator (P8031) currently implies that the person engages in a crime, which is slander and when you use it for a person who is a perpetrator of an accident that isn't a crime potentially qualifying under the legal category of slander in many jurisdictions. Slandering people till you find a more appropriate classification is not a good idea. ChristianKl ❪✉❫ 22:07, 25 December 2020 (UTC)

We define it as "person who carried out a harmful, illegal, or immoral act" so with an accident it would imply: "person who carried out a harmful act" which is not slander. It doesn't read "harmful and illegal". The synonyms mostly imply criminal activity, because we mostly have entries on crimes and criminals. Perhaps we should create "at fault" to describe someone who causes an accident. Any other comments? --RAN (talk) 03:42, 30 December 2020 (UTC)

I think it's worth thinking about a description that would be more clear to include accidents and propose a change tot that description on thae talk page of the property while pinging those involved in the property discussion. ChristianKl ❪✉❫ 14:49, 30 December 2020 (UTC)

Two different people conflated under Q982133

The John Lange who is an American author under the pseudonym "John Norman" (born 1931) and the John Lange who is a Venezuelan artist (born 1932) are two separate people, but I don't know how to disentangle them. John Norman (Q982133) was originally set up for the American author... AnonMoos (talk) 11:42, 26 December 2020 (UTC)

@AnonMoos: I like to use the moveClaims tool for this purpose. --Haansn08 (talk) 00:58, 29 December 2020 (UTC)

I currently have limited time at sporadic intervals to edit with advanced browsers that can handle all the page code on Wikidata. If someone doesn't help me to do this the right way, then next time I have unfettered access, I'm just going to remove everything having to do with Venezuela and art, and let somebody else set up an entry for the other guy. AnonMoos (talk) 03:24, 30 December 2020 (UTC)

•  Done just one reference with three data points belonged to Venezuela guy, see now at John Lange (Q104565065). --RAN (talk) 03:34, 30 December 2020 (UTC)

Thanks for helping (I only dabble in Wikidata from time to time, and don't have constant unfettered Internet access right now, so I couldn't really do it myself in the way it should be done). But you didn't transfer over "Museum of Modern Art artist ID 6802", and I think the 50 years copyright stuff may also be referring to Venezuelan laws... AnonMoos (talk) 06:57, 30 December 2020 (UTC)

The history log says that you did remove it, but it's still displaying when I view the page. What a mess... AnonMoos (talk) 07:09, 30 December 2020 (UTC)

Ring of fire

For Pacific Ring of Fire (Q18783), it currently specifies has characteristic (P1552) -> earthquake (Q7944) and volcanic eruption (Q7692360). That seems a little off, since the Ring of Fire isn't a giant earthquake/volcano itself (we hope), but rather just a region with a tendency (Q82785806) for them. Is there a better way to encode this information? {{u|Sdkb}} ^talk 02:28, 30 December 2020 (UTC)

• Seems right to me, instance_of=earthquake would be wrong, but has_quality=earthquake sounds correct. --RAN (talk) 03:00, 30 December 2020 (UTC)
• Maybe we should instead call the quality "earthquake prone"? ChristianKl ❪✉❫ 12:25, 30 December 2020 (UTC)

Duplicates won't merge

Q58823967 and Q58822965 are clearly the same person, but they won't merge because one Wikipedia has 2 almost identical articles for them. Can someone help fix this- I'm not a regular Wikidata user, so not sure if there's a way to "force" the merge. Joseph2302 (talk) 12:02, 30 December 2020 (UTC)

Merged. --Tagishsimon (talk) 12:08, 30 December 2020 (UTC)

• The way you solve the issue is to look at the almost identical items. The issue here is that one of the articles linked to redirect. After the link to the redirect is removed the merge works (that's what Tagishsimon) did here. ChristianKl ❪✉❫ 12:12, 30 December 2020 (UTC)

Well, not quite. There were two distinct articles. I redirected one to the other b/c both had the same text. Where a language or other wiki is blocking a merge, your options are to fix the issue on the wiki, or to ask someone on the wiki to fix the issue for you. With some effort, the community forum can normally be found. --Tagishsimon (talk) 12:19, 30 December 2020 (UTC)

WordLift URL (P6363)

Seems its links no longer work. Shouldn't that property be fixed or removed ? Bouzinac ^{💬●✒️●💛} 16:31, 29 December 2020 (UTC)

I send a message to the people maintaining it to figure out what is going on. Multichill (talk) 10:18, 30 December 2020 (UTC)

@Bouzinac: Most of them are fixed now. Multichill (talk) 15:34, 30 December 2020 (UTC)

Dittographies in (german language) descriptions (for politicians)

Hello, currently in more than 1.100 wikidata objects we have a duplication of the term "politician" in the german language descriptions:

• Suche nach "Politiker und Politiker"

For example:

• d:Q12056052 - "tschechischer Politiker, Politiker, Politiker und Politiker" -> should be "tschechischer Politiker"
• d:Q12022400 - "tschechischer Politiker, Politiker und Politiker" -> should be "tschechischer Politiker"
• d:Q6277995 - "peruanischer Politiker und Politiker" -> should be "peruanischer Politiker"

Is there a generic approach to find and update such errors?

I could update the > 1.100 labels using Quickstatements, for this I would have to extract the affected objects and the labels (e.g. using SPARQL).

With the query

SELECT DISTINCT ?item ?label ?itemDescription
WHERE
{
  SERVICE wikibase:mwapi
  {
    bd:serviceParam wikibase:endpoint "www.wikidata.org";
                    wikibase:api "Generator";
                    mwapi:generator "search";
                    mwapi:gsrsearch "'Politiker und Politiker'"@de;
                    mwapi:language "de";
                    mwapi:gsrlimit "max".
    ?item wikibase:apiOutputItem mwapi:title.
  }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],de". }
  }

Try it!

With this query, I get a lot of false positive matches (which should not be included, but could be filted out in the downloaded list later on), but also false negatives (missing entries, that actually should be included, like d:Q12022400)

On the other hand, just updating these > 1.100 entries might miss other dittographies with other occupations and/or other languages. So is there a more generic approach to identify and correct affected objects?

Thanks a lot! --M2k~dewiki (talk) 15:27, 30 December 2020 (UTC)

For example there are also entries for "Jurist, Politiker, Jurist und Politiker": Jurist, Politiker, Jurist und Politiker

• d:Q6347372 - "norwegischer Politiker, Jurist, Politiker und Jurist"
• d:Q56311215 - "russischer Politiker, Jurist, Politiker, Jurist und Politiker"
• d:Q5943642 - "kolumbianischer Politiker, Jurist, Politiker, Jurist und Politiker"

Thanks! --M2k~dewiki (talk) 15:41, 30 December 2020 (UTC)

• @Hogü-456: it seems some of those labels where added by you. ChristianKl ❪✉❫ 15:44, 30 December 2020 (UTC)

For example, these errors could be repeated in languages like "de-at" and "de-ch" (see this edit). --M2k~dewiki (talk) 16:37, 30 December 2020 (UTC)

I think that there was a mistake in the way how I created the descriptions before uploading. In the query in that example I Filter the existing descriptions in German [13] a occupation is mentioned more than once if there are more results for the other things what I select. So i have to check and correct it. I think that I can correct it. I downloaded the results and then I can correct it for a specific state. The Items where the word "Politiker" is mentioned more than once have I found with the query. I dont know why the output from the QueryService is like it is with my query. I think it shouldnt be that a statement what is not twice in the Item is mentioned twice in the query result. I think it were good if there are possibiltys to publish spreadsheet files within the Wikimediaprojects used for creating output for the upload so that other users can understand how the data preparation was made.--Hogü-456 (talk) 16:40, 30 December 2020 (UTC)

Generic vs. FamilySearch Historical Records

Do we need separate entries for the generic version of a database and the specific one housed at FamilySearch? The database can be found at multiple locations on the internet, FamilySearch, Ancestry, MyHeritage, and Reclaim the Records, and usually the state's own website. We have California death index (Q28018049) with parts California Death Index 1940-1997 - FamilySearch Historical Records (Q94425079) and California Death Index 1905-1939 - FamilySearch Historical Records (Q94425061). It seems that we could have the generic "California Death Index" and just have the two urls to Familysearch appear in that record. The project was broken into two tranches indexed over two time periods by the State of California. --RAN (talk) 19:48, 30 December 2020 (UTC)

Not sure exactly what you're getting at here, but I think third-party formatter URL (P3303) was designed to support this sort of thing - we have a property for an ID value with its default location linked, and then other sites that use the same ID would be in principle linked via P3303. However, the Wikidata UI doesn't have any specific support for this right now so one would have to construct the links by hand. ArthurPSmith (talk) 20:07, 30 December 2020 (UTC)

• Merging the two entries for the very specific proprietary database into the generic version, would give us one record with urls to the proprietary database. I do not think we need multiple entries. For instance we have one record for a book, and can list multiple places that book is available online. We don't need one entry for Google Books and another entry for the Internet Archive. See for example Social Security Death Index (Q2296741) what a merge would look like.. --RAN (talk) 20:51, 30 December 2020 (UTC)

Need help with P6617 RoMEO publisher ID

For the formatter URL and search formatter URL, they have changed their website link but I can't get the new one to work. It doesn't spell out publisher, it shortens it and then it doesn't work. Thanks in advance for your help! Funandtrvl (talk) 20:52, 30 December 2020 (UTC)

Huggle

I see that Huggle can apparently be used on Wikidata, but I found this page, and I'd like to ask, is Addshore the only person that can use Huggle on Wikidata? JJP...MASTER!_{[talk to] JJP... master?} 21:48, 30 December 2020 (UTC)

Are empty disambiguation pages in scope?

If a disambiguation page has no interwiki links in scope or out of scope per wikidata:notability? If the answer depends, if the disambiguation title is of a person name that is highly likely to have terms, is that variety in scope or out of scope? Thanks for opinions.  – The preceding unsigned comment was added by Billinghurst (talk • contribs).

• Yes, as it may prevent them to be linked to wrong pages (e.g. surname vs disambiguation), and help to link to correct disambiguation pages in other wikis.--GZWDer (talk) 09:07, 30 December 2020 (UTC)
• Disambiguations (lists, templates, etc.) are in my opinion purely Wikimedia internal stuff which isn't notable on its own without sitelinks. --Matěj Suchánek (talk) 10:33, 30 December 2020 (UTC)
• Those are usually being deleted occasionally, without prior discussion. Wikidata:Database reports/to delete/empty disambiguation items is a report to get an overview of what we have. —MisterSynergy (talk) 11:01, 30 December 2020 (UTC)
• Such items fail all of 1, 2 and 3 on Wikidata:Notability, in my opinion. Ghouston (talk) 11:02, 30 December 2020 (UTC)
• I'd check the history of the items to see what happened to the previous site link(s). E.g., on Q297613, they were moved to surname items: should a Wikipedia disambiguation page be linked to a surname item if it only contains surnames? Ghouston (talk) 11:07, 30 December 2020 (UTC)
  • The thing that matters is whether a Wikipedia labels a page as disambiguation page or surname page. ChristianKl ❪✉❫ 13:44, 30 December 2020 (UTC)
    • I've been puzzled in the past by "name disambiguation pages", which seem to be both. E.g., sv:Falkman. That's linked to an item Falkman (Q47783631) which is a conflation of a "surname disambiguation page" and an ordinary disambiguation page. Linking the Swedish and German Wikipedia pages seems desirable, regardless of exactly what template they are using, but then why not with en:Falkman too? Ghouston (talk) 21:43, 30 December 2020 (UTC)
      • A key issue is that there are Wiki's that have both a disambiguation page and a ordinary disambiguation page which means we need both kinds of pages. Consistency in the data model should be the prime concern to allow both kinds of items to get their respective pages. If a particular Wiki wants that a page gets seen by Wikidata as a surname page they can easily use the approritate templates to model it that way internally. To the extend that other interwiki links are desireable, redirects can be used for that. ChristianKl ❪✉❫ 22:39, 30 December 2020 (UTC)
        • It means we end up with some empty disambiguation items as people move sitelinks around to improve interwiki linking. I'm not sure that "surname disambiguation page" items need to be created, since presumably a wiki won't have an article like that and a surname article too for the same name: the sitelinks could go on the surname item. Cases like de:Falkman, which contains only people with that surname, I think could also be linked to the surname item, although there's a chance they'd change in future with the addition of new entries and need to be moved. Ghouston (talk) 02:15, 31 December 2020 (UTC)

Thanks. If I think that such a page is unlikely to be used in any short/medium term, then I will look to nominate for deletion.  — billinghurst sDrewth 06:04, 31 December 2020 (UTC)

Elçin

Elçin is a female given name in Turkish and a male given name in Azerbaijani language. (It is also a surname but we have a separate item for that, for the moment without problems, I guess.) At this moment Elçin (Q26884127) has both male and female names inside and it is also a "name: given and family name" item. Of course this is not correct. We need to separate male and female given names and also have a "name: given and family name" item. In other words, this must be divided into three. (If you look at the talk page of this item, you will see that there is a WP item for male given name in AZ:WP.) I do not know if it is due to year-end, but I feel a heavy fatigue to do things again and again. Could someone with buttons and more internet technology than this scribe who writes with two fingers spare some of her/his time to kindly take care of all this? Never mind for the Commons cat. The two items may use the one same Commons cat. I mean the "given name" cat. I just made it a subcat of Category:Turkish-language feminine given names, Category:Azerbaijani-language masculine given names and Category:Unisex names. We use that formula in Commons for given names which in one language are female and in another male first names. There is no reason at all to bring these two together here as a "unisex name". That is valid for the case of, say, Deniz (Q639085), which is a unisex name in the language/culture it belongs (Turkish). I will accept your help in this as a new year gift. Thanks in advance and Happy New Year to everybody. --E4024 (talk) 14:10, 30 December 2020 (UTC)

• Please look at c:Category:Elçin Sangu at Commons, instead of appearing at letter "S" as defaultsort, due to the absurd situation here at Elçin she appears at "E" as if Elçin were her surname... --E4024 (talk) 02:33, 31 December 2020 (UTC)
  • So Commons recognises that there is a given name Elçin, which is commonly assigned to males in Azerbaijani and females in Turkish. I'm not convinced that creating separate items for "male given name Elçin" and "female given name Elçin" is really a good idea, since there's only one name. Instead, the way it's used in different languages should be somehow declared on a single item. The sorting of c:Category:Elçin Sangu I could fix (it was caused by a redirection of an item in Wikidata). Ghouston (talk) 02:58, 31 December 2020 (UTC)
  • This "given name and surname" item Elçin (Q26884127) shouldn't be linked as a given name, though. A separate given name item should be created. Ghouston (talk) 03:02, 31 December 2020 (UTC)
  • The merge of Elchin (Q76033490) with Elçin (Q26884127) doesn't seem like a good idea: I'll revert it. Ghouston (talk) 03:04, 31 December 2020 (UTC)
  • So Elchin (Q76033490) is restored as a given name, but the gender/language issue isn't resolved. Ghouston (talk) 03:14, 31 December 2020 (UTC)
  • I'm also uncertain if "Elchin" should be treated as a different name (per enwiki sitelink), or the same name with transliteration. Well, the enwiki usage is transliteration, but there may be people actually named "Elchin". Ghouston (talk) 03:23, 31 December 2020 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Let me try to answer a few things, of course after thanking you first for your contributions:

• Language: Not an issue, Turkish and Azerbaijani are among the closest two languages in the world. In Turkey we call our neigbour's language frequently as "Azerbaijani Turkish".
• Gender:The difference comes here, our Azeri neighbours use this name for men while we Turks for women. In Turkey I have met at least one male Elçin, but that exception is most probably due to "family roots" or is similar to "male name borne by a female" case (inverse). If you want to make the given name item "unisex" please go ahead. No problem.
• Elchin (sic) is totally innecessary (I mean here and in real life). In Azerbaijan they use the Latin Alphabet, with a few different letters. Elçin is written as Elçin. Whoever changes it to Elchin makes a mistake.
• Please do not leave anything unresolved. I trust you as everybody else here does. Happy New Year! E4024 (talk) 03:33, 31 December 2020 (UTC)
  • Thanks, and happy new year to you too. There are a lot of examples of "Elchin", such as Elchin Asadov (Q16027885), as well as on en:Elchin. Maybe it's a transliteration from Cyrillic. Ghouston (talk) 03:57, 31 December 2020 (UTC)

Importing US census population

The enwiki template w:Template:US Census population contains fields for all censuses since 1790. Of course, not all are filled always, but it’s still a vast amount of data on more than 30,000 pages. A random selection of 500 pages from the last 5000 of this 30,000 shows that currently 404 out of 500 have no population data at all:

SELECT ?item WHERE {
  SERVICE wikibase:mwapi {
    bd:serviceParam wikibase:endpoint "en.wikipedia.org";
                    wikibase:api "Generator";
                    wikibase:limit 500;
                    mwapi:generator "transcludedin";
                    mwapi:gticontinue 6616366;
                    mwapi:titles "Template:US Census population".
    ?item wikibase:apiOutputItem mwapi:item.
  }
  MINUS { ?item wdt:P1082 []. }
}

Try it!

(The later a page is in the list, the more recent it is, so maybe older articles’ items are not that terrible, but still there’s much data awaiting import.)

I’d like to ask bot owners to import this data. Is it okay? Does it comply with copyright laws? Is there probably a better source than enwiki articles? —Tacsipacsi (talk) 21:03, 27 December 2020 (UTC)

• By law data published by the US government is in the public domain. ChristianKl ❪✉❫ 22:05, 27 December 2020 (UTC)
  • Please no more time-series here. It's really not what Wikidata is good for, and makes items unreadable. Jheald (talk) 10:28, 30 December 2020 (UTC)
    • @Jheald: Then where should time-series go? It’s useful data, for example one can create pretty diagrams based on it, which illustrate how the population grew or shrank over the past decades. Also, as I mentioned, many places have no population data at all—I hope that we can agree that importing the 2010 figures to places having no population data is something Wikidata is for. —Tacsipacsi (talk) 01:31, 31 December 2020 (UTC)
      • @Tacsipacsi: maybe stored as Tabular Data on Commons? Simon Cobb (User:Sic19 ; talk page) 19:18, 31 December 2020 (UTC)
        • @Sic19: Is there an established way to model population in tabular data? Is there a way to link to these pages from Wikidata items? (If I want to display the diagram in an article, the most obvious language-independent way to access the data is through Wikidata.) —Tacsipacsi (talk) 20:25, 31 December 2020 (UTC)
          • @Tacsipacsi: we have tabular population (P4179) for the link from Wikidata to Commons. For example, see Taipei (Q1867), which links to Data:Taipei Population.tab. It does seem to offer a viable alternative to storing this data in Wikidata.

There are a few other tabular data properties, which might give you some idea about how it can be utilised, e.g. weather history (P4150), tabular software version (P4669), tabular case data (P8204). Simon Cobb (User:Sic19 ; talk page) 20:59, 31 December 2020 (UTC)

Yad Vashem import

See for instance Alfred Schwerin (Q85644064) where the Yad_Vashem data was imported as a Yad_Vashem_ID as well as a url. Usually when we have an Identifier, we don't also import the url, it is redundant, can we automate the removal of the urls? --RAN (talk) 03:05, 30 December 2020 (UTC)

i mean probably could be automated, but is it worth the effort to remove them? BrokenSegue (talk) 03:21, 30 December 2020 (UTC)

Sorry about that! Regarding the duplicate references: as far as I can tell, I was working with URLs that included a bunch of duplicates that escaped detection due to a (superflous) parameter. Duplicate references are somewhat annoying, but, if I may say so, possibly better than no references–a far more ubiquous situation?

I never figured out what that extra parameter did, when you truncate it, it has no effect. --RAN (talk) 19:39, 30 December 2020 (UTC)

As to the "described at URL" statement, my thinking at the time was that adds a bit of semantic meaning, specifically a more explicit pointer to a authorative outside source. It's a bit silly for items such as the linked one with only one source, but other items have dozens of external identifiers, many of which are rather shallow (VIAF...).

I've soured a bit on the Yad Vachem data because their data is document-centric, of which they often have several for a single person. I'll be importing a bunch of data from Memorial Book Bundesarchiv ID (P7571) at some point and will avoid these issues, and correct existing ones when I touch the items again. --Matthias Winkelmann (talk) 19:09, 30 December 2020 (UTC)

• @Matthias Winkelmann: Amazingly good work by you. The duplication is small and an easy problem to solve. It would be nicer if Yad Vachem and the USHMM merged their duplicate entries. They should have a unique identifier for each person, not a unique identifier for each entry in a list. If they make a decision in the future to merge, they can use us to see where they have duplication. DAHR, a musician database just switched to using a unique identifier and worked in tandem with us. --RAN (talk) 19:39, 30 December 2020 (UTC)
• May I please remark that the Wikidata item lists 1940 as the death date, whereas the source says he was deported to Riga in 1942.--Ymblanter (talk) 21:10, 31 December 2020 (UTC)

Awards: winner works and winner artists categories

I'm trying to use Wikidata to fill categories of award winners in Catalan Wikipedia with bot. It's usually easy to follow the route from winner article to winners'category by using award received (P166) and category for recipients of this award (P2517) (or topic's main category (P910)). However, when awards are about a work of art (like Oscar or Grammy awards) award received (P166) is inconsistently used in the artists or the works of art items and category for recipients of this award (P2517) may point to a category of artists or a category of works.

For example, both Driving Miss Daisy (Q211373) and Jessica Tandy (Q182104) have award received (P166) pointing to Academy Award for Best Actress (Q103618) and Academy Award for Best Actress (Q103618) has category for recipients of this award (P2517) pointing to Category:Best Actress Academy Award winners (Q7044633), which is a category of actresses (not of films) in all Wikipedias. However, Grammy Award for Album of the Year (Q904528), which is also pointed by performers, has category for recipients of this award (P2517) pointing to Category:Grammy Award for Album of the Year (Q6967283), which is a category of albums, not of performers, or Jean-Luc Godard (Q53001) and other directors have award received (P166) pointing to Golden Bear (Q154590) which has category for recipients of this award (P2517) pointing to Category:Golden Bear winners (Q6376423), which is a category of films, not artists.

Then the question is: For awards about a works and performers, how can we tell whether category for recipients of this award (P2517) points to a category of works or a category of people? If both categories exist, how should they be recorded in the award's item?--Pere prlpz (talk) 17:24, 31 December 2020 (UTC)

I don't think it makes any sense to put awards like Academy Award for Best Actress (Q103618), which are obviously awarded to a particular person, on an item like Driving Miss Daisy (Q211373). The work can be named with a for work (P1686) qualifier on the award statement on the person. An award like Academy Award for Best Picture (Q102427) may seem less clear, but it goes to the producers, so can be handled the same way. I've also noticed this with some Hugo Award (Q188914) statements which have been put on works instead of people. Ghouston (talk) 23:47, 31 December 2020 (UTC)

I must admit that it should be easy to query Wikidata to know how many Oscars were awarded to a given film, even if those Oscars were actually awarded to actors, directors, producers and so, an not to the film. Furthermore, some awards are given to two authors for the same works, and if the awards only went to the author's item, it would be hard to tell if the work of art has gotten one or two awards. And there are awards a lot less clear cut than the best actress ones. Who should have award received (P166) of a literary price awarded for a single work? The work or its author?

Therefore, I think that a way to register the award with the work and its author is needed, but we also need to register which category is for works and which one is for authors and we need to relate them. And please notice that category contains (P4224) is not of much help, because some awards can be given to people, to organizations or even to places.--Pere prlpz (talk) 19:30, 1 January 2021 (UTC)

There is the method of indicating the work, the for work (P1686) qualifier: I expect it would be possible to write a query that finds the awards associated with a particular work just using that. Ghouston (talk) 21:41, 1 January 2021 (UTC)

Such queries won't be helpful if you wanted to display the associated awards from Wikidata using a template in Wikipedia though. I don't know if any Wikipedias actually do that. Ghouston (talk) 21:47, 1 January 2021 (UTC)

I agree that an award like “Award for best actress” goes to the actress, and that for work (P1686) is the way to go.

However, I did quite some work on awards (not sure anymore if that included the Hugo Award (Q188914) or not), and yeah for me it was never clear that “Best Novel” should *not* go the novel itself but to the author: language sources would seem to be all over the place on this (googling around − The Guardian for example did headline “Arkady Martine wins Hugo for best novel”, but there’s plenty of “'A Memory Called Empire' by Arkady Martine Wins Hugo Award” − including Hugo themselves). The same goes for “Best picture”, never occurred to me before that it should not go to the movie (but to the producer? not eg the director?).

This raises the question for me of video games − sources (almost?) always say “A Short Hike (Q66035140) won the Independent Games Festival Award - Seumas McNally Grand Prize (Q16218137)” − if not the game then who? The developer, the publisher, the director? If you look at the BAFTA Games Awards, it’s all over the place: sometimes it’s “Game X development team, developer/publisher” (eg “DEATH STRANDING Development Team – Kojima Productions/Sony Interactive Entertainment Europe”), but some other times they single out one or several people − which make sense for “lone developer” (eg Justin Stander for Katana Zero), but less so some other times − such as “Apex Legends”, a superproduction involving hundreds of people, but somehow they list the Executive Producer (ok makes sense) and the “lead scripter” (?)

Jean-Fred (talk) 09:18, 8 January 2021 (UTC)

It seems clear that the winner may be said to be the artist or the work depending on context and award. Therefore, I think we should be allow to store both the work and the artist as winner, and also to store just one of them, and there should be a way to write down whether every category is for artists or works. The question is how to do it.--Pere prlpz (talk) 20:24, 9 January 2021 (UTC)

I guess category for recipients of this award (P2517) should point to a category of the award winners, not a category for the works. Ghouston (talk) 21:52, 1 January 2021 (UTC)