Wikidata

Wikidata:Project chat/Archive/2021/07

< Wikidata:Project chat‎ | Archive‎ | 2021
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

Reasons for preferred rank

When filling out reason for preferred rank (P7452), what is the difference between generally used form (Q71538638) and most frequent value (Q74524855)?

I'm describing a place that has a bunch of different architectural styles (architectural style (P149)), but with two the most predominant that I've ranked as preferred, and I'm not sure which of those reasons to go with, or if editorial choice (Q71535331) or no preferred rank at all would be better. {{u|Sdkb}} talk 05:41, 29 June 2021 (UTC)

@Sdkb: Since generally used form (Q71538638) is a facet of name (Q82799) and is generally used for preferred names or pseudonyms of humans, I'd say this is not applicable here. If two styles are predominant in that place, I'd use most frequent value (Q74524855). —192.198.151.51 18:20, 1 July 2021 (UTC)
Thanks; that's what I've gone with! {{u|Sdkb}} talk 18:52, 1 July 2021 (UTC)

Should states have new entries when they become a constituent state?

Look at the Canadian provinces of Nova Scotia vs Newfoundland and Labrador. There is one entry for Nova Scotia (Q1952), which is tagged as instance of (P31)crown colony (Q1351282) with an end date and instance of (P31)province of Canada (Q11828004) with a start date. Whereas Newfoundland and Labrador has separate entries for its time as Newfoundland Colony (Q2984260), Dominion of Newfoundland (Q38610), and the province of Newfoundland and Labrador (Q2003). Which of these is the right approach? --Arctic.gnome (talk) 15:21, 30 June 2021 (UTC)

I think that the Newfoundland and Labrador is the best approach. Pmt (talk) 19:58, 30 June 2021 (UTC)
@Arctic.gnome: In general if a more than a few statements would have different values under the different circumstances, then two (N) items are preferable (otherwise you have to add start/end dates to them all); in such cases the items should link to one another with appropriate replaced by (P1366)/replaces (P1365) statements. ArthurPSmith (talk) 16:39, 1 July 2021 (UTC)

snow volleyball (Q15728017)

We need a "player" thingy (property?) for snow volleyball (Q15728017). I am afraid I cannot make this. Can someone do the favour? Thx. --E4024 (talk) 16:28, 1 July 2021 (UTC)

Do you mean an equivalent to beach volleyball player (Q17361156)? --Gymnicus (talk) 17:04, 1 July 2021 (UTC)
why can't you make it?BrokenSegue (talk) 17:07, 1 July 2021 (UTC)
snow volleyball player (Q107392205) --Tagishsimon (talk) 17:09, 1 July 2021 (UTC)
Thanks. I said "I cannot make this" 'cause I thought it was not a Q. My bad. --E4024 (talk) 17:39, 1 July 2021 (UTC)

Wikidata list

Hello. I need help with {{Wikidata list}}. See User:Data Gamer/Check.

1) Column "Πληθυσμός (2011)" (means population). Some items have not population for 2011. I want the query not to show anything there. I want the column to list only the population of 2011 for each item, if the population of 2011 exist for that item.

2) In Greek Language, there is stress (Q181767). The problem exists in column "Ονομασία". If the name starts with a letter that takes ' (for example, Ά, Έ, Ή) then that names are first in the list, and the alphabet order in wrong. That is not a problem in my wikidata list but it's a problem when I am using the same list in Greek Wikipedia. w:el:Επαρχία Λάρνακας#Πληθυσμός, έκταση και υψόμετρο ανά δήμο/κοινότητα.

Thanks. Data Gamer play 22:03, 1 July 2021 (UTC)

bullying (harassment)

‎2001:7d0:81da:f780:81ca:5fa0:3839:7a83, 2001:7d0:81da:f780:9130:432c:2be2:221e, 2001:7d0:81da:f780:9130:432c:2be2:221e
Is editing 'IMA status and/or rank (P579)'. I feel being bullied.
I used a tool: 'auxiliary status: published before 1959'. I changing it to 'described by source (P1343)'. My editing capacity is limited. Let's say 250,000 edits in seven years.
He/She cites the IMA list of minerals as 'grandfathered minerals were first described prior to 1959'. But this is a letter of intent, no decision was made for most of the minerals.
Most minerals were grandfathered with the following publication: IMA/CNMNC List of Mineral Names (March 2007)
I do not know if he/she is blocked, as his/her IP number is changing.
Note: Michael Fleischer/USGS had a list of valid minerals, but this is a USGS list and not a IMA/CNMNC list. IMA/CNMNC, Ernest H. Nickel and Ernst A. J. Burke published a list of valid minerals only on March 2007. Regards --Chris.urs-o (talk) 03:24, 1 July 2021 (UTC)
@Chris.urs-o: If you would like other editors to look into this, it would be useful if you can provide links to items where the situation has arisen. Also, if you believe this has developed to the point of bullying/harrassment, you should open a discussion on the Administrator noticeboard. This Project Chat page can help with editing problems and disputes but for a serious claim of bullying should be looked at by an administrator. I'll open a discussion on the administrators board now. From Hill To Shore (talk) 14:31, 1 July 2021 (UTC)

Minerals grandfathered since 2007?

I think lately the main editorial disagreement on IMA statuses has been about start time qualifier under grandfathered status (edits like [1], [2], [3]). Some third opinions on it would be nice and so I'm opening a subtopic about it here. For reference, I've tried to discuss/explain the issue more thoroughly in user talk, among some related issues.

Chris.urs-o seems to suggest that most (or all) grandfathered minerals are grandfathered since 2007 based on publication time of IMA/CNMNC List of Mineral Names (March 2007) (Q20645057). This source however doesn't say anything explicit about statuses of particular grandfathered (G) minerals before 2007, i.e. whether status was/should be considered different before. Also, IMA (International Mineralogical Association (Q268771)) started to apply mineral statuses around its creation decades earlier, and grandfathered mineral is defined in prefaces of IMA mineral lists (latest version) as a mineral that was grandfathered (i.e. generally considered valid) as of the establishment of the IMA in 1958. "Start time: 2007" contradictingly suggest that minerals were grandfathered as of 2007. Earlier Chris.urs-o oneself also referred to another source[1], where it explicitly says that minerals were considered grandfathered before 2007, and per which "grandfathered mineral" isn't some new category coined retrospectively in 2007. So to me it seems rather straightforward that "start time: 2007" qualifiers for grandfathered status should be omitted as unsubstantiated.

There are also some other issues with IMA statuses, in their current form on Wikidata, which I think would also benefit from some wider attention (see Property talk:P579#Review current values). 2001:7D0:81DA:F780:9986:EB9D:4E0A:82F8 09:36, 2 July 2021 (UTC)

  • @Chris.urs-o: It's unclear to me why you speak of bullying. Apart from that "it's a lot of edits" is no good argument given that edits can be automated (and bot approvals be sought if the edit count is very high). ChristianKl ❪✉❫ 13:21, 2 July 2021 (UTC)

Stellar wind emitted by Sun is called solar wind (Q79833)

Me and Ruslik0 were discussing the revert of my edit to solar wind (Q79833). We struggle to find consensus, which I think calls for input from others. What are your thoughts about using qualifiers in this way? —192.198.151.51 20:33, 1 July 2021 (UTC)

Your edit makes perfect sense to me. Whether or not there is a better qualifier is another thing, but if solar wind is specifically stellar wind originating from Sol then that should be modelled in the data like you say. --SilentSpike (talk) 21:36, 1 July 2021 (UTC)
I agree and restored the original edit. ChristianKl ❪✉❫ 13:12, 2 July 2021 (UTC)

Duplicate entry

I hope this is an appropriate place for a merge request.

These two entries are about the same person, the one and only husband of Duchess Elisabeth Sophie of Mecklenburg (Q448131):

  1. Augustus the Younger, Duke of Brunswick-Lüneburg (Q61963)
  2. Auguste II Herzog von Braunschweig-Wolfenbüttel (Q76328400)

The two entries should be merged, or the latter entry simply deleted and only the first one kept.

Sincerely, Moldur (talk) 07:35, 2 July 2021 (UTC)

Ironically, the latter entry is based on a "peerage" site which itself claims as its source - Wikipedia! Unfortunately, that site did not check that it already had an entry for duke August and it thus created a duplicate, and later this duplicate entry was, again without checking, imported into Wikidata, making duchess Elisabeth Sophie look like a bigamist. - Moldur (talk) 08:36, 2 July 2021 (UTC)
We don't delete entries when there are duplicate entries but merge them to allow people who link to either item (maybe from outside of Wikidata) to find the correct one. ChristianKl ❪✉❫ 10:05, 2 July 2021 (UTC)
@Moldur: See Help:Merge, which will tell you how to make this merge yourself. Feel free to ask again here if you get stuck or want a second opinion on whether a merge is the correct answer. From Hill To Shore (talk) 11:24, 2 July 2021 (UTC)
Thank you! I will study that link and see if I can do that merge myself. - Sincerely, Moldur (talk) 13:30, 2 July 2021 (UTC)

Non-linguistic content

What is code for Non-linguistic content like mul and und for "many languages" and "unknown language"? Eurohunter (talk) 13:14, 5 July 2021 (UTC)

@Eurohunter: "zxx" --Bluemask (talk) 15:17, 5 July 2021 (UTC)
@Bluemask: Thanks. It should be described somewhere. Eurohunter (talk) 21:34, 5 July 2021 (UTC)

Wikidata weekly summary #475

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Discussions
    • Open request for adminship:
  • Events
    • Past:
      • Wikipedia Weekly Network - LIVE Wikidata editing #49 - YouTube
    • Upcoming:
      • SPARQL queries live on Twitch in French by Vigneron, July 6 at 18:00 CEST (UTC+2)
  • Tool of the week
    • User:Nikki/AnchorLinks.js is a userscript that adds a small link before property labels and statement values on entity pages to provide a clickable/copiable link to that section of the page.
  • Other Noteworthy Stuff
    • A job opening for a fullstack developer to work on integrating Wikidata with expert-curated knowledge on invasion biology
  • Development
    • Mismatch Finder: We continued the work on building the basic store that in the future will hold mismatches between Wikidata's data and other databases/websites/catalogs/... These mismatches will then be used in the Mismatch Finder website and other tools to easily review them.
    • Improved the namespace behavior of the CommonsLink constraint so that it now produces less false positives (phab:T237920)
    • Working on adding a magic word to allow pages to be excluded from Special:UnconnectedPages on Wikipedias and co (phab:T97577)
    • Language names in Scots (sco) were updated (phab:T285076)
    • A task to complete language names in Dutch (nl) is being filled (phab:T231748)
    • A language code for an undetermined Mixtec language is being determined (phab:T155419)

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.


Item edit request

Please add “Finland” in Wikipedia entries on Finland (Q33), language code is dag. --62.18.11.223 17:44, 3 July 2021 (UTC)

phab:T285919 is blocking this for now. Vahurzpu (talk) 18:57, 3 July 2021 (UTC)
  Done Vahurzpu (talk) 21:59, 6 July 2021 (UTC)

Google Knowledge Graph ID

Google Knowledge Graph ID (P2671). "To find the Google Knowledge Graph ID, open the page source and look after "/g/XXXX" where XXXXX is the ID" - so where is it? I don't see anything like "/g/" in search result. Eurohunter (talk) 14:45, 6 July 2021 (UTC)

it's not always there unfortunately. you can however click the share icon in the infobox, copy the shortened share URL and then go to it. Then in the URL of the browser you will see something like "kgmid=/m/XXXXX" or "kgmid=/g/XXXXX". Yeah it's not easy. We did have a bot going around populating them for us. We probably should have it setup to run periodically. BrokenSegue (talk) 14:57, 6 July 2021 (UTC)
Could you add it to description in some way? What to do if there is no infobox or there is "wrong item" in the infobox? Eurohunter (talk) 16:04, 6 July 2021 (UTC)
honestly it's kinda complicated and the real answer is "google doesn't make it easy and can change how it works whenever". not sure a full explanation would fit into the description reasonably. if there's no infoboxes then probably your query doesn't correspond to a knowledge id. BrokenSegue (talk) 18:17, 6 July 2021 (UTC)
Not necessarily. Google often has Knowledge Graph IDs but doesn’t show infoboxes in every search. Sometimes the Machine-Readable Entity IDs resolver can help, sometimes not. --Emu (talk) 20:18, 6 July 2021 (UTC) 
the real expert here is probably @Lockal: BrokenSegue (talk) 20:40, 6 July 2021 (UTC)
Thanks for ping, @BrokenSegue, Eurohunter:. Google Knowledge Graph ID is documented mostly in various SEO blogs.
Here is an overview of various ways of retrieving GKG/Freebase ids from Google Services:
Not universal, finds only items with high weights (e. g. which have associated data).

Modelling flooded archaeological sites

I'm working on some archaeological sites in the Middle East. Many sites are flooded after a dam is built in a river and its reservoir starts to fill. For example, Tell Fray (Q7697397) was flooded in 1975 when Tabqa Dam (Q372823) was built and Lake Assad (Q979315) was created. I was looking at some way to model this as significant event (P793) but I couldn't really find anything. I found a few cases where flood (Q8068) was used in this way (for example Orléans (Q6548)) but this doesn't seem to be common and in the cases I found this was about a natural flood and nog a permanent flooding by an artificial lake. Does anyone have an idea how this could be represented? Or is there a property for submerged or underwater or something (couldn't find any)? Thanks! Zoeperkoe (talk) 19:45, 7 July 2021 (UTC)

My contribution to the art is at Poolburn Cemetery (Q105696464) - specifically Q105696464#P576 where I've qualified dissolved, abolished or demolished date (P576) with has cause (P828) = inundation (Q301869). It gets a constraint violation flag b/c P828 has an IMO ill thought out 'property scope constraint'. significant event (P793) sounds good, too. There's also located in or next to body of water (P206) which can presumably be used to indicate the lake / reservoir in which the site is now found; that could be qualified with a start date. --Tagishsimon (talk) 21:05, 7 July 2021 (UTC)
Maybe that constraint shouldn't be there. There's also has immediate cause (P1478), which, at the moment, doesn't have any constraints. According to Help:Modeling causes, the cause would be "construction of the Tabqa dam", anyway.
I like dissolved, abolished or demolished date (P576) for the cemetery, since the inundation caused the regular use of the cemetery to cease. For the archaeological sites, I'd use significant event (P793), since the sites are still there and, if the dam were to be removed, they would still be able to be studied. --Azertus (talk) 23:31, 7 July 2021 (UTC)
I hadn't seen inundation (Q301869). That would work well with significant event (P793). Would there be a good way to link it to the dam/lake that would be flooded (I would say that the immediate cause was not the construction of the dam, but the filling of the lake; there are dams that have been (partially) built but where the reservoir was never filled. --Zoeperkoe (talk) 06:46, 8 July 2021 (UTC)
That's what I was trying to say: (underlying) has cause (P828) the construction and immediate cause the flooding. You don't like located in or next to body of water (P206) to link it to the lake? --Azertus (talk) 10:22, 8 July 2021 (UTC)
Ah, yes, I might not have read your answer close enough ;) As for located in or next to body of water (P206), the aliases focus mostly on locations along the shore, so I wasn't sure whether something that's underwater would qualify, but the talk page on the property seems to indicate that underwater can be described by P206 as well. I'll play around with it. Thanks for the ideas! Zoeperkoe (talk) 10:59, 8 July 2021 (UTC)
Yes. <<in ... body of water>> does what it says on the tin: Mariana Trench (Q510) is the posterchild - https://www.wikidata.org/wiki/Q510#P206. --Tagishsimon (talk) 11:19, 8 July 2021 (UTC)

Are chemicals materials?

I was looking at the entry for this ship and it has <designed to carry> = <phosphate 3-ion>, which is in the chemistry type hierarchy, and the web page noted that this was an error because <phosphate 3-ion> isn't a subclass of any of the enumerated things <designed to carry> should refer to. Is this right? Since any compound can be a bulk material shouldn't eg: Q72070508 be a subclass of <material>?

It is an incorrect statement. There should be phosphates (Q46220103) or more precisely: phosphate salt (Q76559025). Wostr (talk) 19:08, 8 July 2021 (UTC)

Single value constraint on LoC and MARC vocabularies ID (P4801) should be revisited

I wrote up an explanation of why on the discussion page for the property, but to summarize: the biggest reason I see is that there is considerable overlap in the places represented by separate MARC Country Codes and MARC List for Geographic Area codes, and there is value in recording them both.

The single value constraint could be lifted, or two separate properties could be created for the above vocabularies. The latter seems clearer to me from a data modeling standpoint.

https://id.loc.gov/vocabulary/countries says about itself: "The MARC country codes are not the same as the ISO 3166 country codes."

If it is preferred that ISO 3166 country codes be used instead of MARC country codes, that should be clarified on the property and some constraint/warning should be added to flag addition of IDs with the URI pattern id.loc.gov/vocabulary/countries

This came to my attention when I added the Geographic Area Code n-us-or to Oregon (Q824), which already had the country code (oru) recorded. --Infomuse (talk) 00:19, 9 July 2021 (UTC)

Wider usage of P789?

Hi! Could edibility (P789) have a wider usage than mushrooms alone? I have placed this taxon is source of (P1672) on a number of instances of taxon (Q16521) to record that the item is a source of an plant as food (Q9323487). P789 would allow a better way to differenciate between items with more or less edibility, but perhaps it is a complex task to use it that way? The property is interconnected with a template used in many wikipedias. -- Abuluntu (👨🏼‍💻💬) 01:17, 4 July 2021 (UTC)

Isn't it too vaguely defined? I mean you can eat many things, even poisonous substances. You can only recommend them for eating if they are nontoxic, which is way better defined. --SCIdude (talk) 14:46, 4 July 2021 (UTC)
Thanks for your opinion. The property is a value of how edible a mushroom is, ranging from not edible (which doesn’t necessarily have to mean toxic) to choice mushroom (a recommendation to eat it if you can find it). It does have problems with it because some mushrooms require a certain treatment to become choice mushrooms. Anyway. I’ll follow up this subject with the approach recommended below. -- Abuluntu (👨🏼‍💻💬) 06:53, 9 July 2021 (UTC)
If you think there should be a different usage, make a proposal what you think the scope of the property should be. Then raise the issue on the talk page of the property, and ping all people that were involved in the property proposal and mention it here.
Given that it does used in template it's useful to understand whether those templates get only used on mushrooms or also elsewhere. It's also worth investigating whether there's an organization that has already standardized edibility of plants, so that we might copy their approach. ChristianKl ❪✉❫ 21:32, 4 July 2021 (UTC)
Thanks for tips on the right approach. -- Abuluntu (👨🏼‍💻💬) 06:53, 9 July 2021 (UTC)

how search various properties of a particular country or domain

i am unable to find data of an artist belonging to iceland. my online search results will be useful if i have little bit of knowledge of website urls.

how do i search various properties of a particular country or domain (***.is)? -Agyaanapan (talk) 06:05, 9 July 2021 (UTC)

Wrap text in table cell

The label for American Sign Language (Q14759) in "ase" is "M528x523S14c02497x497S14c0a472x500S2e85e483x478 M525x535S2e748483x510S10011501x466S2e704510x500S10019476x475 M551x515S1dc50504x485S1dc58474x485S26512449x501S26506536x501".

Is there a way to make this wrap into several lines at Help:Wikimedia_language_codes/lists/all/in_nl#ase ?

I tried adding "<span style="word-wrap: break-word">" at [3], but that didn't quite work out. --- Jura 10:42, 16 July 2021 (UTC)

This section was archived on a request by: see template. --- Jura 10:42, 16 July 2021 (UTC)

Question about notifications

Hi, "newbie question" here. Are there any scripts we can use to notify users when they do something "wrong"? Regards.SirEdimon (talk) 05:04, 10 July 2021 (UTC)

User:Bene*/userwarn.js. But for most honest mistakes, undo with an appropriate message would probably be a better idea. --Emu (talk) 08:39, 11 July 2021 (UTC)

Should multiple objects for one and the same person generally be merged?

This general question came up to my mind when I stumbled upon a person that has multiple objects but which can't be seen as mere aliases cause although both objects describe the same person (same birth date, still living) they still differentiate the persons profession and name. Both names are not the real name of that person and my request is NOT about naming that person's real name but about maybe connecting these two objects or even merging them.

Is there a general rule about it? It might be obvious for example to merge two objects of the same person with different pen names of the same author (using aliases) or a person with different stage names for different professions like a singer who is also an actor. But what for example (my example) if that person is a known singer (object 1) and also a know porn actor (object 2)? Should or shouldn't these objects be connected or even merged because they are describing the same person in public but not the same profession in public.? Weapon X (talk) 18:43, 10 July 2021 (UTC)

The general rule in Wikidata is one item per person. This is different from standards like RDA (generally one record for the real name, one record for each pseudonym). There may be exceptions for privacy reasons as you already pointed out. --Emu (talk) 08:35, 11 July 2021 (UTC)
One item with instance of (P31) = human (Q5). --- Jura 08:42, 11 July 2021 (UTC)

Occupation

I still have doubts about the "occupation" property. When someone is a "stage actor" and "TV actor", I guess we should only use those two, however most users also add "actor". Same thing exists with physicians; for example an ophtalmologist is linked to "ophtalmologist", "physician" and "scientist". In the end an ophtalmologist "is" a physician and a physician "is" a scientist! Maybe because I worked for a long time in Wikimedia Commons, I get to think in Cat-a-lot terms, but still I need help to be convinced either to do what the others do or to be told "hey, you're right"... Thanks for answers. --E4024 (talk) 16:56, 11 July 2021 (UTC)

In general, WD depends on the subclass of (P279) class trees, so for ophthalmologist (Q12013238) https://www.wikidata.org/wiki/Q12013238#P279 which takes us via medical specialist (Q3332438) to physician (Q39631). All things being equal, it's redundant, but not harmful, for the ophtamologist's item to have physician (Q39631) if it already has ophthalmologist (Q12013238). Much the same holds for the actor class tree. Whether or not "most users also add" I would tend to dispute. Not sure if this answers your concern. --Tagishsimon (talk) 18:14, 11 July 2021 (UTC)
There are also cases where people change occupation over the period of their career. They may start out as a physician (Q39631) but then specialise as an ophthalmologist (Q12013238). For a complete record, we would need to include both occupations in the item. From Hill To Shore (talk) 21:21, 11 July 2021 (UTC)
We also need references, so if a source mentions them as a scientist, we can add and should add that. I don't think people regularly take a source calling someone an ophtalmologist and add the occupation of physician, for example. That is something we definitely should not do. --Azertus (talk) 22:07, 11 July 2021 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘

  • Thanks for the opinions, I see that I am the only confused party. :) Speaking about occupations, I have two more questions. i) I have seen that we have two cases for academics, "academic" and "university teacher". Perhaps the second one was intended for, say, an "English teacher" at a Medicine School, but I have no idea. What I observe is that people (like me) get confused and use them in an arbitrary way.

ii) The occupation "diplomat" needs a "grade" or "degree" thingy, just like the academics. (We write academic and then add academic degree "professor", for example.) We should be able to add, as "diplomatic grade" titles like "Ambassador, Consul General, or even Honorary Consul (although I doubt that the last one is part of a professional career, but certainly it is an occupation). E4024 (talk) 00:00, 12 July 2021 (UTC)

Ambassador, Consul General, or even Honorary Consul tend to be encoded as position held (P39) values - e.g. Q15090723#P39. Use of "academic" and "university teacher" (and various others) is a huge mess. --Tagishsimon (talk) 00:29, 12 July 2021 (UTC)

Connecting Vice President of the United States (Q11699) to President of the United States (Q11696)

Anyone have any suggestions as to how to link the Vice President of the United States (Q11699) to President of the United States (Q11696)? The infobox on Wikipedia shows this connection, but I don't see it for the wikidata entry. Gettinwikiwidit (talk) 21:23, 9 July 2021 (UTC)

They're linked via https://www.wikidata.org/wiki/Q11696#P2098 - substitute/deputy/replacement of office/officeholder (P2098) ... there does not seem to be an inverse property (cf. second-in-command to (Q66779118)) ... the first q. is, do we want/need bidirectional links? If so, then afaics a new property would be required. --Tagishsimon (talk) 01:48, 9 July 2021 (UTC)
Sorry if I wasn't clear. I wasn't asking how to link the offices but how to link the people. How can I see that Joe Biden (Q6279)'s vice president is Kamala Harris (Q10853588)? Is the only answer to do date math? This can be complicated by the fact that the same President can have multiple Vice Presidents (as did Abraham Lincoln (Q91)) or even Vice Presidents who served more than one President (as did John C. Calhoun (Q207191)) though the terms invariably are bound by the President's term. Gettinwikiwidit (talk) 21:23, 9 July 2021 (UTC)
Maybe like this? --Azertus (talk) 10:48, 11 July 2021 (UTC)
For future reference, that's Joe Biden (Q6279)significant person (P3342)Kamala Harris (Q10853588)object has role (P3831)Vice President of the United States (Q11699)start time (P580)20 January 2021 --Azertus (talk) 11:53, 11 July 2021 (UTC)
@Azertus: That looks reasonable to me. If there are no objections, I don't mind adding all of these. Incidentally, it's interesting that apparently John Adams (Q11806) was sworn as Vice President of the United States (Q11699) in before George Washington (Q23) was sworn in as President of the United States (Q11696). This also raises an old question about how to distinguish being sworn into office from winning office. The difference in both Wikipedia and Wikidata is often obscured. Gettinwikiwidit (talk) 09:49, 12 July 2021 (UTC)

Şeyhülislam

I guess Sheikhul Islam (Q16138292) and Sheikh-ul-Islam (Q1410729) should be merged, but ZH (Chinese WP) has two articles. Is it a mistake or the articles treat different things? --E4024 (talk) 00:16, 12 July 2021 (UTC)

I don't think so. The Thai Chularajamontri (Q16138292) and 2nd Chinese article refer to Chularajamontri, which seems to be the Thai version of the general concept of Sheikh-ul-Islam (Q1410729). At most they're subclasses, IMO. --Azertus (talk) 08:59, 12 July 2021 (UTC)

Muslim Caliphate

The information regarding the Muslim Caliphate on wikipedia is wrong and misleading. it should be either corrected or deleted. This is person Mirza Masroor is not even a muslim. How can he be called a Caliph.  – The preceding unsigned comment was added by [[User:|?]] ([[User talk:|talk]] • contribs).

Go to EN:WP to ask help, but I doubt you can get much, as I observe anything related to Islam is handled in a peculiar way there (my purely personal opinion only). --E4024 (talk) 14:12, 12 July 2021 (UTC)
This particular issue seems to be no Wikidata or Wikipedia issue but a Google issue and about which Wikipedia page Google shows given specific query. ChristianKl ❪✉❫ 14:23, 12 July 2021 (UTC)

modelling data for authority and multivocality

Hi I have a rather basic question. I want to show that a native label (P1705), a place name, was recorded by Christopher Columbus (Q7322) in Christopher Columbus' journal (Q893385) and also give that statement a conventional book reference. Is there a standard way to model this? MassiveEartha (talk) 11:32, 12 July 2021 (UTC)

You might say <stated in> Christopher Columbus' journal (Q893385) and statement supported by (P3680) for the conventional book reference. - PKM (talk) 21:35, 12 July 2021 (UTC)

Wikidata weekly summary #476

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Discussions
    • Open request for adminship:
      • Emu (RfP scheduled to end after 13 July 2021 11:49 UTC)
      • -akko (RfP scheduled to end after 16 July 2021 06:31 UTC)
    • Closed request for adminship:
  • Events
    • Past:
      • LIBER 2021 - Panel Discussion: Why use Wikidata or not - YouTube
      • SPARQL Wikidata divers. Session of Tuesday, June 29 (in French) - YouTube
      • SPARQL Wikidata Wikisource. Tuesday, July 6 session (in French) - YouTube
    • Upcoming:
      • The next Wikidata+Wikibase office hours will take place on Wednesday 28th July 2021 at 16:00 UTC (18:00 Berlin time) in the Wikidata Telegram group. This month we will also be having a guest presentation about Toolhub by Srishti Sethi from the Wikimedia Foundation.
      • Next Linked Data for Libraries LD4 Wikidata Affinity Group call: Carlo Bianchini (University of Pavia), Stefano Bargioni (Pontifical University Santa Croce (Rome)), and Camillo Carlo Pellizzari di San Girolamo (University of Pisa, Scuola Normale Superiore) will discuss their recent project and article “Beyond VIAF: Wikidata as a Complementary Tool for Authority Control in Libraries”.; [4], July 13th.
      • Upcoming Search Platform Office Hours—July 14th, 2021. Come and ask anything related to Wikimedia search, Wikidata Query Service, Wikimedia Commons Query Service, etc.!
      • Online Wikidata meetup in Swedish #71, July 18
  • Tool of the week
    • Schafe vorm Fenster is a project to build up a calendar for rural villages in Vorpommern-Greifswald using Wikidata to get images and short descriptions for the villages. (See example)
  • Development
    • Mismatch Finder (a tool to review mismatches between Wikidata and other databases): We continued initial development of the store part of the tool. We focused on the upload of new mismatches. (phab:project/view/5422)
    • Discussion on how do describe the types of mismatches reviewed in Mismatch Finder (phab:T285849)
    • We introduced a magic word that can be used to exclude a page from Special:UnconnectedPages on Wikipedia and co. This will make the special page more useful to find pages on Wikipedia and co that should be added to an Item on Wikidata as sitelinks. (phab:T97577).
    • We reduced the number of Wikidata edits that show up in the watchlist and recent changes of Wikipedia and co by not triggering entries for a number of Wikidata edits that do not influence the article (phab:T286193)
    • Property Suggester: We are reviewing patches by a student working on an improved Property Suggester.
    • Working on improving how deprecated statements are handled when checking “type” and “value type” constraints" (phab:T170401)
    • Working on fixing a bug where the entity suggestions are opened when a valid value is already selected (phab:T285102)
    • Wikidata-Wikibase federation (Federation v2): We investigated How to refer to entities that have the same IDs as the source wikibase entities. You can read the ADR here


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.


IFPNI author ID

I request adding an "IFPNI author ID" property to International Fossil Plant Names Index (Q60315739), in the same way as the equivalent IPNI author ID (P586) in International Plant Names Index (Q922063). The IFPNI item already includes an IFPNI species ID (P6341) property for taxon names of species, but none for the authors of those taxon names. –Tommy Kronkvist (talk), 15:38, 6 July 2021 (UTC).

@Tommy Kronkvist: If you need help, feel free to ping me at my Talk page. You can look here for inspiration. Proposing a new property may seem daunting at first, but it's actually not that bad. --Azertus (talk) 00:59, 13 July 2021 (UTC)

UEFA stadium categories

Hello. How can I add that Santiago Bernabéu Stadium (Q164027) is rank as Category 4 in UEFA stadium categories (Q192484)? Data Gamer play 16:56, 10 July 2021 (UTC)

@Data Gamer: For this you would first have to apply for a property for the data object UEFA stadium categories (Q192484), which probably have the same name, and then you create a data object for each of the four categories and then you can add this information using the property and the data objects. --Gymnicus (talk) 13:55, 12 July 2021 (UTC)
@Data Gamer, Gymnicus: I guess they could also create a statement like instance of (P31)UEFA cat. 4 stadium, where the latter is a subclass of association football venue (Q1154710). But a new property is probably the preferred option. --Azertus (talk) 00:55, 13 July 2021 (UTC)

Emperors

Emperor (Q39018) and emperor (Q938153) seem to be a bit mixed up, both WD and WP(s)wise. Making one with "E" and the other with "e" does not help much, either. --E4024 (talk) 00:25, 13 July 2021 (UTC)

The most confusing part is caused by this edit. Reverting. --Azertus (talk) 00:33, 13 July 2021 (UTC)

Checking my work

Hi! Could a more experienced user check my work at Cueva de las Manos (Q223385) and provide feedback if necessary? Thanks! Best, Tyrone Madera (talk) 04:16, 10 July 2021 (UTC)

@Tyrone Madera: I'm not the most experienced user, but I checked your work and didn't see any obvious mistakes or shortcomings. Another user and I changed the instance of (P31) statements, but those weren't added by you. I made sure that cave with prehistoric art (Q11269813) leads to "cave", so I left out the latter. I've also made Cueva de las Manos (Q223385) a model item of the former. Maybe it can become a showcase item? --Azertus (talk) 00:52, 13 July 2021 (UTC)
@Azertus: Thanks, and that would be awesome! How do we get it to be a showcase item? Tyrone Madera (talk) 05:25, 13 July 2021 (UTC)

What Are Some Ways To Get Backlinks From Wikidata?

https://www.youtube.com/watch?v=tlbnXfWebrM

Apologies if this has been discussed before, but I recently ran across this short video where a group of SEO workers (Semantic Mastery) discuss the process of getting Wikidata to link to their clients. I found it interesting to see how our processes are perceived from the outside. Bovlb (talk) 16:14, 11 July 2021 (UTC)

I wonder what are they doing if they claim they keep getting "banned" from Wikidata. And can you blame anyone, especially newcomers, if they can't find Wikidata:List of policies and guidelines, which like so much of the infrastructure of Wikidata, seems to be obscure and arcane rather than prominently displayed and linked to from every single Help and Welcome page? -Animalparty (talk) 17:32, 11 July 2021 (UTC)
From the way they talk, they were probably trying to build a social media profile on Wikidata. If I am interpreting things correctly, one of them also says how he chose not to add links out to his website (possibly as a way to avoid accusations of self promotion). His entry would almost certainly have been deleted as unsourced spam. Depending on the size of his organisation/quality of his website, the weblink may have been sufficient to delay deletion long enough for additional sources to be added. However, I have little sympathy for this type of editor; they quite clearly discuss using paid editors to build Wikipedia and Wikidata entries for marketing purposes. While they may occasionally align with our goals to build Wikidata, they could just as easily pay someone to "clean" an entry and remove unflattering statements. From Hill To Shore (talk) 17:54, 11 July 2021 (UTC)
@Animalparty: I don't know what they're doing to keep getting banned, but I suspect it goes something like this: New user creates one or more items that aren't obviously notable. Items get deleted. User can't find items, or any record of what happened to them, so recreates them. Account gets blocked for spamming. User creates new account and tries again. Swiftly blocked for spamming/block evasion/sock puppetry. Repeat ad naseum. (See User:Bovlb/DUI#Recreation)
@From Hill To Shore: Agreed that the omission of potential identifiers is an odd choice, indicating that they really don't understand our notability policies. Remember that Wikipedias will often delete articles for "sounding promotional". I sometimes think about writing an essay "How to create an item on Wikidata so that it won't get deleted". Bovlb (talk) 21:20, 11 July 2021 (UTC)
You should do that. I've been regularly editing Wikidata for years and I still struggle to find solid documentation of policies, guidelines, best practices, WikiProjects, etc. It still seems like the Wild West, where anything goes until somebody shoots you, or like a bunch of blind monks doing the best they can, with little knowledge of what monks in another cloister are doing. On Wikipedia, nearly every article has a link to a WikiProject where users, especially new users, can discuss how to improve articles, and instructional and policy pages are plentiful, largely interconnected, and highly visible. On Wikidata it seems like the powers that be just don't much care about informing users. As the number of Wikidata items grows to the billions, it may be a lost cause for humans to maintain, and maybe we'd better create some benevolent AI robot overlords to curate the data instead. -Animalparty (talk) 21:33, 11 July 2021 (UTC)
@Animalparty: I put down some thoughts in User:Bovlb/How to create an item on Wikidata so that it won't get deleted. Bovlb (talk) 00:46, 12 July 2021 (UTC)
@Animalparty: If people who come to Wikidata for the purposes of getting backlinks and they end of getting banned our process works very well. Saying that we are able to detect and remove his attempt to put backlinks into Wikidata for his clients within 24 hours is a great compliment. ChristianKl ❪✉❫ 14:16, 12 July 2021 (UTC)
Cool. Doesn't make Wikidata's policies and guidelines any more transparent. -Animalparty (talk) 17:48, 12 July 2021 (UTC)
@Animalparty: It also doesn't clean my kitchen sink. To the extend that good faithed editors have issues editing Wikidata it's worth investigating how the issues arrise and fix them. To the extend that bad faith editors are deterred from Wikidata, there's no problem.  – The preceding unsigned comment was added by ChristianKl (talk • contribs).
I find it interesting that so called SEO workers were unable to figure out the notability policy. Seems like very basic stuff for someone working in that space. It sounds to me like the issue they're having is not engaging with the community/platform, but purely trying to exploit it for profit by spamming (got nothing against SEO workers, but even the most basic attempt to communicate with the community would illuminate to them the reasons for their experience if they really wanted to have a symbiotic relationship with the platform). Thanks for sharing an interesting link. --SilentSpike (talk) 18:43, 12 July 2021 (UTC)
I do think they are roughly aware of the notability policy. Thing is that the SEO community has its meetings and conferences where "Being in Wikidata" apparently is a thing in the past years, with talks and hands-on workshops related to Wikidata. The motivation is that a Wikidata item supposedly somehow boosts your Google ranking, and this is what they are actually after.
In their items, they do address the notability issue and try to make their items "meeting" the requirements with all sorts of tricks—sometimes funny and creative, but often not successful in the sense that we would keep the item if we look at it. However, I think that ultimately most of them are aware that they simply (ab)use this project for their own perceived benefits, rather than contributing useful knowledge here. —MisterSynergy (talk) 20:04, 12 July 2021 (UTC)
The rabbit hole just gets deeper. In this video, they claim that $1000-$2000 will buy anyone a Wikipedia page, but here you can buy a single link on Wikipedia for only $75. This article is actually fairly well-informed (apart from some weirdness about properties), but includes the horrifying advice to work with multiple accounts at once, making sure to keep their edit patterns separate, and faking up users pages to "build the impression that the account is managed by a real human being". Bovlb (talk) 02:18, 13 July 2021 (UTC)
I'm pretty sure the main motivation comes from the goal of obtaining a presence in Google's Knowledge Panel, the infobox like part in Google's web seach results. While (I believe) you cannot ask Google to add one, it appears that having a Wikidata item and a Wikipedia page helps getting one. At least that's how it's understood in the SEO world, apparently. [5] whym (talk) 14:21, 13 July 2021 (UTC)

Dominica vs Dominican Republic

There are 346 items about humans with country of citizenship (P27) set to Dominica (Q784) and place of birth (P19) set to somewhere in Dominican Republic (Q786) or viceversa. See the query: 

SELECT DISTINCT ?item WHERE {
  ?item wdt:P31 wd:Q5.
  VALUES ?countryCitizenship {
    wd:Q784
    wd:Q786
  }
  VALUES ?countryBirth {
    wd:Q784
    wd:Q786
  }
  ?item wdt:P27 ?countryCitizenship.
  {
    { ?item (wdt:P19/wdt:P17) ?countryBirth. }
    UNION
    { ?item wdt:P19 ?countryBirth. }
  }
  FILTER(?countryCitizenship != ?countryBirth)
}
Try it!

Some of them might be correct, but in the absence of references, I assume most are data entry errors. MarioGom (talk) 11:52, 12 July 2021 (UTC)

Yes. Some patterns: https://w.wiki/3d9s --Tagishsimon (talk) 12:38, 12 July 2021 (UTC)
Good luck with some SPARQL which doesn't produce hundreds of thousands of false positives. Messy business, countries & citizenships. --Tagishsimon (talk) 13:08, 12 July 2021 (UTC)
I think checking if the P27 has a reference is a good idea. --- Jura 13:34, 12 July 2021 (UTC)
SELECT DISTINCT ?item 
{ 
  ?item wdt:P27 wd:Q784 .
  ?item wdt:P19/wdt:P17? wd:Q786 .
  FILTER NOT EXISTS { ?item wdt:P27 wd:Q786 }
  ?item wdt:P31 wd:Q5.
  ?item p:P27 ?claim .
  ?claim ps:P27 wd:Q784 .
  OPTIONAL { ?claim prov:wasDerivedFrom ?source . ?source ?p ?v . FILTER NOT EXISTS { ?v wdt:P31 wd:Q10876391 } } 
  FILTER(!bound(?source)) 
}
Try it!

Sample query for a {{Complex constraint}} --- Jura 15:59, 12 July 2021 (UTC)

If the family name sounds spanish-like, the probability of being "Rep. dominicana" is more important than for "Dominica". Bouzinac 💬●✒️●💛 16:16, 12 July 2021 (UTC)

Bey and Beg

Bey and Beg are the same thing, dear colleagues. This is a Turkish word. Leave aside the history, even today urban people in Turkey use "Bey", and rural people say "Beg". Colleagues in a handful of Wikipedias should merge their Bey (Q181217) and Beg (Q3683091) articles to be able to merge them also here. --E4024 (talk) 00:54, 13 July 2021 (UTC)

As you observe, the issue is with language wikis, and you added said to be the same as (P460) - wikidata's only remedy - in January 2021‎; so it's unclear why you have come here to advise us of this instance of a common problem. Tens of thousands of item duplicates exist, which could be merged if language wikis permitted. But most often, they do not. --Tagishsimon (talk) 02:29, 13 July 2021 (UTC)
Why would it be "unclear"? I'm bringing an "issue" to the attention of people. Thanks for your interest. --E4024 (talk) 03:27, 13 July 2021 (UTC)
It's unclear because the people who need to be alerted to this are the language wikipedians of the involved languages, not WD users. There are tens of thousands of instances of this issue in WD. Bringing them to this forum is not likely to effect any change. --Tagishsimon (talk) 12:29, 13 July 2021 (UTC)

Use of P1672

I have started a discussion on the use of this taxon is source of (P1672) with collectable wild foods. Please join the discussion if you have opinions. An example of intended use is Arctium tomentosum (Q854816) -- Abuluntu (👨🏼‍💻💬) 12:04, 13 July 2021 (UTC)

Wikimedia Deutschland to receive a grant from Arcadia to work on minority languages

Hello all,

We would like to give you a heads up on a quite exciting project that will be starting soon. Wikimedia Deutschland just received a grant from the organization Arcadia. The goal of this grant is to make our software more usable by cultures underrepresented in technology, people of the Global South and speakers of minority languages. This project will be defined over the coming months and will run for 3 years, ideally in close collaboration with other Wikimedia organizations.

Our first milestone will be to support the creation of a new development team that will function together with Wikimedia Deutschland to improve parts of our codebase and to develop new functionality. With this experiment, we would like to share and transfer the knowledge of the Wikidata development team to other organizations around the globe.

Here you can find the press release related to this announcement. If you have any questions, feel free to reach out to us. Lea Lacroix (WMDE) (talk) 11:52, 8 July 2021 (UTC)

  • Saying that the new development team will function together with Wikimedia Deutschland sounds like it's a separate entity outside of WMDE. If so, how is it structured?
When I think about the Global South, on of the main aspects is that many people don't have computers and use mobile phone to access the web. Wikidata currently isn't able to be edited via mobile phone, so that seems like an obvious place for development effort.
There are two ways to go about that goal. One way is to work on the existing codebase to make it possible to edit Wikidata via the browser. The other is about developing an app that can edit Wikidata with the main goal of being usable by people in the global south.
The next problem is that of how to build language communities inside of Wikidata for minority languages. If we look for example at Kadai Kopi, currently the project chat for that language doesn't have any posts in Kadai Kopi but a lot of English announcements. That means that if a speaker of Kadai Kopi goes to that page they will unlikely think it's a good place to talk to other people in Kadai Kopi. I'd love to have a small Wikidata community that speaks Kadai Kopi and works together at shared interests. I have no idea what the shared interests of speakers of Kadai Kopi happen to be, but I would expect that there are shared interests were they could work together. Maybe is the solution to have a way that automatically puts the project chat of minority languages on the watchlist of speakers of that language to make it easier for them to find each other but I don't know what the best way happens to be. I think it would be great when the new team thinks about possible solutions to this problem and then presents the solutions for further discussion and feedback before putting them into code (likely a solution would be a mix of code and community effort). ChristianKl ❪✉❫ 15:15, 8 July 2021 (UTC)
Thanks a lot Christian for raising these points. At this point the specifics are very much open to discussion. We want to figure them out with the group we’ll be working with together. The exact structure is yet to be determined and will be influenced by our choice of partner for this project. The ideas you mentioned are all worth looking at as we explore this further with the group and community. Lea Lacroix (WMDE) (talk) 09:09, 9 July 2021 (UTC)
Here's the press release related to the grant. Lea Lacroix (WMDE) (talk) 09:48, 12 July 2021 (UTC)
For Mediawiki there are possibilities to install it on a own Computer without deeper knowledge of programming but with using an external company that offers this service. I think it were great if this will be possible for Wikibase too and as a software package from the Wikimedia Foundation or one of the organisations related to it or from a Community member and not from an external company. A important feature that I wish it would be added are from my point of view forms. I think it should be possible to enter data with a form and that with wider possibilities as at the moment with cradle as an external tool. Wikibase is an interesting Software and I hope that I can install it in the future on my own computer. As far as I understand the press release one goal is making the developing of own features for the minority communities easier. I see that in a wider view and think that if you improve it here then it can help all people who use Wikibase. I support this and hope that Wikifunctions will make it easier. I suggest that you create a page for the topic for what you received support from the Arcadia Fund so that it is possible to discuss it there further with also the possibility to make suggestions about possible new features.--Hogü-456 (talk) 19:42, 13 July 2021 (UTC)

succession of statements and identifiers

Is there a way to change the order of statements (and identifiers) so that similar items can get the same sequence? It would be much easier to compare and see what's missing. --Dutchy45 (talk) 07:17, 13 July 2021 (UTC)

@Dutchy45 Statements are usually ordered according to MediaWiki:Wikibase-SortedProperties; if you’re noticing different statement order on different items, it’s probably due to statements that aren’t listed there yet. Can you give some examples? Lucas Werkmeister (WMDE) (talk) 13:09, 14 July 2021 (UTC)
@Lucas Werkmeister Boy I feel like a fool! You are absolutely right about statements not being listed yet. I jumped the gun and should have checked better. Thank you for enlightening me! --Dutchy45 (talk) 17:14, 14 July 2021 (UTC)

Upcoming project on Cafes & Co-work spaces in Abuja Nigeria

Hey guys, I recently put in for a project on the above subject matter. Kindly check it out here and hit me up if you'd like to participate or help out. Thanks --OtuNwachinemere (talk) 15:55, 14 July 2021 (UTC)

On the Structure of Knowledge - From Universe to the particles

I have this theory that everything (except, arguably, the universe object) is a part of something and so correctly composing the "has part" and "part of" structures is important to indexing knowledge.

I'm thinking that, starting from Universe (Q1) we should be able to follow "has part" properties from Universe down through observable universe, galaxy clusters, milky way, earth, objects, matter, down to particles that make things up.

I've never participated in the wiki community and so I apologize if I don't follow some protocol, but this was an idea I had a few months ago and I haven't seen any changes yet and I am curious about what other people think.

Update (7/11/2021)

I haven't really seen any arguments attacking the idea itself (some possible ad hominems). I've been thinking about how to get the work done a lot as well. It seems like the changes I am proposing will alter potentially every item. That's kind of an intimidating point by itself. If I get to a point where I have a strategy that is worth sharing, I will do so.

Update (7/12/2021)

I am going to look in to creating some kind of visualizer to see the compositional relationships. If I find gaps, Ill consider reporting them. One of the premises in my mind that I am interested in validating against Wikidata is that things like "psychology" are a part of the universe in the sense that they are a part of the mind, which is a part of the universe. Does wikidata share in this interpretation? If not, how do they categorize this data? Because I would argue that their interpretation is invalid if it doesn't conclude something like "Psychology PART OF Universe" in however many transitive degrees of composition.

It takes a considerable amount of hubris to rock up to a database that's had 1,458,427,319 edits since its inception, by tens of thousands of users, and which has 12.8 billion triples, and without so much as bothering to familiarise yourself with the basics of properties used to form predicates, decide you've stumbled across need for a change which, in your own words, "will alter potentially every item" ... and not take that as a cue that, just maybe, there's a flaw in your thinking. There is certainly a very great deal wrong and inconsistent and missing in WDs class trees. I live in hope that you learn enough to make the situation better and not worse.
In other news, it is normal - conventional - to add new content to discussion threads at the bottom of the thread, rather than (if I may say so, self-importantly) at the top of the thread. Just as it is conventional to familiarise yourself with a domain - say, WD, for example - before getting all stampy footed with your straw men - "I would argue that their interpretation is invalid". Anyway; have fun. --Tagishsimon (talk) 01:02, 13 July 2021 (UTC) Seriously though, why would my hubris concern you in anyway? Stay on topic man. If my argument is invalid state specifically where it is. Attack the argument; leave me alone. Also wouldn't this style of editing be considered rude to you? I think it is. Beard.spy.42 (talk) 01:09, 13 July 2021 (UTC)


Please don't go around adding those properties to lots of things until you more fully explain what you intend to do. The "part of" relationship is not always best captured by that property. For example San Francisco (Q62) is part of California (Q99) but shouldn't express it as such. BrokenSegue (talk) 20:22, 9 July 2021 (UTC)
I won't touch anything until I see some support. But if I do, then I will make a bunch of changes (maybe procedurally).
P361 would be the inverse of P527. So I'd desire to express something like "California HAS PART (P527) San Francisco" and so, "San Francisco PART OF (P361) California"
What do you think? Beard.spy.42 (talk) 20:36, 9 July 2021 (UTC)
I think you need to learn a lot more about wikidata; in this instance located in the administrative territorial entity (P131). You won't see any support because it really is a very poor idea. --Tagishsimon (talk) 20:41, 9 July 2021 (UTC)
Yeah but, that's just like your opinion, man.
Why is this a very poor idea? I need something concrete before I put it to sleep. Beard.spy.42 (talk) 20:44, 9 July 2021 (UTC)
Like I see, P131 is a very precise thing. But what I am suggesting does not contradict P131, but works in conjunction with it. Beard.spy.42 (talk) 20:45, 9 July 2021 (UTC)
Well, in this case, because we have other properties than 'part of' which link California with San Francisco and provide contextual information that would be missing from a 'part of' relation. And more generally, because WD has decompositions of the sort you seek, and which do not rely on has part / part of. You'll start to meet some of these as you start to learn about the project. --Tagishsimon (talk) 20:50, 9 July 2021 (UTC)
In my opinion it is about the truth or falseness of the predicate statements that I am making. For example, the predicate "San Francisco is a part of California" is a true predicate and so it belongs in the database. Are there explicit concerns about redundancy somewhere that I am missing?
Also what are these compositional properties that I seek? Because, yes, I do actually seek them. Beard.spy.42 (talk) 21:01, 9 July 2021 (UTC)
@Beard.spy.42: In the case of the California example a part of (P361) statement is redundant since located in the administrative territorial entity (P131) has statement subproperty of (P1647) part of (P361). So the relation is already expressed in a more precise way. However, I don't necessarily disagree with your premise, I think the point being made is just that "part of" is the most general relationship and more specific subproperties should be used where appropriate. --SilentSpike (talk) 22:04, 9 July 2021 (UTC)
The point is that, in the end, true statements like "San Francisco is part of U.S.A." are not in Wikidata because they can be derived by simple rules from statements in Wikidata. There are near infinite true statements---you don't want them all listed, you just list the basics and let logic do the rest. Maybe you want to read about https://en.wikipedia.org/wiki/Knowledge_representation_and_reasoning? --SCIdude (talk) 14:38, 10 July 2021 (UTC)
Thanks for the link, though I am not sure what specific argument you want me to glean from the information. It is generally along the lines of how I am thinking about wikidata.
With regards to "San Francisco is a part of the U.S.A" - this is a transitively true statement, and I agree, we don't want to make our data schizophrenic by over-connecting all of the pieces, but the idea should be to make explicit things that are already explicit - for example the fact that "San Francisco is in California" and "California is in the U.S.A.". We could call these "First Degree" relationships. Something like "San Francisco is in U.S.A." is a 2nd-degree kind of relationship because of transitivity, and so lets say that anything higher than degree 1 shouldn't go within the database. If we accept this, then this would mean removing certain predicates already present in the database, e.g. : Q1 currently has the predicate that "Matter PART OF Universe" - Personally, I'd remove that.
Further, I am not sure we can reasonably derive composition from the other relationships present without complicating our data further. Like how would you show me that "San Francisco PART OF Universe" using your methods? Beard.spy.42 (talk) 20:15, 10 July 2021 (UTC)
There are (at least) 25 paths linking San Francisco with The Universe; which is to say, showing that "San Francisco PART OF Universe".
SELECT *
WHERE 
{
  wd:Q62 wdt:P31/wdt:P279*/wdt:P361* wd:Q1.
  hint:Prior hint:gearing "forward".
}
Try it!
--Tagishsimon (talk) 20:55, 10 July 2021 (UTC)
You're using P361 (the property that I am discussing). If anything this is in favor of what I am proposing (not in opposition).
Also try Q2225 (electron) to Q1 (universe). Beard.spy.42 (talk) 21:01, 10 July 2021 (UTC)
No, Beard.spy.42: you try Q2225 (electron) to Q1 (universe). The report service is here. I'll go back to your "so lets say that anything higher than degree 1 shouldn't go within the database". That's pretty much where WD is, and that's why we don't need your idea of linking everything with part of. --Tagishsimon (talk) 21:07, 10 July 2021 (UTC)
That is definitely not where it is at. See Q2225 to Q1. Beard.spy.42 (talk) 21:10, 10 July 2021 (UTC)
Have you understood the statement above, quote: "a part of (P361) statement is redundant since located in the administrative territorial entity (P131) has statement subproperty of (P1647) part of (P361)"? Because if so, what you are complaining about is that you need to learn how to use the query (or other) tool to get what you want. But we are all in the same boat, you can't expect someone else doing the work for you. --SCIdude (talk) 06:35, 11 July 2021 (UTC)
That is indeed User:Tagishsimon's opinion, "man", but you would do well to heed it. Your time on Wikidata will be much less painful if you moderate your current attitude. I have left a series of introductory links on your talk page. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:40, 11 July 2021 (UTC


@Beard.spy.42: Just a heads up that it's unconventional to edit a new response into a previous point in the discussion. For what it's worth, I don't have any qualms with your premise, but that's because it's not a new idea in the slightest. It's part of the fundamental structure of Wikidata's data modelling - see Help:Basic_membership_properties. The point I want to emphasise that I think you may be overlooking is that the majority of these relationships are already in the data, it's just not immediately obvious because of relational complexity.

I don't want to discourage you from contributing to Wikidata and also believe there has been some miscommunication in this thread (it's best to follow WD:AGF). For what it's worth, these kind of ideas are a great way to test/validate the data model by querying to find unexpected gaps. Overall though I would discourage you from pursuing any kind of mass edit without first seeking to better understand the data that's already there (otherwise your efforts will be wasted). --SilentSpike (talk) 19:11, 12 July 2021 (UTC)

These are true words, thank you for stating them SilentSpike. Beard.spy.42 (talk) 00:26, 13 July 2021 (UTC)

I think this is a fantastic idea. I do think you should be open to the idea that it doesn't have to directly be through part of (P361) and has part(s) (P527). Other properties linking objects should be fine as long as that property is listed as a subproperty of (P1647) part of/has part(s). But yes, every physical object should eventually link back to the universe. Supertrinko (talk) 21:33, 13 July 2021 (UTC)

Thank you for the supporting statement. I am open to the idea of representing composition using other properties, but I am still flipping mental coins about it. It's just that if I can deduce the predicate A PART OF B why not just make it explicit for others who come across the data? And I shrug, and think: the only reason not to is that I am too lazy to do the work and there are some concerns about over connecting the objects which I am not sure is actually a legitimate concern (it might be). Beard.spy.42 (talk) 21:32, 15 July 2021 (UTC)

Docent title (sic)

Is "Docent title" (Q85859833) not the same thing with associate professor (Q9344260)? --E4024 (talk) 15:58, 14 July 2021 (UTC)

Strictly speaking, not, they belong to different systems (the Russian one has four titles, the American one three titles), though they are often considered as equivalent (I denote myself in my CV as an associate professor for a certain period of my life, and this is generally accepted).--Ymblanter (talk) 17:21, 14 July 2021 (UTC)
Well, help me then, Ymblanter, as an academic and experienced Wikidatan. Ms Mahpeyker Yönsel (Q107528407), other than being an artist, is an academic with the title doçent in Turkish, which is in between "professor" and "yardımcı doçent" (an "assistant with a doctorate") and we use it as "associate professor" in Turkish universities, like METU, that make education in English. I cannot add her academic grade, rank or whatever it is called to her item without receiving from the system an exclamation mark in response! Will you please give 10 seconds of your life to this and do it for me? Thank you very much in advance. --E4024 (talk) 19:15, 14 July 2021 (UTC)
I think you need Q462390.--Ymblanter (talk) 19:35, 14 July 2021 (UTC)
[6] Could you please add a reference?--Ymblanter (talk) 19:37, 14 July 2021 (UTC)
Thanks, I added, but I still do not understand the relations between "academic rank", "degree", do we also have "grade"? Why do we have "professor" and "full professor"? How can we make a distinction, I mean which one to use, if in a country you do not have such a difference... Whatever, at another section above, titled "Occupation", another admin has rightly used the word "mess" for this area... Thanx again. --E4024 (talk) 19:50, 14 July 2021 (UTC)
Mostly because many countries apply systems with different ranks from each others, and we also need to account for ranks no longer in use. The "full professor" name is really not the best, as the item actually correspond mostly to the Spanish/Latin American title of Catédratico, which is above that of what would be referred in English as a regular Professor (if we had to, we might draw a parallel with tenured vs. untenured professors). Certain university posts (e.g. faculty rectors) may require that their holder be catédraticos, not merely titular professors (). It doesn't help that in this particular cases, several wikilinks have since been redirected.Circeus (talk) 18:04, 15 July 2021 (UTC)

Merge request

Please merge 2021–22 Serie A (Q104599603) and 2021–22 Serie A (Q107538887). Thanks in advance!!! --2001:B07:6442:8903:C9B1:F8AD:FF5D:AC26 14:12, 15 July 2021 (UTC)

done! Circeus (talk) 18:05, 15 July 2021 (UTC)

Data donation? - Name authority data from the Schoenberg Database of Manuscripts

Hello Wikidata community,

My name is L.P. Coladangelo, and I am a doctoral student with the College of Communication and Information at Kent State University. I am currently working as a fellow of the LEADING project, a fellowship focused on data science projects through Drexel University. My host site is the Schoenberg Institute for Manuscript Studies at University of Pennsylvania. The project involves the Schoenberg Database of Manuscripts (SDBM), a dataset focused on mostly provenance related data about premodern manuscripts.

One of the goals of the fellowship project is to increase exposure/reuse of the SDBM data. We are hoping to contribute the structured/linked data in the SDBM regarding names and places to Wikidata, in order to enrich both linked datasets. This would likely involve a data donation of the relevant URIs from the SDBM for their name authorities as well as proposal (and creation/approval) of a identifier property for SDBM ID (similar to an identifier property for other authorities, e.g., LC authority ID, VIAF ID, or ULAN ID). That way, we would be able to donate the URIs in the SDBM as IDs and to add other linked data to Wikidata entities (using existing properties, where relevant) to link the two datasets.

I would welcome any and all feedback from the community, especially those interested in manuscripts and provenance data, to suggest the best way to go about carrying out this project as well as any assistance that can be provided.

Thank you in advance!

-L.P.

131.123.18.142 15:01, 15 July 2021 (UTC)

Very interesting proposal, I would surely like to cooperate as far as my knowledge and my free time allow! First step is obviously the IDs, I have just proposed them: Wikidata:Property proposal/SDBM IDs. See you soon, --Epìdosis 19:49, 15 July 2021 (UTC)
Hi there, appreciate the help, I thought I already has a proposal in the pipeline Wikidata:Property proposal/Schoenberg Database of Manuscripts ID, but I appreciate that you made that happen. I'm new to Wikidata, so maybe whatever I created didn't post yet or I didn't do it correctly? LPC47 (talk) 20:07, 15 July 2021 (UTC)
Hi @LPC47:, I confess I didn't notice your proposal. In fact it was correct, except for one aspect: a property cannot have contemporarily two formatter URLs, each working for a part of the IDs; in these cases two separate IDs are necessary. Let's now wait for further feedback in the proposal. If you have any doubts, you can ask me! See you soon, --Epìdosis 20:56, 15 July 2021 (UTC)
@Epìdosis:, I genuinely appreciate you catching that difference in the URIs, and therefore correctly proposing two properties. I think with the very strong proposals you created (much appreciated!) and some of the tweaks I made, they look excellent. I think the folks at the Schoenberg will be very happy to add their data once everything is in place. Thank you again! LPC47 (talk) 21:01, 15 July 2021 (UTC)

P373 / Commons category links

Hi all. Wikidata:Properties_for_deletion#Commons_category_(P373) has been running since 2019, it would be really good if it could be closed one way or the other. I've explained many times why I think this data duplication is bad - and why we should just be using the sitelinks - but if we really must keep this property then I can work on more bot scripts to maintain it (i.e., more copy-pasting from the sitelinks...). I asked at the admin noticeboard before, so asking here now in the hope it gets a wider audience / to see if other people have new arguments... Thanks. Mike Peel (talk) 13:05, 9 July 2021 (UTC)

Who can close the proposal? Perhaps we should tag some of the active admins and ask them to close the proposal, especially those who haven't taken sides in the discussion yet. Ideally, they should go through the "Keep" votes and see if all were properly answered/disproved. Personally I only see relevant arguments FOR deletion, but I don't dare to close the discussion because I've already voted (also I'm not an admin).Vojtěch Dostál (talk) 09:07, 11 July 2021 (UTC)
Pinging @MisterSynergy, Mahir256, HakanIST, Lymantria, Hasley: and @Minorax, Ymblanter, Bovlb: as they don't seem to have been involved in the discussions. Also open to a non-admin close, but agree that it shouldn't be closed by someone that participated. Thanks. Mike Peel (talk) 18:36, 11 July 2021 (UTC)
Anyone? Thanks. Mike Peel (talk) 08:41, 16 July 2021 (UTC)
It is not the only property deletion debate from 2019 that is still open on that page, and requiring closure. Indeed, there are some there from 2018. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:47, 11 July 2021 (UTC)

what is the right way to set preferred rank for official website (P856)

i am trying to update new website location on Aggelika Korovessi (Q4692185). archive website says error 301 since feb 2020. how do i update preferred rank for new location or website? -Agyaanapan (talk) 12:40, 16 July 2021 (UTC)

It's explained at Help:Ranking#How_to_apply_ranks. --- Jura 13:28, 16 July 2021 (UTC)
hello Jura, that was super easy. thank you -Agyaanapan (talk) 16:23, 16 July 2021 (UTC)

Neuerscheinung: what is it god for?

Can someone with German knowledge confirm whether my edit of new release (Q1286737) based off the de:wiki article is accurate? Circeus (talk) 18:20, 15 July 2021 (UTC)

This change is wrong, Circeus. --Succu (talk) 19:27, 15 July 2021 (UTC)
What did I misread in the German article? Just telling me it's wrong is not going to be very useful given my terrible gTranslate-assisted German. Circeus (talk) 20:59, 15 July 2021 (UTC)
@Circeus: I made some correction in the description. "Neuerscheinung" is a term from the publishing sector: at a certain (recurring) date a publisher presents all his latest products (e.g. books). These are called "Neuerscheinung". It does not refer to the latest version of a work/calendar/magazine/product. I hope I could clarify the meaning (I'm not sure about "latest release" - my German dictionary suggests "new release" as a translation of "Neuerscheinung") - Valentina.Anitnelav (talk) 21:46, 15 July 2021 (UTC)
@succu, Valentina.Anitnelav: There is currently no German article associated with version, edition or translation (Q3331189), which corresponds to en:Edition (book) (de:erstausgabe is the one that comes closest). Do you think this article might fill that gap? Circeus (talk) 17:29, 16 July 2021 (UTC)
@Circeus: No, I don't think so. new release (Q1286737) is not that explicitly linked to the idea of a certain edition of a work. The distinction between work and edition does not play a role for that term. It is a term from advertising, not from bibliography. - Valentina.Anitnelav (talk) 18:21, 16 July 2021 (UTC)
@Circeus:: I've made some changes too. Hope they fit better in. --Succu (talk) 21:03, 16 July 2021 (UTC)
@Succu: Thanks for your additions. GND 4485351-8 offers the following translations or quasi-synonym (Q2122467): new release, lançamento, novità editoriale, publication nouvelle, lanzamiento, novelty, новинка. --Kolja21 (talk) 21:32, 16 July 2021 (UTC)
Thank you Succu. I'm glad that removing the awkward connection to musical release (Q2031291) that used to be there was correct. Circeus (talk) 22:36, 16 July 2021 (UTC)

poulet (Q213984) link to Spanish Wikipedia

Hi. I was on w:Macaw and clicked on the side bar where there are links from Wikidata to other Wikimedia sites to get to the Spanish Wikipedia article. It's not a big problem but something seems to be up with the link on Wikidata's poulet (Q213984) to Spanish Wikipedia -- the item says the link is to w:es:Ara (animal) but the sidebar takes the user to w:es:Guacamayo (ave) which the article was moved from a while ago and redirects to the proper page. I'm curious what's going on here -- why does Wikidata say one thing but the sidebar takes the user somewhere else? The Spanish Wikipedia page's sidebar takes the user to w:Ara (bird)???

Idk this is weird someone else might get it thanks DemonDays64 | Talk to me 22:54, 15 July 2021 (UTC) (please ping on reply)

@DemonDays64: The article en:Macaw has an old style sitelink to es:Guacamayo (ave). This means that the link between these articles is created in English Wikipedia, not in Wikidata. Also, en:Macaw is linked to macaw (Q219035), not to poulet (Q213984). --Shinnin (talk) 23:55, 16 July 2021 (UTC)

Duplicate Vladimir Barykin

Q107573685 = Q21635886 Vladimir Barykin (talk) 11:16, 19 July 2021 (UTC)

Merged them into the older one (Q21635886). You can now edit this one and correct the VIAF and/or other things. Mirer (talk) 11:55, 19 July 2021 (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) 13:44, 22 July 2021 (UTC)

Two "housing estate"

Should we delete one and merge them? Rtnf (talk) 10:12, 16 July 2021 (UTC)

There is clearly a need to better qualify group of houses (Q2282602), which has a separate article from housing estate (Q12104567) in german and polish only (I can't tell whether basque and japanese are associated with tyhe correct item). Polish and German wikipedias clearly make some sort of distinction, but the item names and description (especially in English) completely fail to properly explain that distinction. Circeus (talk) 17:35, 16 July 2021 (UTC)
So, there is a difference between Großwohnsiedlung/Wielki zespół mieszkaniowy and Siedlung/osiedle mieszkaniowe , but somehow they are getting translated as "housing estate" in English. I understand now. Thanks Rtnf (talk) 08:12, 17 July 2021 (UTC)
It's possible that there is also a difference between the German and Polish terms and not all on Q2282602 are equivalent. --- Jura 08:39, 17 July 2021 (UTC)
The article in Japanese Q2282602 is about housing estate in the Weimar Republic. Afaz (talk) 12:38, 17 July 2021 (UTC)

language in script

There are few items that are instances of language in script (Q64362969). A list is now at Help:Wikimedia language codes/lists/language in script.

These are to describe the language when used with a given script: e.g. "aeb-Arab" is Tunisian Arabic in Arabic script, while "aeb-Latn" is for Tunisian Arabic in Latin script.

The script subtag is described at w:IETF language tag.

It seems that @Pintoch: created a few of them, later some were merged with items about the alphabet used in that language, e.g. "Uyghur Latin alphabet" (Q986283) and "Uyghur in Latin script" (Q22338145).

I don't think the two should be merged, otherwise language in script (Q64362969) and alphabet (Q9779) end up on the same item. It's likely that (some) Wikipedia has just one article on the both concepts.

@Liuxinyu970226, Infovarius: you might want to comment as well. --- Jura 11:58, 17 July 2021 (UTC)

Merging name items

I believe merging a "female given name" and a "unisex given name" item, a recent example is here is not a good idea. Please see Talk:Q47860848,

I would even prefer deleting all three "female given name", "male given name" and "unisex given name" properties and use only "given name" than mixing up things. For example, Elçin may be a unisex name in Azerbaijani Turkish, but in Turkey/Turkish it is only a female name; unless carried by some male who has Azerbaijani roots... --E4024 (talk) 02:32, 12 July 2021 (UTC)

There are some discussions on the topic at Wikidata_talk:WikiProject_Names. Ghouston (talk) 07:51, 12 July 2021 (UTC)
Having different items is particularly useful. I have used them before to automatically detect incoherent relations that surface data entry errors, where people enter wrong sex property by incorrect inference from name. If these are merged, it would be important to properly represent in statements and qualifiers which languages they are commonly used as female, male or unisex. MarioGom (talk) 11:58, 12 July 2021 (UTC)
I find Wikidata's information about "female given name" and "male given name" generally useful when I encouter a name in the wild and want to look up whether it's associated with a gender. ChristianKl ❪✉❫ 20:53, 17 July 2021 (UTC)

Shouldn't Q5460604 and Q43375360 be the same?

Q5460604 and Q43375360. Why are there 2 'Wikipedia:Vital articles' and EN:WP:VA is in the one with few other languages? — XComhghall (talk) 08:34, 17 July 2021 (UTC)

frwiki has different links on each item, so they can't be merged. From Hill To Shore (talk) 08:44, 17 July 2021 (UTC)
What's actually going on is that on the one hand there's meta:List of articles every Wikipedia should have (a list of 1000 articles that is intended to apply to all wikis) and on the other end there is en:Wikipedia:Vital articles (lists of 10, 100, 1000, 10,000 and 50,000 articles that should aimed to be featured in that specific language).
Vital articles are technically project specific (e.g. en:Wikipedia:Vital articles: "This list is tailored to the English-language Wikipedia."). Many Wikipedia, especially of smaller projects, have used the Meta 1000-articles list and called it "vital articles" without tailoring it or expanding it (or if they do, it's hard to tell), hence the existence of both cases in French, Italian and Hebrew, where the meta page has be translated as a separate item. The Cherokee page should be on the other item, and I'm not clear what's going on with the Afrikaans item. Circeus (talk) 23:05, 17 July 2021 (UTC)

canonical list of "Xwiki pages with Wikidata item"?

hi! sometimes, i need a list of all Xwiki articles or pages with Wikidata item to do further filtering etc. for testing purposes, i ran these queries on replicas:

wbc_entity_usage (on client)
SELECT distinct eu_entity_id
from page p
join wbc_entity_usage wb on p.page_id=wb.eu_page_id
where eu_aspect='S'
page_props (on client)
select pp_value
from page_props
where pp_propname='wikibase_item'
wb_items_per_site (wikidata)
SELECT ips_item_id
FROM wb_items_per_site
where ips_site_id='XXwiki'

all these queries did return different result sets with various issues: redirects to merged WD items, page on wiki was deleted, maybe something more. some links were simply missing (and sitelinks were added few months ago). already understood problems with wbc_entity_usage, that the results are bad because of item usage with Lua.

i assume, the problem is known, but what could be done? do i really need to get links from wikidata json dump? that should have better results, right? :) "use api", you say? not for my usecases, i would say. and that would be more slow and probably put more strain on servers. or is there some other source? --Edgars2007 (talk) 13:13, 17 July 2021 (UTC)

  • How many pages are you looking for? If it's lvwiki, I suppose it would be "only" ca. 100,000. This seems to be not entirely impossible with Query Service.
SELECT * { [ ] schema:about ?item ; schema:isPartOf <https://lv.wikipedia.org/> ; schema:name ?page } LIMIT 100000
Try it!
If in addition you filter in one way or the other, it could be doable. --- Jura 17:23, 17 July 2021 (UTC)
ah, right, sparql. yeah, that would also work in some cases, but doesn't it have the same data problem as all others mentioned? :) --Edgars2007 (talk) 10:25, 18 July 2021 (UTC)
I'm not sure how it compares to "wb_items_per_site (wikidata)" (equal?), but I think it should be better than the other two. --- Jura 10:30, 18 July 2021 (UTC)

Fun with Mismatches: typology

 


For the Mismatch Finder, above a sample screenshot of differences that can be reviewed.

An external entity is compared to a Wikidata item by some key (generally an external identifier). The screen presents Wikidata values and ExternalValues that differ.

  • Differences happen when one is incorrect, one is more precise or both are equally "true".
  • It can also be that the key is applied to the incorrect item or external entity or that these conflate several concepts.

In phab:T285849#7198666, I tried to summarize how these can currently occur.

Here are items for each case:

  1. ExternalValue incorrect (Q107563872): add ExternalValue with deprecated rank
  2. Wikidata value incorrect (Q107563862): set Wikidata value to deprecated rank, add ExternalValue
  3. ExternalValue and Wikidata value equal (Q107564007): add ExternalValue with same rank as Wikidata value
  4. prefer Wikidata value (Q107563853): set Wikidata value preferred, add ExternalValue with normal rank
  5. prefer ExternalValue (Q107563856): add ExternalValue value with preferred rank, set Wikidata to normal rank
  6. key error (Q107563942): key is applied to the wrong Wikidata item or External Entity
  7. key error on Wikidata (Q107564015): key is present on the wrong Wikidata item
  8. ExternalValue entity is conflation (Q107563892): don't add ExternalValue to Wikidata
  9. Wikidata item is conflation (Q107563887): add P31=conflation to Wikidata item, don't add External value
  10. it probably needs an additional item for "other".

Maybe there are other cases that should be considered, WDYT?

See Wikidata:Database_reports/identical_birth_and_death_dates/1 for another summary of mismatches (or matches). Help:Conflation_of_two_people about conflations in that field. About ranks, see Help:Ranking.--- Jura 10:13, 18 July 2021 (UTC)

Commons categories

Asking here in light of no replies on the norwegian chat, I'm assuming the question I posted was too special-case for anyone to make heads or tails of.

Basically I wanted to know what was the proper way to link up Commons categories and Wikidata, and especially so when it comes to ships where you have two categories, one for IMO number and one for Ship name+year built. One of the categories is linked by interwiki, but there is also Commons category (P373) and category for ship name (P7782). I suspect there should be two wikidata entities, one instance of commons category and the other one subclass of ship. It ends up being a lot of cross-linking, and it is not at all obvious how exactly it should be done.

I often come across entities where a commons category have been wrongly turned into a ship thereby causing a conflict since there should not be two entities using the same IMO number. It would be nice to know how to properly fix such a case, since at present I'm not really able to, and I don't feel like winging it. Thanks. --Infrastruktur wdt:P31 wd:Q5 (T | C) 17:27, 18 July 2021 (UTC)

@Infrastruktur: for an example of how it works, see Astrid (Q750516) and Category:Astrid (ship, 1918) (Q83564910), the main item uses the IMO Commons category sitelink, then category for ship name (P7782) links to the ship name category item. The Commons category links are in the sitelinks (Commons category (P373) doesn't do much here), this also powers the infoboxes in the linked Commons categories. I have a bot task that will auto-create items for IMOs and ship names at the start of every month, so mostly it should just needing cleaning up in cases where things have gotten into a mess. Thanks. Mike Peel (talk) 19:05, 18 July 2021 (UTC)

Item validation criteria

Hello folks,

TL;DR: what do you think of the 3 validation criteria below?

I'm excited to let you know that the soweego 2 project has just started: m:Grants:Project/Hjfocs/soweego_2. To cut a long story short, soweego links Wikidata to large third-party catalogs.

The next step will be all about synchronization of Wikidata to a given target catalog through a set of validation criteria. Let me paste below some key parts of the project proposal:

  1. existence: whether a target identifier found in a given Wikidata item is still available in the target catalog;
  2. links: to what extent all URLs available in a Wikidata item overlap with those in the corresponding target catalog entry;
  3. metadata: to what extent relevant statements available in a Wikidata item overlap with those in the corresponding target catalog entry.

These criteria would respectively trigger a set of actions. As a toy example:

  1. Elvis Presley (Q303) has a MusicBrainz identifier 01809552, which does not exist in MusicBrainz anymore.
    Action = mark the identifier statement with a deprecated rank;
  2. Elvis Presley (Q303) has 7 URLs, MusicBrainz 01809552 has 8 URLs, and 3 overlap.
    Action = add 5 URLs from MusicBrainz to Elvis Presley (Q303) and submit 4 URLs from Wikidata to the MusicBrainz community;
  3. Wikidata states that Elvis Presley (Q303) was born on January 8, 1935 in Tupelo, while MusicBrainz states that 01809552 was born in 1934 in Memphis.
    Action = add 2 referenced statements with MusicBrainz values to Elvis Presley (Q303) and notify 2 Wikidata values to the MusicBrainz community.

In case of either full or no overlap in criteria 2 and 3, the Wikidata identifier statement should be marked with a preferred or a deprecated rank respectively.

Please note that Soweego_bot already has an approved task for criterion 2, together with a set of test edits. In addition, we performed (then reverted) a set of test edits for criterion 1.

I'm glad to hear any thoughts about the validation criteria, keeping in mind that the more generic the better.

Stay tuned for more rock'n'roll! With love --Hjfocs (talk) 15:46, 6 July 2021 (UTC)

I don't think deprecating an identifier that is no longer valid is proper. Instead mark it with an end time (P582). The identifier was presumably right at some point. BrokenSegue (talk) 18:19, 6 July 2021 (UTC)
What do you mean by URLs (criterion 2)? Say MusicBrainz links to https://snaccooperative.org/ark:/99166/w6cp7v21 (it actually links to https://snaccooperative.org/ark:/99166/w6fb512h, which is a soft redirect to the former). Wikidata has Elvis Presley (Q303)SNAC ARK ID (P3430)w6cp7v21. Does that mean your bot will add Elvis Presley (Q303)described at URL (P973)https://snaccooperative.org/ark:/99166/w6fb512h (even though property SNAC ARK ID (P3430) already provides that link)? What would it do in the actual case (where the identifiers don't match)? --Azertus (talk) 19:34, 6 July 2021 (UTC)
Hello @Jura1: here's an initial test edit, where a YouTube URL available in MusicBrainz (at the time of the edit) was converted into a proper Wikidata identifier. For those URLs where soweego couldn't find a valid formatter URL (P1630), the bot adds a (described at URL (P973), URL) claim, see this edit.
With respect to the reference node, I'm actually thinking of a richer one: I'd add (based on heuristic (P887), artificial intelligence (Q11660)) as per a past discussion of ours, plus the target catalog identifier where the claim is stated, like (MusicBrainz artist ID (P434), 06d3fe7f-b5c7-4af9-a96b-5a41675e29ce).
I made a couple of test edits to illustrate everything: please have a look at Q13406268#P244 and Q13406268#P973.
Do you think everything looks fine? Thanks a lot for your feedback! --Hjfocs (talk) 13:54, 12 July 2021 (UTC)
@Hjfocs: It does not look fine to me. Picking up one of your test edits - this edit - from earlier, why are we adding some random person's blog ("Hello, I am DjPault. This blog is dedicated to Olivia Newton-John. I have been hopelessly devoted to Olivia ever since my aunt took me to see "Grease" in 1978 when I was eight years old.") to a WD item? Sorry to say that all that I've seen on this thread leads me to believe that your software is most likely to append huge quantities of low quality, wrong, and completely inappropriate values to WD items. It seems to be miles removed from something that should be used in production. It is, in short, not a very clever bit of so-called AI, but rather some sort of GIGO machine. --Tagishsimon (talk) 14:04, 12 July 2021 (UTC)
@Tagishsimon: I believe you are barking at the wrong tree: soweego processes data available in what are supposed to be trustworthy external catalogs. Regarding the example you mention, maybe you can reach out to the MusicBrainz community and ask why they maintain such links.
Said that, if you are aware of any URL blacklist (or whitelist, viceversa) that can be used to automatically filter an URL like that, I'm happy to inject them into soweego. Or maybe you can start building one and look for consensus, instead of complaining here. Thanks for your comment! --Hjfocs (talk) 14:29, 12 July 2021 (UTC)
Why is this being discussed here rather than at bot approval? Also, if musicbrainz's data (or any external catalog) is too unreliable then we shouldn't sync with it. Determining if it's reliable enough is...hard and subject to debate. I think the burden of showing it is reliable should fall on the operator. That said, it doesn't mean that the bot is technically doing the wrong thing (GIGO) and nor is there a burden on the commentators to find a URL blacklist or anything to improve the bot. BrokenSegue (talk) 14:38, 12 July 2021 (UTC)
By the way, I remember to have used this Wikidata:Primary_sources_tool/URL_blacklist in the past, and am more than happy to use it again: it would help to filter those blogspot URLs for instance. --Hjfocs (talk) 14:45, 12 July 2021 (UTC)
@Hjfocs: "I believe you are barking at the wrong tree: soweego processes data available in what are supposed to be trustworthy external catalogs" is offensive nonsense. You are purposing to import junk data into WD on the basis that you think the content of an external source is 'trustworthy', for reasons you do not elucidate. The external source, or, at least, the content you have used as an exemplar of your process, is not trustworthy. It's rubbish. The assumptions underlying soweego are false, the software will do damage, and you are unwilling to address this central problem. There is a world of difference between a useful external source, such as MusicBrainz; and abstracting its content with no sanity filter applied: that's what will lead to adding crappy data into WD. The community would be most grateful if you did not do this. --Tagishsimon (talk) 18:04, 12 July 2021 (UTC)
In general, the more aggressive and ill-disposed someone gets about something, the less likely I am to engage. Therefore, I'm afraid I can't contribute further to this specific point of discussion, which I think is outside the scope of the initial topic by the way (as also raised by other participants).
Said that, data quality is a central focus for soweego without any doubt: I believe extensive work has been made towards that direction, demonstrated by the available documentation material. Concentrating the discussion around one single example just wipes out all the past efforts. In any case, no worries: the team will do its best to avoid the addition of low-quality URLs via blacklist filtering. Thanks for your time --Hjfocs (talk) 09:10, 13 July 2021 (UTC)
@Hjfocs: In regards to a black-/whitelist, I think it would be a good idea to split up the project. In a first phase you would only add URLs for which a property exists. Essentially, this is using the list of external identifiers as a whitelist. You could generate some statistics on the URLs that could be added in a second phase, like prevalence of domains. Based on that list, new properties could be proposed or domains could be whitelisted, etc. --Azertus (talk) 13:48, 19 July 2021 (UTC)
@Azertus: I fully agree with your idea, thanks a lot for sharing. Let me integrate it into my thoughts:
I'd love to help with the creation of new relevant identifier properties. Cheers --Hjfocs (talk) 15:11, 19 July 2021 (UTC)
  • @Hjfocs: The sample you gave above is about Elvis Presley (Q303): to make this more predictable can you do the edits mentioned in the sample:
Also, can you correct the reminder of the description (deprecated rank isn't the one to use). Also It's odd that the reference is MB ID "01809552", but that doesn't exist any more. --- Jura 15:09, 19 July 2021 (UTC)
Hey @Jura1: that's just a toy example for illustration purposes, please don't interpret the values as real ones. Anyway, I think I can translate it into a real example and do the actual edits on Elvis Presley (Q303).
With respect to using end time (P582) instead of the deprecated rank, I'm not sure this is the way to go, and would love to hear more supporting references before taking such a decision. My rationale for deprecating an identifier statement that no longer exists in the target catalog is the impact on the truthy statements. I believe this is the most used dataset to query Wikidata: see for instance Wikidata:SPARQL_query_service/queries/examples#Showcase_Queries, where lots of queries use the truthy prefix wdt. Leaving a no longer valid identifier with a normal rank plus end time (P582) would entail either more complicated queries that should take into account end time (P582), or just wrong results in case of the simpler truthy queries.
Cheers --Hjfocs (talk) 15:44, 19 July 2021 (UTC)
@Jura1: I added 3 real URLs coming from MusicBrainz (Q14005) which could be then converted into identifiers: Q303#P214, Q303#P6960, Q303#P5404. I also added Q303#P20, which interestingly contradicts the current statement. Hope this helps shed more light! Cheers --Hjfocs (talk) 16:20, 19 July 2021 (UTC)

Property for brand name

Hello! I'm missing a property to record the brand name of a company if it differes from its (offical) name. (For example, the Switzerland's postal operator is officially called "Die Schweizerische Post" in German (or "Die Schweizerische Post AG" including the legal form), but it's German brand name is "Die Post".) Best regards --Dafadllyn (talk) 19:04, 18 July 2021 (UTC)

Other than adding it as an alias, the possibility is that it's best done by creating a discrete item for the brand name, and linking it to the company using owned by (P127), as per Stagecoach London (Q7596856). I suppose one could also add it to a company using, for instance, official name (P1448) with a qualifier of has characteristic (P1552) taking the value brand name (Q2519914). --Tagishsimon (talk) 19:19, 18 July 2021 (UTC)
Thank you for you reply! In my opinion creating discrete items makes much sense if the company owns brand names for products it sells, but less so for the brand name of the company itself. I think it should be possible to record the brand name of the company directly on the item itself as it is done with official name (P1448) or alternative name (P4970). --Dafadllyn (talk) 18:56, 19 July 2021 (UTC)
PS: There was a warning saying that "has quality is not a valid qualifier for official name". --Dafadllyn (talk) 19:02, 19 July 2021 (UTC)
Where a warning is inappropriate, the constraint should be altered, IMO; especially with respect to qualifiers. The pattern tends to be a user/small group imagines the set of valid qualifiers, fails to imagine other equally valid qualifier use cases. --Tagishsimon (talk) 19:13, 19 July 2021 (UTC)

Merge Request

Please merge municipality types in Switzerland (Q2101494) into municipality of Switzerland (Q70208) --Kolduxrep (talk) 23:06, 18 July 2021 (UTC)

No. They are about different concepts. The first is about the different types of municipalities in Switzerland, the second focuses on one of those types (all afaics). Besides, language wikis like DE wiki have articles on both, so the two items cannot be merged. But, mainly, despite their name; different things, different in scope. --Tagishsimon (talk) 23:11, 18 July 2021 (UTC)
Oh, I see. Would it be possible to rename municipality types in Switzerland (Q2101494) to municipality type in Switzerland for example? I think it is quite confusing and had to correct a few mislabeled items to municipality of Switzerland (Q70208). --Kolduxrep (talk) 08:35, 19 July 2021 (UTC)
@Kolduxrep: Absolutely, yes. One or both together of the label and the description should indicate the subject/scope of the item/article, and the way in which it differs from other item/articles. I've added the 'type' word; the descriptions for both items already made the distinction clear, but it's no bad thing IMO to be pedantic with the label. --Tagishsimon (talk) 08:45, 19 July 2021 (UTC)
@Tagishsimon: Thank you very much for your effort! The distinction should be clear now. --Kolduxrep (talk) 09:45, 19 July 2021 (UTC)

Wikidata weekly summary #478

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Events
    • Past:
      • Next Linked Data for Libraries LD4 Wikidata Affinity Group call: Carlo Bianchini (University of Pavia), Stefano Bargioni (Pontifical University Santa Croce (Rome)), and Camillo Carlo Pellizzari di San Girolamo (University of Pisa, Scuola Normale Superiore) will discuss their recent project and article “Beyond VIAF: Wikidata as a Complementary Tool for Authority Control in Libraries”.; July 13th.
    • Upcoming:
  • Tool of the week
    • searchy.toolforge.org is a semantic search engine to find articles on a topic and filter results by metascientific info (gender and region).
  • Development
    • Mismatch Finder: Continued working on the store part. It can import first mismatches now.
    • Fixing an issue where month and year are sometimes parsed as the first day of the month (phab:T233105)
    • Fixing an issue with adding sitelinks for some of the newer Wikipedias (phab:T285919)
    • Working on improving the Lua usage tracking for redirected Items (phab:T280910)
    • SPARQL was amended on the 4th July to use the report #title, if present, as the anchor for the 'Try it' link. (example)

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.

Why are some varieties of English not selectable in native labels?

I've edited couple Australia related items and it seems that for certain native label properties such as Property:P1559 do not allow selection of Australian English, but does allow Canadian, British or American varieties. Is there a reason for this? How do I label it with Australian English? Melmann (talk) 10:15, 16 July 2021 (UTC)

Thanks for pointing the way. Found phab:T286862 but no movement for 3 years... Melmann (talk) 08:13, 20 July 2021 (UTC)

Duplicate art paintings

These two imaqes are of the same painting. Q5634388 (cropped and jpg) and Q50907959 (framed and tiff). The museum accession number is identical, BHC0594. Do they need to be merged or kept separate? Why? The painting has an aquatint version, how should it be linked? Also the painting is one of a set of two; how should the companion piece (Q50919662) be linked? Broichmore (talk) 09:40, 20 July 2021 (UTC)

@Broichmore: The first is about a ship, the second a painting. They should not be merged. Thanks. Mike Peel (talk) 09:45, 20 July 2021 (UTC)
In terms of linking the two paintings, they are already linked as structured data through having the same creator. I am not sure if there is another property to link related pieces of work. There are the properties for follows and followed by if they are clearly part of the same series of works. From Hill To Shore (talk) 11:36, 20 July 2021 (UTC)

Modelling archaeological sites

I am currently in a discussion (which you can read here) with user Bouzinac about an edit he made to the item on Babylon (Q5684). His idea was to split it into two items: one about the ancient city, and one about the archaeological site (which is now here Babylon (Q100329356)). This split wasn't carried out very neatly (to be honest) but that's not the point I want to discuss.

Bouzinac has argued that, in principle, every item about an archaeological site should be split in two: one item about the entity as archaeological site, and one item about the entity as city/human settlement/graveyard/whatever it was in the past.

I would say that, instead of splitting them, it is better to model them with things like start/end dates. So rather than having two items about the city/archaeological site of Babylon, there would be one item with instance of=archaeological site and instance of=city/human settlement/whatever, qualified where necessary, with start/end dates (for a simple example, see Norşuntepe (Q2001555)).

Since this would not only affect Babylon, but tens of thousands of other items, I would like to get some more input on this. Thanks! Zoeperkoe (talk) 08:09, 20 July 2021 (UTC)

Start and end dates on P31 don't really work that well. --- Jura 09:21, 20 July 2021 (UTC)
@Jura1: Out of curiosity; why don't start/end dates work well on P31? Zoeperkoe (talk) 14:00, 20 July 2021 (UTC)
  • Property proposals are a way to find consensus about how things are modelled and there's no reason to change the existing consensus to have separate items here.
Generally, it's easier to query information in statements then information in qualifiers. If you mix different entities together in one item that's likely to lead to different combinations of items to be structured differently and make it harder to interact with the data. ChristianKl ❪✉❫ 09:56, 20 July 2021 (UTC)
@ChristianKl: This property seems to assume that there's a clear cut distinction between when a site was used, and when it became an archaeological site. Also, if you take the logic of this property proposal to its full conclusion, it would mean that you would have to make a new item every time an entity changes what it is. Take the following hypothetical simplified example (and to be clear, this is not far-fetched, this kind of occupation history is common in thousands of sites in the Middle East alone; look up the concept of tell (Q755017): a site was first occupied in the Neolithic period as a small farming village. Then it was abandoned. Then it was reoccupied on exactly the same spot during the Early Bronze Age as a small city with city walls. This settlement was destroyed by fire. Then it was reoccupied again in the Iron Age, first as a small fortress, which was destroyed by fire and then immediately reoccupied as a small village. This village was abandoned. Between the Iron Age and early Islamic period, there's evidence at the site of use by nomadic pastoralists (sheep/goat herders) who did not build houses or anything. Then the site was used as a cemetery during the early Islamic period. Finally, the site was reoccupied, on top of all these older layers, in the Ottoman period, and continues to be occupied until today. The builders of this village reused some of the older material from the site in the construction of their houses (mudbricks, stone foundations). In 1920, some archaeologists came and started to dig in the middle of this village to excavate the older layers. In 2010, archaeologists came again to this village, and as part of their excavation also re-excavated the 1920's excavation (this actually happens quite a bit nowadays). What would be your suggestion to model this site history in Wikidata?
Very simple : Q3265963#P1366 ==> Q270273#P1366 ==> Paris (Q90) Bouzinac 💬●✒️●💛 14:00, 20 July 2021 (UTC)
Yes, that could solve the issue with sequential occupation layers. But how does the excavation part fit in? If we take the same site with or without excavation, would you suggest to model this as follows?
Site XYZ without excavation:
  • Item 1: site occupied during period 1
  • Item 2: site occupied during period 2
  • Item 3: site as modern village
And the same site with excavation would be:
Best, Zoeperkoe (talk) 14:19, 20 July 2021 (UTC)
On some further thought, I can see where this is going. Irrespective of whether there are excavations, any site that is classified as an archaeological site would have an item with P31=archaeological site and P9047 with links to every separate occupation layer, is that correct? Not sure whether I fully agree with this solution but I guess it is one way to solve this. Interesting. Zoeperkoe (talk) 15:16, 20 July 2021 (UTC)

List of properties for humans?

Hi everyone, I wanted to double-check a property for a human item, but unfortunately the page for the list of properties for human is empty. Is this intentional or some error? I searched briefly in the archives (of the project chat) to see if there was something about it, but I couldn't find anything. Can someone enlighten me please? :-) --Soylacarli (talk) 21:13, 20 July 2021 (UTC)

@Soylacarli: Perhaps you are supposed to use Special:WhatLinksHere/Q18608871. Ghouston (talk) 00:57, 21 July 2021 (UTC)

Policy on good summaries?

Hi I recently found a bot request with code that used pywikibot and did not specify a summary. The result is edits with no summary detailing which code/script/tool was used. I wonder if we have a policy currently on good summaries? Good summaries are vital for quality assurance, code discoverability and innovation IMO.--So9q (talk) 09:35, 21 July 2021 (UTC)

Pratar du om manuella redigeringar eller redigeringar med bot/script?
Förutsätter inte möjligheten att skapa sin egen redigeringssammanfattning det att man redigerar med script? 98.128.181.19 09:51, 21 July 2021 (UTC)
From Wikidata:Bots#Approval process: When working in namespaces that allow for customized edit summaries, bots should always use descriptive edit summaries that indicate what task is being performed and indicate that the action is being performed by a bot. —MisterSynergy (talk) 11:16, 21 July 2021 (UTC)
  • I think one of the problems is that some bots may have relied on automated edit summaries and these changed some time back and may have become much less descriptive.
Accordingly, more care is needed from the bot operators and bots authorized in the past might no longer do it properly. --- Jura 11:20, 21 July 2021 (UTC)

Disa albomagentea vs Disa albomagentea (Orchidaceae), a new species from the Hottentots Holland Mountains in the Cape Floristic Region, South Africa

Hi there! This is the same specie! Can somebody please rectify. See [7] Regards. Oesjaar (talk) 15:37, 21 July 2021 (UTC)

have different instance of (P31), one is about an article. --- Jura 15:56, 21 July 2021 (UTC)

Language names (in English and other languages)

At Help:Wikimedia language codes/lists/all/in nl, there is now an improved version of the list of language names used by Wikidata and their code. That is:

  • The language name in English as it appears in the interface
  • The English label of the item for the language
  • The title of the English Wikipedia article about the language
  • The codes used by WMF and IETF for the language

It also shows the name in an additional language, here Dutch (nl). If that language is missing, it falls back to language code (or English).

You probably noticed that Wikidata is the main user of these (non-English) language names, so if they aren't complete or up-to-date, it's probably because we haven't requested an update yet.

The list can be adapted to other languages and help us complete the language names. For Dutch, there is already a ticket at phab:T231748.

The list also includes the language name in the language itself (autonym), but this isn't directly used on Wikidata. For codes that are missing, please remember Help:Monolingual_text_languages. --- Jura 08:56, 12 July 2021 (UTC)

  • For endonym (Q1266782), there is now, in addition the native label (P1705) value and MediaWiki version, the item label and WP article (if one exists or can be added - not all are termbox languages).
It seems that some P1705 values combine several scripts and/or transliterations. Ideally, these would probably be split/moved to a transliteration or transcription (P2440) qualifier. --- Jura 12:23, 16 July 2021 (UTC)

Merge proposal

2021 Moldovan Super Cup (Q107608229) and 2021 Moldovan Super Cup (Q107337407) contain the same information, only in different languages, and should be merged. Sørhaug (talk) 11:04, 22 July 2021 (UTC)

Record producer

Hi all. I could not find which thingy to use for the "instance of" for record producing companies. I could find "record label" or "record producer" (person). What do we use for the company? Several companies in this area in my watchlist use "company" or "record label". For example the so famous Atlantic Records (Q202440) does not even qualify as a "company" but only as "record label". I think we do need something that would fill this gap, and not only for records but also for "CD"s, "cassettes", etc; I mean something like "music company", "musical production company"? Thanks for your input. --E4024 (talk) 14:29, 22 July 2021 (UTC)

sounds good. go ahead and make it. probably should be a subclass of production company (Q11396960). BrokenSegue (talk) 15:04, 22 July 2021 (UTC)
wait we already have music production company (Q58318936). Is that not enough? BrokenSegue (talk) 15:10, 22 July 2021 (UTC)
Just what I was afraid of! When I saw "music company" did not produce any results at the search box, I thought about trying "music production company" but said to myself "If Atlantic Records (Q202440) item is made this way, there would be no chance" and did not look... My bad. Thanks BrokenSegue. --E4024 (talk) 15:15, 22 July 2021 (UTC)

@E4024: Why do you think that Atlantic Records (Q202440) is a music production company (Q58318936) if actually it's record label (Q18127)? Eurohunter (talk) 16:51, 22 July 2021 (UTC)

  • Wow, we also had record company (Q2442401)! I was indeed very surprised that it looked like we lacked something here instead of having redundant items, the latter of which is the most accustomed situation in WD. Your question has the answer inside, Hunter: We have seen many records produced by this company (and that is why they have that label on them, I guess :) therefore I thought so and I still think so. If you think differently, do as you wish. If you have a look at my contributions, you will see that except very rare cases I only state what I think is correct but avoid edit wars. The item is all yours. --E4024 (talk) 17:46, 22 July 2021 (UTC)

Removing a general value when a precise alternative is in place?

Cf. date of birth and date of death fields in Q380282. There is a precise YYYY-MM-DD date (2 refs) and a not-so-precise YYYY only date (9 refs). Is it ok to remove the YYYY entry, with all its refs, altogether - or what is the best course of action?  – The preceding unsigned comment was added by Retired electrician (talk • contribs).

@Retired electrician: The best practice is setting preferred rank to the most precise value (cfr. Help:Ranking); the qualifier reason for preferred rank (P7452)most precise value (Q71536040) can be applied to it. --Epìdosis 10:06, 19 July 2021 (UTC)
A bot was set up to perform setting the preferred rank for these entries, does anyone know the status of it? The problem was that all values of equal rank show up in the infoboxes, so they display the year only and the precise date. --RAN (talk) 18:18, 19 July 2021 (UTC)
@Richard Arthur Norton (1958- ): sounds like you are talking about my bot. It has set preferred rank on lots of items but there's an issue with the SPARQL query which is limiting its progress(see my section below about this). BrokenSegue (talk) 23:54, 22 July 2021 (UTC)

Stuck on a merge

I have read the instructions for the merge gadget but I am not clear which I would be merging into which.

Dominant Species (Q36899015) is intended to be the boardgame item and has a few statements. However it does not link to the Wikipedia page.

Dominant Species (Q105905489) has a link to the Wikipedia page but has no other data. (I really found out about Wikidata by following the bug report in the page stats page for Dominant Species.Slimy asparagus (talk) 23:05, 22 July 2021 (UTC)

Merged. --Tagishsimon (talk) 23:13, 22 July 2021 (UTC)
Thanks. Slimy asparagus (talk) 23:17, 22 July 2021 (UTC)
@Slimy asparagus: Generally, merge into the older item, with the smaller Q number, which is the default behaviour of the merge gadget. Ghouston (talk) 01:42, 23 July 2021 (UTC)

Guidelines on categories of humans

Hi all, are there guidelines for items about categories of humans? For instance, Category:Nurses (Q6144739) is very messy when you look at the interlingal links: sometimes the link is towards a gendered category, sometimes a gender neutral category. Pdehaye (talk) 18:35, 17 July 2021 (UTC)

  • Normally, sitelinks should be to categories with similar scope. The above item doesn't look like it was much edited. So if you think it should be sorted out, please go ahead. --- Jura 19:03, 17 July 2021 (UTC)
    • The question was prompted by a debate I started when I wanted to make some changes on French wikipedia, naively seeking to change the naming of a category to reflect the more gender inclusive phrasing and writing used in French-speaking Switzerland. This prompted a huge debate originally focused on my use of gender-inclusive language (a recurrent issue on French wikipedia, where it is frowned upon), but now evolving into a poll on the strategy to adopt with respect to gender and category hierarchies. In the end, I care less about the naming of the category or even the strategy used for picking a hierarchy than properly modeling those hierarchies (i.e. not collapsing nuances across languages). My motivation is mostly around not perpetuating gender biases in tools based on wikidata outputs. Pdehaye (talk) 13:22, 23 July 2021 (UTC)

Cast members

What is the inverse statement for "cast member" (in a TV series, such as Hercai)? Part of, participant in? Is there something like "notable work" for actors? Notable role? Please see Akın Akınözü (Q65307943). Thx. --E4024 (talk) 02:26, 23 July 2021 (UTC)

@E4024: there is no inverse statement for "cast member" due to maintenance issues. However, you can use notable work (P800) for notable roles as in Q39476#P800. There is also an inverse label for "cast member" property - cast member of (Q66318312). To access it, you can enable relateditems gadget in Special:Preferences#mw-prefsection-gadgets. --Lockal (talk) 09:13, 23 July 2021 (UTC)

Most recent Alexa rank data

I am looking at the Alexa rank (P1661) for BoardGameGeek (Q887528). There are two entries. I was hoping to be able to pick the latest one. The best I have been able to do is:

{{wikidata|property|references|P1661}}

This returns the first value (1983) not the most recent (1745). That is not surprising in of itself, but I was not able to refine it by playing with the parameters. Also I obviously need the the date of the record too, which I assumed was a qualifier.

So I was hoping to produce something (which updates when a new Alexa record comes into wikidata) that ends up looking like:

As of 10 March 2019, Boardgamegeek.com has an Alexa rank of 1,745[6].

Slimy asparagus (talk) 07:07, 23 July 2021 (UTC) Figured it out eventually:

As of {{wikidata|qualifier|preferred|P1661|P585}}, boardgamegeek.com has an [[Alexa rank]] of {{wikidata|property|references|preferred|P1661}}.Slimy asparagus (talk) 14:49, 23 July 2021 (UTC)

Okay I am getting more suspicious of Alexa data. Looking into I realise that the Alexa bot was turned off in 2018, the developer has gone awol and that there were concerns about the data it was generating. I think any data where the Alexa site was not archived cannot be verified and should be deleted. Slimy asparagus (talk) 23:28, 23 July 2021 (UTC)

family (Q8436)

To use with families (Q8436) do we have something like family picture or family grave picture? We do have group photograph (Q70077691) which we may use for family photos taken in life, but what can we do for those that we only have a family cementery? Please see Koç family (Q6435455)... Thanx. (BTW I hate to make items with exclamation marks!) --E4024 (talk) 14:09, 23 July 2021 (UTC)

Wrong data

They add wrong data over and over (2021 Bahrain Grand Prix (Q104188668), Mercedes-Benz in Formula One (Q65954812)). I have to watch it and revert all the time otherwise it will be destroyed atlest once a month? Eurohunter (talk) 13:15, 21 July 2021 (UTC)

The information you removed - at least, the first instance I looked at - diff was not wrong. Perhaps there is a debate to be had, or analysis to be done, as to how dates are associated with GP items, but there is nothing obviously objectionable to adding a point in time (P585) statement with the date of the GP, whether or not there are start time (P580) and end time (P582) statements. WD has ample room for diversity in the way data is held, and diversity provides richer means by which data can be queried. WD has much less room for users who are sufficiently convinced that their particular, but limited, viewpoint is right that they edit war to remove valid statements. --Tagishsimon (talk) 15:07, 21 July 2021 (UTC)
@Tagishsimon: But Grand Prix lasts 3 days. It could be used only for separate items for free practises, qualifying and race sessions. Eurohunter (talk) 19:57, 21 July 2021 (UTC)
It's an interesting question; is the item for the GP weekend, or for the race itself. If we look at the set of GP items, it's clear that the vast majority - 1306 - have a P585, whereas only 20 & 21 have P580 & P582 - see https://w.wiki/3fTF - and so it seems that the race, rather than the period in which practise and qualifying takes place, is what's being modelled. --Tagishsimon (talk) 21:29, 21 July 2021 (UTC)
No. Practise, qualifying and race are all parts of one Grand Prix. If I would not add exact periods to 2021 GP's then they would still keep adding one date - maybe someone would realise before or after 2050. Eurohunter (talk) 16:47, 22 July 2021 (UTC)
No. That's a dogmatic misreading of the situation. Without getting hung up on the words, there is a GP weekend with components including practise, qualifying and race. It's quite fine for WD to have items on the weekends, and on the components. Right now, whether you like it or not, 1306 GP items are modelled as the race - the thing that happens on Sunday afternoon/evening. 20 are modelled as the weekend. The solution, if this is a thing of concern to you, would be to create distinct weekend and race items for each GP. The solution is not to suggest that your concern with the representation of the weekend trumps the long established and overwhelming WD status quo of modelling Sunday's race. You need to get your head around the fact that other people see things differently from you, and their views are wholly legitimate and sane. --Tagishsimon (talk) 16:53, 22 July 2021 (UTC)
@Tagishsimon: I agree we definitelly should have items for Grand Prix weekends (P580 and P582) and each components (P585). 1306 items is to be verified and fixed - it's just because I know myself how hard is to find exact dates from early years of F1 so it was easy just to add one date so making mess then it was easy to keep this mess same like in case of relases (albums and its tracks). Additionally bot probably just imported race date from ENWP from race details part of infobox... Maybe it's evem the main reason. It was always confusing me - it's better to fix it now than in 10 years. Eurohunter (talk) 17:15, 22 July 2021 (UTC)
I can see the logic behind using start time (P580) and end time (P582) on Grand Prix items, but removing point in time (P585) loses us important information. The item no longer contains information about when the race actually starts. For example en:2021 Bahrain Grand Prix uses the race date as the official date rather than the date the race meet started. Additionally, Wikidata:WikiProject Motorsports suggests that point in time (P585) is used on all items representing an auto race. One option is to include a qualifier on point in time (P585) to show what exactly it refers to. For example:
point in time
  24 July 2021 edit
significant event Race Start
0 references
add reference


add value

In the future, items may be created to represent Grand Prix weekends and have the race items linked to them with part of (P361) and have start time (P580) and end time (P582) moved the the race meet items. This is something that is not unique to F1 races (e.g. NASCAR races meets could be modelled the same way). Does this sound like a reasonable solution? Teester (talk) 14:14, 24 July 2021 (UTC)

Query Service Timeout Change?

I've noticed some of my bot scripts have started failing recently. Seems like the timeout for the SPARQL query service was reduced to 1 minute. It definitely used to be longer than that. Was there an announcement about this? Any hope of granting bots a longer timeout? Or some other work around? BrokenSegue (talk) 15:05, 22 July 2021 (UTC)

Hey BrokenSegue, thanks for reaching out. We're not aware of any changes to the timeout recently; our understanding is that the timeout has been 60 seconds at least going back to 2018, if not further back.
As a general rule, WDQS (the SPARQL query service) presents unique performance challenges due to the nature of the underlying datastructures; as such having sane timeout values is one part of our strategy to maintain high service availability. For that same reason we don't generally increase the timeout for individual users or automated queries specifically, because we're already dealing with the set of unique performance challenges that come up when running a publicly available SPARQL service.
It's hard to give great advice without knowing your use-case but generally I'd recommend trying to see if your needs can be met with less expensive queries that will finish within the timeout window. RKemper (WMF) (talk) 20:34, 22 July 2021 (UTC)
@RKemper (WMF): Thanks for the response. I'm somewhere between very and 100% sure that the queries didn't used to timeout at exactly 60 seconds. Perhaps there was a bug that allowed longer queries or how the timeout is being enforced changed. And queries that used to reliably complete no longer do. In any case it sounds like I'm not going to get a fix here. I'll try to find a work around. BrokenSegue (talk) 22:38, 22 July 2021 (UTC)
  • "Wikidata:Database reports/items with P569 greater than P570" stopped running because of timeouts. So the query was broken into smaller parts and then the parts concatenated so that when one finished, the next would start, each running under 60 seconds. --RAN (talk) 01:37, 23 July 2021 (UTC)
    yeah that's probably what I'll do. It's just frustrating and inefficient. BrokenSegue (talk) 02:02, 23 July 2021 (UTC)
  • I agree, we may need to double the time allowed as the database size doubles. --RAN (talk) 03:13, 25 July 2021 (UTC)

merge rules info. issue about Regina Pessoa-Pureur (Q60729701) Regina Pessoa-Pureur (Q88384947)

where can i find info about merge rules or acceptable criteria or suggestions. Regina Pessoa-Pureur (Q60729701) and Regina Pessoa-Pureur (Q88384947), contain identical names, one w/o identifiers. i am assuming both can be merged. --Agyaanapan (talk) 10:40, 24 July 2021 (UTC)

I've managed to expand Regina Pessoa-Pureur (Q60729701) based on a large number of linked items. Regina Pessoa-Pureur (Q88384947) is now the inferior item as we only have the ORCID and the information isn't available to public view. If someone can see what is in that ORCID record and confirm they are the same individual, then a merge would be appropriate. However, without that confirmation, merging would be a bad idea as we could be conflating the IDs of two different researcers with the same name. There are many cases of duplicate names among researchers, especially if the records cover several generations. One researcher may be named after a grandparent or cousin who worked in the same field. From Hill To Shore (talk) 11:25, 24 July 2021 (UTC)
hello From Hill To Shore, after data is updated on orcid id, we need to decide whether it should be merged or keep seperately. so we have to search online thoroughly before merging items. -Agyaanapan (talk) 12:58, 24 July 2021 (UTC)
For such items, you may search EuropePMC to see how the ID is used there.--GZWDer (talk) 04:13, 25 July 2021 (UTC)
I think they can be merged. The only usage I can find of that ORCID (using a web search) is on [8] and the corresponding Wikidata item Cytoskeleton as a Target of Quinolinic Acid Neurotoxicity: Insight from Animal Models (Q39396792) links to Regina Pessoa-Pureur (Q60729701). Ghouston (talk) 04:59, 25 July 2021 (UTC)

Video

I want to show a video with a length of 00:07:05 minutes, it is on my DVD-RW drive (D :). It is from 1913 and how can I represent it?. Regards, Wname1 (talk) 12:04, 25 July 2021 (UTC)

Wikidata does not store videos. Please see Commons instead.--GZWDer (talk) 15:26, 25 July 2021 (UTC)

Sports: Failed shots at goal

Hi there,

I'm trying to make 2011 Rugby World Cup Final (Q4622275) a bit more complete by putting more of the game events by focusing on significant event (P793) instead of points/goal scored by (P1363). As per the wikipedia page 2011 Rugby World Cup Final, there were four failed penalty goals, one failed drop kick, and one failed conversion kick.

I asked this over at Wikiproject Sports, but since there's been no response, I thought I'd ask a wider audience. I'd like some advice on how to record a "missed goal" as a significant event (P793) in a match. I gave a few potential ideas, but I'm open to anything.

  1. I could record two values. I could create a new item "Conversion kick" to differentiate from a conversion (Q63884263), if only the first exists, then a kick on goal was attempted, but only if the second exists were points scored.
  2. I could use the qualifier number of points/goals/set scored (P1351) only if the kick was successful. If that qualifier is not there, then we assume it failed.
  3. I could create two new items. "converted try" and "missed conversion".

I think my preferred is the third option, because it's the only one that doesn't require you to have trust that the full information has been entered. I've created a template at User:Supertrinko/Sandbox that shows what I'm going for.

Maybe outside of sports someone has experience with recording successful and unsuccessful events that could be replicated in sports items? Supertrinko (talk) 21:33, 25 July 2021 (UTC)

Advice on using initial or name for given name (P735)

For Charles W. Penrose (Q107631467), most sources I have found use W. (Q19803523) as the second name while Library of Congress authority ID (P244) identifies the name as Wilkinson (Q97690211). On the principle of mapping rival claims, I have recorded both options against given name (P735) but with the same series ordinal (P1545). Is this a good way to handle the situation, or will it cause problems? From Hill To Shore (talk) 10:02, 24 July 2021 (UTC)

@From Hill To Shore: I agree with the solution.Vojtěch Dostál (talk) 07:48, 26 July 2021 (UTC)

Gyumri / Akyaka border post (Q67893109)

I want to make a call to those users who have come here to make a database to edit Gyumri / Akyaka border post (Q67893109). I was fired away from a couple of WPs for fighting against or in other cases I abandoned voluntarily another couple of WPs to escape from such POV "contributions". We cannot have an item's "description" like "Meetings have been held, agreements have been signed, but until further notice, this border is still closed. Very, very closed." This is abuse of Wikidata for propaganda. --E4024 (talk) 13:25, 24 July 2021 (UTC)

This was fixed by Mahir256 Vojtěch Dostál (talk) 07:50, 26 July 2021 (UTC)

Items with two "identical" date of birth statements

Hello, we have quite a few items with several date of birth/date of death statements, with year precision, same calendar, without qualifiers, yet pointing to the same year. The underlying cause is that +1942-00-00T00:00:00Z","precision":9" and "+1942-01-01T00:00:00Z","precision":9" are understood as different dates. I've tried to design a query for some of these: 

select distinct ?item where {
  ?item p:P569 ?s1, ?s2 filter(?s1 != ?s2) .
  ?s1 ps:P569 ?v1 ; wikibase:rank ?rank1 ; psv:P569/wikibase:timePrecision ?prec1 ; psv:P569/wikibase:timeCalendarModel ?cal1 filter (?rank1 != wikibase:DeprecatedRank && ?prec1 = "9"^^xsd:integer ). 
  ?s2 ps:P569 ?v2 ; wikibase:rank ?rank2 ; psv:P569/wikibase:timePrecision ?prec2 ; psv:P569/wikibase:timeCalendarModel ?cal2 filter ( ?cal1 = ?cal2 && year(?v1) = year(?v2) && ?rank2 != wikibase:DeprecatedRank && ?prec2 = "9"^^xsd:integer ) . 
  minus {
    ?s1 ?pq1 [] .
    ?qualifier1 wikibase:qualifier ?pq1.
  }
  minus {
    ?s2 ?pq2 [] .
    ?qualifier2 wikibase:qualifier ?pq2.
  }
} limit 10
Try it!

Would you agree that results of this query should get the corresponding statements merged into one? Vojtěch Dostál (talk) 07:25, 26 July 2021 (UTC)

Yes! --Epìdosis 12:36, 26 July 2021 (UTC)

Wikidata weekly summary #479

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • Added a tags parameter to the wbmergeitems API so Item merges can now also get an edit tag (T286778)
    • Mismatch Finder: Continued working on the store part. This week we focused on uploading a file with potential mismatches and the UI for getting a list of all uploaded mismatch files.
    • Addressing comments from the security review of the Query Builder so we can hopefully soon move it from the test system to its proper place under query.wikidata.org

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.

Merge the City of Ornbau

For Ornbau there are two records: https://www.wikidata.org/wiki/Q572242 https://www.wikidata.org/wiki/Q32209001

One is described as town, the other as main settlement. I live there since my birth. It's just a town.

Another question: Can I add the instance of City (Q515). It has just 1653 people, but's it's official a city (since 16th century). Zottmann (talk) 15:07, 20 July 2021 (UTC)

They're not the same. Ornbau (Q572242) is the entire stadt (i.e. this area). Ornbau (Q32209001) is the main place within that area, excluding the other places like Gern (Ornbau) (Q1514337), Oberndorf (Ornbau) (Q2010515) and Haag (Ornbau) (Q1566275). There's no need to add city (Q515), it's redundant because urban municipality in Germany (Q42744322) is already a subclass of it. - Nikki (talk) 15:47, 26 July 2021 (UTC)

Diana alcocer

Buenas tardes, quiero poner una queja ya que busqué Fernando Sanclemente Alzate y sale narcotraficante, cosa que no es cierta!!! Y quiero poner la queja y el cambio de modo inmediato

Well, good luck with that. It does not seem to be a wikidata issue; perhaps https://es.wikipedia.org/wiki/Fernando_Sanclemente ? As a general observation, if your complaint does not specify - for instance by URL - the location of the supposed problem, you have not done all that you could or should have done in making the complaint, and you leave the reader of the complaint wondering where the problem is. Wikidata? A language Wikipedia? Google? Who knows. --Tagishsimon (talk) 23:27, 26 July 2021 (UTC)

river water body (Q30092769)

Hi all. Nothing other than "river" (Q4022) links here. Do we really need this? P.S. Please do not give me angry replies, I did not open a DR (deletion request), just asking... --E4024 (talk) 23:58, 26 July 2021 (UTC)

On the face of it, yes we do, b/c Q30092769#P973, where the term is defined. Both it and Q4022 claim they're different from each other, although on scant authority. I think you'd need to be certain about the definition in the Eionet Data Dictionary before you could consider a merge. --Tagishsimon (talk) 00:17, 27 July 2021 (UTC)

Wikidata error

I attempting to add Wikipedia entry on Seediq (Q716686) but this error appear: "A page "Taroko" could not be found on "trvwiki". The external client site "trvwiki" did not provide page information for page "Taroko"." The page on trvwiki actually exist. Why there is this error? --2001:B07:6442:8903:E9DC:FE05:45FE:A635 15:03, 26 July 2021 (UTC)

trv:Taroko does not exist as far as I can see Vojtěch Dostál (talk) 20:11, 26 July 2021 (UTC)
So why there is a link in this page (it is near to the end of the page, number 272)? --2001:B07:6442:8903:5588:25D6:AB9B:45B9 07:42, 27 July 2021 (UTC)
Because someone added the link there without checking if the page exists or not.Vojtěch Dostál (talk) 12:39, 27 July 2021 (UTC)

Types of division (Q334453)

Division is one of the parts into which a business (?), organization or company so there are two or three types of division and they are company division (Q107658553) and division (Q107658606) ("organization division")? Eurohunter (talk) 10:10, 27 July 2021 (UTC)

Questions for subclass of (P279) and inverse property label item (Q65932995)

1. How there can be inverse property label item (Q65932995) (mostly unused)? How is generalized by (Q99419468) inverse property item for generalization of (P7719) while it is item not property? "inverse property item" and "inverse label for property" (what ever it means) are used interchangeably. If there are anthem (P85) and is anthem of (Q65933124) I could understood if they both are properties or is anthem of (Q65933124) would be used as qualifier. Eurohunter (talk) 10:48, 27 July 2021 (UTC)

2. Why there is no inverse propery for subclass of (P279)? If item is element of higer class then why there is no information about its lower class? You will only know about higher class with subclass of (P279) so "you can go only one way and can't back". I understand that subclass of (P279) is indirectly way for me to find higer class of item but I can't find its lower class in this way. Eurohunter (talk) 10:48, 27 July 2021 (UTC)

Help:Ranking

According to other help pages and content of Help:Ranking it should be under Help:Ranks. Eurohunter (talk) 12:05, 27 July 2021 (UTC)

Operating area of EDRi

I'm trying to fix the operating area of EDRi Q1376533, which was incorrectly stated as the European Union, while it is actually all memberstates of the Council of Europe. I'm getting an error, which I'm not sure how to fix. BFG (talk) 22:56, 26 July 2021 (UTC)

Sorted diff --Tagishsimon (talk) 23:14, 26 July 2021 (UTC)
Thanks Tagishsimon. BFG (talk) 18:33, 27 July 2021 (UTC)

Issue with undoing a merge

I recently discovered that someone merged the items for two different people on Wikidata. I went to the "View history" tab and clicked on the option to undo this merge. However, I can now only find the record that I undid the merge on (Q61830856) and not the other record that had been merged with it (Q107553212). When I search for the latter QID, it brings up the former, as if the two were still merged.

Is the newer item (Q107553212) just permanently merged with the other one or did I miss a step to undo the merge? Or does it take a while for the records to become unmerged?

If I can't recover the original item (Q107553212), I may have to recreate it with the information I have about this person. But it would be nice to recover the original item before it got merged.

Jeannette Ho (talk) 23:06, 27 July 2021 (UTC)

You have to go to the other one that is now a redirect and restore it to before the merge. Teasing apart a bad merge is a two step process, one step for each of the two entries. --RAN (talk) 23:10, 27 July 2021 (UTC) See: https://www.wikidata.org/w/index.php?title=Q107553212&oldid=1461821433 for the pre-merge version of one that is now a redirect. If the information is correct at the above link, save it, and it will be restored to the point before the merger.

Thanks, I have gone to the redirect and restored it. Now I can retrieve both items.

Jeannette Ho (talk) 23:32, 27 July 2021 (UTC)

Disa cornuta vs Disa aemula

Guys, the way these species are set up in Wikidata indicates that they are synonyms. They are not see [10]] and [11]. The South African botanist John Charles Manning also indicates that these two species are completely independant in his book Field guide to Fynbos dated 2018 ISBN 978 1 77584 590 4. Can I continue to rectify this mess? Regards. Oesjaar (talk) 09:12, 25 July 2021 (UTC)

@Oesjaar: These are independent taxon names and should not be merged under current scheme of things in Wikidata. I guess it would be useful to indicate that some sources consider them synonynous, using taxon synonym (P1420) if I am not mistaken. Vojtěch Dostál (talk) 07:53, 26 July 2021 (UTC)
Rectified! Oesjaar (talk) 19:18, 28 July 2021 (UTC)

Sprint records

Turkish athlete Pınar Saka (Q7265034) has the Turkish record for women's 400 m run with 52.99 that she broke in 2018, in New York. Got no idea how this fact can be incorporated to her item. Any good Samaritan around to help? Thanks in advance. --E4024 (talk) 15:30, 28 July 2021 (UTC)

See Kim Collins (Q311298), where something similar is entered as properties of a personal best (P2415) statement, which is a solution I rather like. There's always record held (P1000), but it would need a new item "Women's 400 metres Turkish national record". Circeus (talk) 15:59, 28 July 2021 (UTC)

Nicolas Louis de Lacaille (Q202703) vs Nicolas Louis de la Caille (Q58589263)

Guys, according to me this is the same person. Please advise on the way forward. I am out of my depth here. Regards. Oesjaar (talk) 19:17, 28 July 2021 (UTC)

Nicolas Louis de Lacaille (Q202703) is a person and Nicolas-Louis De La Caille: Astronomer and Geodesist (Q58589263) is an article about said person. --Emu (talk) 19:43, 28 July 2021 (UTC)

Long dagger ~ short dagger ~ poignard

I have create a new item Q107567655 for pugnale/puñal/poignard because in Romance languages there's a difference between them and a daga/dague (Q182780). In English probably both would be called a dagger, the long one (about 60 cm/25", a short sword) and the short one (about 25 cm/10", a double-edge knife), but I'm not sure. I'm pretty sure instead that poignard/poniard (Q2100658) in English mainly refers to a French dagger used in the past. Do you have some reliable sources to know what kind of weapon today a dagger means in English? Thanks in advance.--Carnby (talk) 12:45, 23 July 2021 (UTC)

@Carnby:, I don't think that the word 'dagger' is used colloquially in English today, as an actual weapon. Any modern references are likely to be theatrical (as in Macbeth Is this a dagger which I see before me, The handle toward my hand? Come, let me clutch thee. I have thee not, and yet I see thee still.). There is also a class of detective fiction called "cloak and dagger" (from a typical silhouette image on the cover - classically the dagger looks like a Fairbairn–Sykes fighting knife). In ordinary usage, it is just a knife - exceptionally a hunting knife, combat knife or just a kitchen knife. FWIW (no expertise whatever) I would class the 60cm item as a short sword and the 25cm one as a dagger or a knife. BTW, don't forget the dagger mark in typography († ‡). --John Maynard Friedman (talk) 11:13, 29 July 2021 (UTC)

Notification error for P1889

If someone is trying to merge both items with same different from (P1889) it should result in edit error with notification window and further steps. So if both items has different from (P1889) I can imagine there probably should pop up big blue window (like Blue Screen) covering whole item under {{Q}} with text such as "Both items use different from (P1889) (with instance of (P31) linked for both items - showing difference) Are you sure you want to merge these items?" with answer "Yes/No" and if someone click "Yes" then next big window should pop up with text such as "Provide reason why both items should be merged" resulting in edit description added for both items (one of them became redirect in this case). Eurohunter (talk) 09:30, 27 July 2021 (UTC)

Otherwise they will be merging different items for no reason. Eurohunter (talk) 09:31, 27 July 2021 (UTC)

@Eurohunter: If one item links to the other in any way (for example with different from (P1889)), you will not be allowed to merge them anyway. Vojtěch Dostál (talk) 18:04, 28 July 2021 (UTC)
@Vojtěch Dostál: So maybe there should be confirmation if someone want to remove different from (P1889) before merge? Eurohunter (talk) 07:46, 29 July 2021 (UTC)

Unconnected taxonomy templates

Hello there. In Turkish Wikipedia, we have so many taxonomy templates that are not linked to Wikidata. All of them have the almost same name as in English. For example, tr:Şablon:Taksonomi/Tigrisoma and en:Template:Taxonomy/Tigrisoma.

It would take too long to connect them all one by one. But I couldn't do it with any speedly way. Can anyone help about it? --ToprakM ✉ 19:58, 27 July 2021 (UTC)

maybe try Wikidata:Bot_requests. Seems straightforward. BrokenSegue (talk) 20:17, 27 July 2021 (UTC)
I did. Thank you. --ToprakM ✉ 13:27, 29 July 2021 (UTC)

Short descriptions edited unintentionally via en.wikipedia short description editor

I have just received a message thanking me for my 100th edit to Wikidata. I wasn't aware of having done any! I have now realised that, when I used the 'short descriptions editor' widget in en.wikipedia and chose the option "edit and import", I was actually editing the description on wikidata without any intention to do so. In principle this should not be a problem - except that en.wiki has a silly limit of 40 characters on the length of descriptions, imposed by a badly specified mobile app. Consequently I have changed the original useful 'natural sentence' descriptions into ultra-terse, verb-free phrases. They are not 'wrong', just rather cryptic. I've had a look at the history of some articles I changed here, to see if I could revert to the status quo ante but I can't see the previous version. Is there a way to undo my vandalism? --John Maynard Friedman (talk) 09:56, 29 July 2021 (UTC)

  • After reviewing all your edits: you have not actually overwritten any existing descriptions—all edits were made to items with empty English descriptions
  • Some of your descriptions do not meet all of the recommendations from Help:Description (capitalization, use of articles, etc.); while this is maybe not ideal, it is not totally uncommon and your contributions are clearly a net-positive contribution to this project
  • So, no vandalism here. Keep on going… ;-)
—MisterSynergy (talk) 11:59, 29 July 2021 (UTC)

SABR ID (baseball)

At Property:P2482 the scheme for the ID has changed. For Babe Ruth (Q213812), for example, it switched from "9dcdd01c" to "babe-ruth", but the old scheme still works. Can someone figure out how to go from one scheme to the other. For Clark Calvin Griffith (Q1095705) I had to use the url because I cannot figure how to find the old alphanumeric value from the new alphabetic value. --RAN (talk) 14:30, 27 July 2021 (UTC)

I tried to figure it out and couldn't. Maybe we need a new property that uses the slug. BrokenSegue (talk) 19:46, 29 July 2021 (UTC)

Notability vs. referencing (or "why isn't that a FAQ?")

Though I am here around for a bit longer than the average this question comes up from time to time, and now I have reached the threshold to ask.

I cite an example. There is an actress, well known, with her articles and WD entry. She has a spouse whose name is known/public but he is not notable. Since spouse (P26) requires a real reference (to a human) I cannot add the property since he does not exist as an object. If I do not add spouse then she looks like unmarried, which is false. What am I supposed to do in such cases? I cannot add a non-notable person and I cannot add spouse without an object to reference to.

This problem arises in various places where there are non-notable but known persons are involved, and a few times for other objects, like public, known but insignificant mentions of companies, intitiutions and such. What are we expected to do? And why is not that in the FAQ? :-) Thanks! grin ✎ 16:07, 28 July 2021 (UTC)

  • if the "name is known/public" then you must have a reference, a news article or a podcast, or an interview in any medium. Just create a reference like "stated in=New York Times" and the url for the article. Notability at Wikidata means that some reliable medium has recorded the information and another person can check the reference and come to the same conclusion. It doesn't mean "famous". We want to weed out false entries and deprecate false information that creeps into source material. --RAN (talk) 16:20, 28 July 2021 (UTC)
    @RAN: I am not sure I see your point. The question was not about references: both the marriage and the name of the spouse is referenced. It's just there isn't even 3 lines about the man so nobody would start writing an article, which is another way to say "non-notable". (As a sidenote we do know the name of the person [and often occupation] but about nothing else, including birthdate, background, etc.) grin ✎ 19:42, 28 July 2021 (UTC)
The absence of a spouse (P26) statement doesn’t mean that the person is unmarried as Wikidata operates under a w:en:Open-world assumption. Also, if the spouse in question is publicly known and there are independent sources about the marriage, then there is a good chance that said spouse will be notable under WD:N #3. If there is only information about the marriage itself but not the spouse, a spouse (P26) with unknown status might come in handy. --Emu (talk) 16:24, 28 July 2021 (UTC)
I would rather chosen a more generic approach of the problem, since the specifics are just examples.
Anyway, in this specific example I do know the name and occupation of the man (and have it referenced well) and nothing else. Nobody will start writing an oneliner article (well, okay, nobody should :-)), so the man will not have a WD entry due to lack of public data (which is possibly the same as saying "non-notable"). Setting it unknown when the person is known would be improper (=false information).
I am not sure about your open-world comment: I clearly do know that she is married, so WD shall contain that as well but I do not see a way how to do it. (And even that is specific since the question is not just about spouse (P26).) grin ✎ 19:50, 28 July 2021 (UTC)
(After edit conflict) I am not sure why you are using an example of writing an article. This is not Wikipedia. For the level of information you have suggested, a new item is valid under WD:N #3. Create the new item with the name and link it to the original item as a spouse (with references). There is no need to expand the item further. If more structured data becomes available, another user will expand it. If no more data becomes available, the item still serves the purpose of making a useful contribution to the notable spouse's item. From Hill To Shore (talk) 21:37, 28 July 2021 (UTC)
<unknown value> may just mean that there is no Wikidata item for the object/person (see Help:Statements#Unknown_or_no_values. There may be reasons against the creation of Wikidata items for non-notable spouses (e.g. legal or social reasons) that justify the use of <unknown value>. There has been a discussion about the same topic in 2019 (Wikidata:Project_chat/Archive/2019/06#Non-notable_spouses_of_notable_people). This might be of interest to you. - Valentina.Anitnelav (talk) 21:42, 28 July 2021 (UTC)
Thank you, that was a very useful pointer indeed. I also strongly dislike the use of <unknown> instead of <some value> but I see #2 is popular so I'll follow that, and start creating bazillions of referenced nobodies. :-/ [Nah, not really, but it feels like that.] The main problem with that is that since there's only a name it's almost impossible to pick duplicates as there is no data to correlate, and it's really not possible to say which of the nine hundred "John Smith" are indeed the same. Well, I guess that's a problem for another day. grin ✎ 16:06, 29 July 2021 (UTC)
  • I think you are conflating Wikidata notability with English Wikipedia notability, no one here cares if anyone is going to create an English Wikipedia article on a spouse. As people continue to point out, if it is factually correct, a spouse can have an entry. Even when a person's name is not known, we have entries for them. Search for "unnamed child" and you will see entries for children that have died at birth, that are important in the numbering system in noble families. --RAN (talk) 21:35, 28 July 2021 (UTC)
    While I see strong duplication and validation problems with this approach I can accept that.
    I thank you all for your valuable input and help, and I again suggest making this into a FAQ. grin ✎ 16:08, 29 July 2021 (UTC)

What to do about an incomplete source?

The item for Steve Giddens has a start time of 1991 for his title of FIDE Master, with a reference that just says stated in ratings.fide.com. There's no URL; the place where ratings.fide.com usually specifies the year a title was awarded, the "Titles" section of the player's profile page, is empty; and I can't find this information anywhere else, either. Should I remove the reference? Joriki (talk) 06:48, 29 July 2021 (UTC)

@Trade: Your department, iirc. --Tagishsimon (talk) 07:56, 29 July 2021 (UTC)
What does this have to do with me? @Tagishsimon:--Trade (talk) 15:55, 29 July 2021 (UTC)
I thought you were curating chess players. Perhaps not. --Tagishsimon (talk) 19:16, 29 July 2021 (UTC)
Ah. User:Steak. There you go. --Tagishsimon (talk) 19:17, 29 July 2021 (UTC)
If only they also included a date/URL we could check the database on the archive. Seems it was added in this edit but the archive doesn't go back that far for this page. Is it possible his FIDE Master date was inferred from his historical rating data? Rudimentary research here suggests his rating wasn't high enough until a few years later (though he could still be a master with a lower rating). Given we can't source this I'd say removing the date is fine. BrokenSegue (talk) 19:31, 29 July 2021 (UTC)
but I see no reason to remove the reference. BrokenSegue (talk) 19:33, 29 July 2021 (UTC)

Messed up interwikis

For en:Eid al-Ghadir there are several messed up interwikis. E.g. Azerbaijani Wikipedia equivalent Qədir-Xum which is about the same holiday is not linked to Eid al-Ghadir and its interwikis are different from those at Eid al-Ghadir (e.g. Spanish Ghadir al-Jumm linked to Qədir-Xum is not the same thing as Eid al-Ghadir). Could someone sort those irrelevant or possibly duplicate interwikis? Brandmeister (talk) 14:29, 29 July 2021 (UTC)

I have moved the site link to az:Qədir-Xum. Has that fixed everything or are there any others which are wrong? — Martin (MSGJ · talk) 21:13, 29 July 2021 (UTC)
Looks ok, thanks. Brandmeister (talk) 06:29, 30 July 2021 (UTC)

Get notifications for a single discussion

Hello, all.

Soon (early August), the Beta Feature for "Discussion tools" here will be updated. You will be able to subscribe to individual sections on a talk page at more wikis. If you enable the Beta Feature, then you will get this. Otherwise, you won't see it.

You can test this now by adding ?dtenable=1 to the end of the talk page's URL. For example, if you click on https://www.wikidata.org/wiki/Wikidata:Project_chat?dtenable=1 you will see new [subscribe] buttons. If you click to subscribe to this thread, then every time someone adds a new comment, you will get a notice via Special:Notifications. (It won't annoy you with separate notifications for typo fixes or additions to comments, just for new comments.)

I'll be subscribing to this thread, so please feel free to subscribe and reply here, if you want to test it out.

I have found this especially helpful for cross-wiki communications, so I have asked the Editing team to prioritize Wikidata and Commons for this feature. I am very interested in learning what you all think, and if there are changes that would help you. You can reply here, ping me to another page, or post your thoughts to mw:Talk:Talk pages project/Notifications (the central page for this feature).

Thanks, Whatamidoing (WMF) (talk) 19:44, 30 July 2021 (UTC)

New project on meta against cross-wiki promotional spam

Hello,

Not sure it's the best place to write this, please tell me where I should send the message!

A small group of wikipedians started a new meta project to keep promotion in check. m:Wikiproject:Antispam

We want to coordinate efforts against cross-wiki spam for "reputation" and (self) advertisement. And looking for contacts on projects, people who will tell their communities when there's a promotional campaign.

Maybe you've already seen a few single-purpose accounts create simultaneously many translations of the same promotional content. Please let us know about suspicious edits-- or you can also join us!

Best, --Bédévore (talk) 21:40, 30 July 2021 (UTC)

Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2021/07&oldid=1474774968"