Wikidata

Wikidata:Project chat/Archive/2021/11

< Wikidata:Project chat‎ | Archive‎ | 2021

Type constraint issue with software

ScummVM (Q145568) has a warning on each of its versions that say

type constraint

Entities using the software version identifier property should be instances of one of the following classes (or of one of their subclasses), but ScummVM currently isn't:

• software

However its entry for instance of (P31) (synomym "is a") includes the value software (Q7397). So I don't understand how ScummVM "isn't" a software. Is the type constraint not working properly, or is there a different issue? –Thunderforge (talk) 18:55, 1 November 2021 (UTC)

@Thunderforge: this seems to have resolved itself. sometimes there's caching delays that cause these warnings even after the issue is resolved. BrokenSegue (talk) 21:01, 1 November 2021 (UTC)
I confirm that it is now resolved. Thunderforge (talk) 21:34, 1 November 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. SilentSpike (talk) 09:20, 2 November 2021 (UTC)

Conflation needs to be teased apart

Abraham van der Neer (Q108750378)  – The preceding unsigned comment was added by Richard Arthur Norton (1958- ) (talk • contribs) at 07:09, 30 October 2021‎ (UTC).

Wikidata Map from WikidataCon 2021

 

Hi all.

Wikidata has come a long way in the last 9 years and I wanted to put this updated Wikidata Map in the Project Chat so that anyone not attending the WikidataCon birthday present session would see it.

You can play around with the map in a UI, the code can be found on GitHub and there is a Commons category with a collection of snapshots, comparisons and animations.

I recommend checking this animation out at full screen.

You can also find a static snapshot from last week in the thumbnail to the right.

·addshore· talk to me! 17:51, 31 October 2021 (UTC)

@Addshore: What's with the square grid in Mexico? And 100% Bangladesh? 71.204.166.188 23:05, 31 October 2021 (UTC)
I have no idea, I can say that for Mexico it appeared between 2018 and 2019, or at least that's when it became very apparent. For Bangladesh you can see part of it really light up between 2018 and 2019, and the rest fills in between 2019 and 2020. If you lower the intensity to 1 its not solid white (one of the reasons the intensity is important). On intensity 1 a pixel will only appear as white if it has 100 items stacked up on top of it. On the default intensity of 5 you would only need 20 to reach a white pixel.
·addshore· talk to me! 15:20, 1 November 2021 (UTC)

Which properties should I use for setting the main topic of a recurring event (Q15275719)?

I've identified two potential properties for storing that information

The following is a list of examples where I would use that property

I'm not sure which property I should use.

I've already asked this question in the Wikidata Telegram channel, but I just thought that I could get more insight here. From now on, I will be posting similar questions only here so as to people with the same questions can find them.

I'd be more inclined to main subject, but I can see a case for field of work. Note well, though, that it is not an exclusive choice. If in real doubt, add both. Duplication of values across multiple properties is not of itself a bad thing; if it's 50:50, then duplication serves the user who seeks for info based on main subject and serves the user seeking for info via field of work. There isn't a downside. --Tagishsimon (talk) 05:39, 1 November 2021 (UTC)

Specific properties vs qualified properties

Are there any guidelines or principles when choosing between specific properties and more generic properties with appropriate qualifiers? For example:

How are decisions made on an appropriate data model? What factors are in play? For example, is there a deliberate trade-off being made between model consistency and query complexity or speeds? Ease of access from client wikis via Lua? If so, are there objective measures that can be used to decide? Inductiveload (talk) 15:22, 1 November 2021 (UTC)

I think generally, if a specific property exists for a conceptual relation, it should be used; constraint checks can be much more precise in such cases, for one thing. Generic properties of course need to be used with qualifiers when we don't have such a property. ArthurPSmith (talk) 16:57, 1 November 2021 (UTC)

Wikidata weekly summary #492

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • WikidataCon! \o/
    • Continued the work on the first version of the Mismatch Finder. We are getting closer to the polishing phase now and will have something ready in the next weeks.
    • Lua access for Lexemes is now ready to test on English Beta Wiktionary.
    • Concluded work on the improved behind-the-scenes system for notifying Wikipedias and co about Wikidata edits that affect them. Nothing should have changed for you.
    • Started working on a new implementation of the search box to be ready for the upcoming skin changes the WMF is working on.

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.

English vocabulary

Hello, when dealing approximations with dates, are "approx. approximately approximate value about" correct wordings in English ? I mean circa(Q5727902) is not equivalent to approximately(Q60070514) and confusions should be avoided. Bouzinac 💬●✒️●💛 22:34, 1 November 2021 (UTC)

Shortcuts are eligible for items or not?

^^ Liuxinyu970226 (talk) 02:18, 2 November 2021 (UTC)

Probably. It's possible to construe them as being within WD:N criteria 1 (or, at least, the do not seem to be explicitly excluded). IMO a reasonable case can be made under WD:N criteria 3, since their being on WD would allow for analysis, Listeria reports, &c. Might be an interesting issue of how shortcuts items would work ... WP:THIS on wikipedia_1 very possibly does not go to the same sort of place as WP:THIS on wikipedia_2, so do sitelinks live on the same item b/c same string value, or on different items b/c different target. --Tagishsimon (talk) 02:28, 2 November 2021 (UTC)

Can we remove all the PDB structure ID (P638) from small molecules / ligands?

I believe linking occurrences of small molecules such as nicotinic acid (Q134658) to PDB is against the point of P638. Can we:

  • get ProteinBoxBot to stop adding them;
  • find out some way to revert the additions; and
  • somehow add a constraint to warn people about this?

For people who do want to see the molecules in PDB, we have PDB ligand ID (P3636).

--Artoria2e5 (talk) 11:27, 30 October 2021 (UTC)

@Artoria2e5: probably it'd better to contact bot operators directly or write on the bot discussion page. Or at least ping them. Wostr (talk) 18:57, 30 October 2021 (UTC)
Pinging Andrawaag. --Artoria2e5 (talk) 14:43, 31 October 2021 (UTC)
* @Artoria2e5: The edits in the examples are 5 years old and that specific bot should not be running. Do you have an example of recent edits? I will dive into this myself as well.
* Reverting/Removing those statements with a bot is straightforward, once the candidate QIDs are identified. How can we identify those items that are in need for such an update? --Andrawaag (talk) 09:01, 2 November 2021 (UTC)
I agree with this request, in general. PDB Ligand databases point to all PDBs so if there is a ligand entry, the PDBs are unnecessary. The constraint to be added would be straightforward. --SCIdude (talk) 07:19, 31 October 2021 (UTC)

Meet the new Movement Charter Drafting Committee members

More languages • Please help translate to your language

The Movement Charter Drafting Committee election and selection processes are complete.

The committee will convene soon to start its work. The committee can appoint up to three more members to bridge diversity and expertise gaps.

If you are interested in engaging with Movement Charter drafting process, follow the updates on Meta and join the Telegram group.

With thanks from the Movement Strategy and Governance team--YKo (WMF) (talk) 04:41, 2 November 2021 (UTC)

Statements that stopped being true at an unclear time

Hi, Wikipedian here trying to make some Wikidata edits. The UK company Scotsman Group (Q20877059) changed its name from the G1 Group at some point in the past year. While the fact that the name changed is well supported by various sources, I can't find a great source for when exactly the change happened, other than by doing things that border on original research (like looking at Companies House filings, or looking at the Wayback Machine to see when the website changed).

That leaves me with a problem: as best as I can tell, the correct way to "remove" the company's old logo is to add an end date to that statement. But I'm not sure what to put as the end date: do I put a best guess, even if it's poorly sourced (Companies House, the UK's register of companies, shows the legal name change happening on June 1); do I put a vague-but-probably-correct date (such as just the year 2021); or is there some way to mark a statement as "outdated, but we don't know when"?

Thanks! Gaelan (talk) 12:59, 29 October 2021 (UTC)

If you don't know the exact end date, you can use earliest end date (P8554) and latest date (P1326) as qualifiers to give a general idea when the statement stopped being true. --Shinnin (talk) 13:11, 29 October 2021 (UTC)


(EC) Quite a lot of ground to cover:
  • 'things that border on original research' are okay. This is not WP. A deep dig into companies house records might elicit a document which serves as a solid reference, and if so, that's a good thing.
  • You say 'Companies House, the UK's register of companies, shows the legal name change happening on June 1'. I'd be inclined to use June 1 and a reference pointing back to the CH page which specifies same
  • But yeah, situation in which you do not know. You could put a broad date for end time (P582) (20201; 2020s; 2000s) and then perhaps some more pointers using latest date (P1326) & earliest date (P1319)
  • Beyond use of a broad date, I dunno. Users will, presumably, be looking for end time (P582) so it follows that we need to do something with that.
  • Is it better to have a guesstimate end date with no reference, than no end date, in a situation in which it's clear something has ended? Probably. One could use maybe use inferred from (P3452) to point to a value which sets out the basis for the date. --Tagishsimon (talk) 13:20, 29 October 2021 (UTC)
  • Maybe it's worth repeating the basics: while the qualifier property:P582 is labelled "end time", one can't actually enter a precise time of the day, but only day-, month- or year-precision values. "end year" is a perfectly valid value and for many purposes, it's actually the most useful one. See Help:Dates for more. --- Jura 08:47, 30 October 2021 (UTC)
The special value unknown value Help is designed for that. But for a date an a low precision is often a better choice. author  TomT0m / talk page 12:26, 2 November 2021 (UTC)

How do I delete a redirect?

Hi everyone, I accidentally merged Q16575483 (previously "mantra yoga" in Italian) into Q78089383 (interventional study) instead of Q31086271 (mantra yoga). I tried to undo the merge, but it didn't restore the item, so I manually re-created the properties of Q16575483 in Q31086271. But Q16575483 still redirects to Q78089383 (interventional study). How do I delete that redirect and ideally manually create the redirect to the correct item? -- Sparklingbright (talk) 13:03, 2 November 2021 (UTC)

You should be able to undo the edit that was made to the target, and then undo the edit that created the redirect! ·addshore· talk to me! 13:16, 2 November 2021 (UTC)

Neither the undo nor restore operation seemed to address the redirect issue. Is there any way to do this manually? -- Sparklingbright (talk) 16:50, 2 November 2021 (UTC)

@Sparklingbright: Rolling back was hung up on the fact that Q78089383 was carrying the it: sitelink which was also on Q16575483, the solution for which was to temporarily remove the sitelink from Q78089383 such that Q16575483 would roll back; then delete the sitelink from Q16575483 and readd it to Q78089383. I've rolled the redirect back; please check that you are happy with the statements & sitelinks on all of the items, and we're done. --Tagishsimon (talk) 18:14, 2 November 2021 (UTC)
@Tagishsimon: Thanks for the explanation and for the fix, looks good to me now! -- Sparklingbright (talk) 20:25, 2 November 2021 (UTC)

Wikidata will be on read-only for a few minutes on 11 November, 06:00 UTC

A maintenance operation will be performed on Thursday November 11, starting at 06:00 UTC. It should only last for a few minutes. During this time, saving changes will not be possible at Wikidata.

For more details about the operation and on all impacted services, please check on Phabricator.

A banner will be displayed 30 minutes before the operation.

Please help your community to be aware of this maintenance operation.

Thank you! Trizek (WMF) (talk) 15:51, 2 November 2021 (UTC)

add better source (Q105160679)

The place of birth was added to the data object Hans-Peter Hannighofer (Q105160679) with a Wikipedia source. I would ask that this source either be replaced with the following source or that the following source be added. In the article WM-Debüt mit Bronze is the following sentence: „Hannighofer begann als Leichtathlet in seiner Geburtsstadt Arnstadt und in Erfurt beim ASV.“ (“Hannighofer began as a track and field athlete in his hometown Arnstadt and in Erfurt with the ASV.”) The Google Translator has now translated „Geburtsstadt“ to “hometown”, which may not be the perfect translation. „Geburtsstadt“ should rather be translated as city of birth. --Gymnicus (talk) 12:24, 3 November 2021 (UTC)

  • Why did you add the year precision date with deprecated rank? If there is a more precise date, it would generally have normal rank and the other preferred. --- Jura 13:03, 3 November 2021 (UTC)
    @Jura1: That surprised me too, but judging by the version history, I did the editing. Since I don't do this normal like that, it must have happened by accident. It is of course true that the more accurate date gets the highlighted rank and the less accurate date gets a normal rank. --Gymnicus (talk) 13:21, 3 November 2021 (UTC)

Merging Q3480844 and Q97690891

The two data objects Edda Garðarsdóttir (Q3480844) and Edda Garðarsdóttir (Q97690891) should be merged. At first glance, it probably seems difficult to see it. But if you look at the identifier Soccerdonna player ID (P4381) in the data object Edda Garðarsdóttir (Q97690891), you will find the information that the player played in the country game against Germany on October 14, 1999. This game can also be found in the KSI database under the link [1]. With the KSI link and the link [2] from Soccerdonna, where this international game is also described that the KSI profile [3] and the Soccerdonna profile [4] are about the same lady. The data object Edda Garðarsdóttir (Q3480844) does not contain the KSÍ player ID (P6495), but it does in the linked German Wikipedia article. --Gymnicus (talk) 13:15, 3 November 2021 (UTC)

Contested merge: Maarten van Dis

The October 2020 merge of Q2180675 and Q99519260 is being discussed at en:Wikipedia:Reference desk/Humanities#Did an obscure Dutch rower take this photograph?. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:17, 3 November 2021 (UTC)

User:BeneBot*/RfD-stats

Bot’s counting requests became strange. Bot holder isn’t active since 2019. 217.117.125.83 17:10, 3 November 2021 (UTC)

I deflagged the bot since Bene* indeed has not edited any Wikimedia projects since 2019. The bot was inactive since 2016 anyway.--Ymblanter (talk) 17:35, 3 November 2021 (UTC)
The page is usually being updated by User:Pasleim's User:DeltaBot after it added templates to the WD:RfD page itself. The script is here. If I execute it from Toolforge manually, it does indeed work as expected. Not sure which issues DeltaBot has. —MisterSynergy (talk) 18:51, 3 November 2021 (UTC)

Chess moves database

The world needs a searchable editable database of chessboards, pieces and the next move and description. This will provide a simple way to play chess by easily searching for matching chess boards and pieces and evaluating a description of the next move. Allow a computer hours to find the next move and then record this move in the database. The next time this chess board is queried a description and move will be provided in seconds. Eventually, the Wikidata:chess fills up with nearly all possible chess moves.

Umm, Shannon number (Q1953800) much? --Tagishsimon (talk) 02:16, 2 November 2021 (UTC)

And of course this only appears to be undoable; the main point, I see, involves the computer working out the next move for hours; Basically that move then become the preferred move and in a clever insight most other possible moves are labeled as unwise without taking up any storage space. The first iteration allows only for the best possible move. The second iteration allows for the two next best moves. This idea goes into the future when molecular storage and quantum computing becomes a reality. So we start today building the basics. In the future, two dimensional chess will become trivial and we will be playing three dimensional chess.

Claude gently spins in his grave. --Tagishsimon (talk) 16:56, 2 November 2021 (UTC)

So to begin, I suggest starting at a point that is trivial -- Tic-Tac-Toe -- and learn from it. Then progress to the next nearly trivial game -- checkers --. Then on to chess. I know the benefits here out weight the negatives. Children learning tic-tac-toe at age 6 do not consider it trivial but by age 10 the game becomes trivial. Let me start a Tic-Tac-Toe moves database.

Yes, I agree. This is not the place for these ideas--Withdraw. Thank you for you consideration. SageOfNotMuch (talk) 01:02, 4 November 2021 (UTC)

Tic-Tac-Toe moves database

Let us begin the journey of solving chess by starting with the trivial game of Tic-Tac-Toe. I am in the process of creating an open source App using Xcode and SwiftUI for the Apple App Store which will edit tic-tac-toe moves. Once completed, I will then code a search method that will seek out and display all the available next moves; with the winning move at the top of the search list. Wikidata:Tic-Tac-Toe displays web pages of next moves.

No it does not. See WD:N. Perhaps get yourself your own wikibase. --Tagishsimon (talk) 17:14, 3 November 2021 (UTC)

Yes, I agree. This is not the place for these ideas--Withdraw. Thank you for you consideration. SageOfNotMuch (talk) 01:03, 4 November 2021 (UTC)

JupyterHub / PAWS

I tried following the instructions in the JupyterHub docs but I couldn't get them to work.

I was looking for a way to remotely start up a server and if possible have a script run automatically upon server start as well. What I've tried so far is to generate a token and use curl to call the REST API. I get a HTTP 302 "Found" message. 

curl -v -X POST -H "Authorization: token muhsekrittoken" "http://hub.paws.wmcloud.org/hub/api/users/Infrastruktur/server"

But no servers will start. I was hoping someone here might be of assistance. --Infrastruktur wdt:P31 wd:Q5 (T | C) 15:38, 3 November 2021 (UTC)

Issue resolved. --Infrastruktur wdt:P31 wd:Q5 (T | C) 22:23, 4 November 2021 (UTC)

GeoNames in Canada

Hi guys! In GeoNames ID (P1566) big problem with code: Wikidata:Database reports/Constraint violations/P1566#Canada (Q16). Need help! Машъал (talk) 16:24, 4 November 2021 (UTC)

Probably because GeoNames is festering heap of complete garbage. It & the cebwiki / svwiki imports should be thrown off WD, entirely, ASAP. Right now it's just poisoning the well. --Tagishsimon (talk) 16:29, 4 November 2021 (UTC)
And United Kingdom: Wikidata:Database reports/Constraint violations/P1566#United Kingdom (Q145)… Машъал (talk) 16:37, 4 November 2021 (UTC)

Here, just to give an illustration of how poor GeoNames is, are GeoNames coords for the Shetland Islands. 56 of 293 have already been marked as deprected becuse they do not point at the object they claimed to represent - https://w.wiki/4LSE . A couple of minutes spent with the map shows that many more, which still have normal rank, are also wrong - they plain do not point to the headland or island they're meant to mark; or else they're very poorly positioned relative to the subject - https://w.wiki/4LSJ . So, if, conservatively, 25% of the data we're acquiring from this source is just plain wrong, why would we continue to give it space in the database? Are our standards that low? Any old crap so long as we have a record - never mind its contents? --Tagishsimon (talk) 17:43, 4 November 2021 (UTC)

Best practices for sourcing values of country of citizenship (P27)

Lately I've been adding references to claims about country of citizenship (P27) by including place of birth (Q1322263) as a value for inferred from (P3452). Is that something that's typically done and/or advisable? Express statements about someone's country of citizenship are rare. Sometimes library authorities say "associated country", but that's not the same thing (may be work location, longtime residence but not country of citizenship, etc). Inferring citizenship from place of birth is the best I can usually do. AleatoryPonderings (talk) 21:39, 4 November 2021 (UTC)

Please use based on heuristic (P887)inferred from place of birth (Q91770864) instead. Saying inferred from (P3452)place of birth (Q1322263) implies that the source of the citizenship information is the place of birth (Q1322263) item itself (it would be akin to putting reference URL (P854)https://www.wikidata.org/wiki/Q1322263).
--Quesotiotyo (talk) 02:19, 5 November 2021 (UTC)

Another editor being extremely rude

I am having difficulty with another editor who is accusing me of vandalism for no good reason. When I complained she became aggressive and reverted my work, again for no good reason. Please can anyone advise what I can do about it. AmirahBreen (talk) 23:40, 3 November 2021 (UTC)

Having dippied into https://www.wikidata.org/wiki/Topic:Wjnv0wdtbvidfcgw it seems to me that you are being extremely rude, AmirahBreen. Perhaps best if you step away from wikidata for an hour or so and think it over --Tagishsimon (talk) 23:44, 3 November 2021 (UTC)
Oh, so it's not rude to accuse someone of vandalism then, is it? Why did this user accuse me of vandalism? I have not vandalised the project at all. If she had not been so rude and accusatory she could have helped me to understand if there is some rule on this wiki I was unaware of. She has been rude and accusatory from the very start of her interaction with me. AmirahBreen (talk) 00:35, 4 November 2021 (UTC)
tbh, AmirahBreen, we try not to do drama on wikidata. Just give it a rest. --Tagishsimon (talk) 00:39, 4 November 2021 (UTC)
Isn't it dramatic to accuse an inexperienced editor of vandalism just because they made a mistake and didn't know the rules and then when they try to tell you that they were not vandalising accuse them again and start a wiki war with them. AmirahBreen (talk) 00:44, 4 November 2021 (UTC)
I added another crop now, explained it on the discussion page and reset the bad picture. Just to see, that the reset is reverted again after my start of the discussion on the talk page (an no, I don't wanna read any more accusations on my talk page - this one is stopped by me). This should be re-reverted by someone else. Mirer (talk) 23:58, 3 November 2021 (UTC)
And another edit in one of the old pictures and another false accusation on my talk page. My AGF is running out fast. Mirer (talk) 00:08, 4 November 2021 (UTC)
Exactly, that is what it is an 'old picture'. Infact the picture is very old and I have uploaded a much more recent one. If you have any problems with that then please stop reverting my edits and discuss on the talk page to reach consensus. AmirahBreen (talk) 00:24, 4 November 2021 (UTC)
I didn't revert ... just tried to reset the pictures to "normal" after you disqualified them and tried to add a new one and started the discussion. But after seeing this discussion[5] about your COI to the object, I'm not wondering anymore.
Could please an admin reset the pictures until consensus is reached? Mirer (talk) 00:41, 4 November 2021 (UTC)
There is no conflict of interest to the subject. The discussion you have linked to is an old discussion which established that there is no conflict of interest. AmirahBreen (talk) 00:48, 4 November 2021 (UTC)

Isn't it a case of trolling ? Wouldn't it warrant a temporary block ? Kpjas (talk) 10:42, 4 November 2021 (UTC)

Agreed. Support short block of User:AmirahBreen — Martin (MSGJ · talk) 12:35, 4 November 2021 (UTC)
The worst picture (inserted by a user with a clear cut COI) is still featured in the object Q270783. On the discussion page there is still just trolling and filibustering and insulting me with another set of links to guidelines. This has to be ended by other users/admins!
The only thing I set back was the rank of "deprecated" for the picture which was featured/prefered till yesterday. I only set it to normal and let the new picture (of low qualitiy) be the prefered one to prevent another edit war by the user.
I'm sorry, but other have to step up and decide ... Mirer (talk) 09:03, 5 November 2021 (UTC)

I have reverted the item to the state it was in before this edit war. I ask User:AmirahBreen to refrain from amending the image unless they can demonstrate clear consensus to do so. The issue here - apart from AmirahBreen's conduct, and the general indifference of the WD community to this particular edit war - is that WD does not, afaik, have any policy on image choice for P18. In the instant case, the dispute appears to be one between the use of a high-resolution and IMO aesthetically pleasing but dated image; and a more recent image which is of low resolution and of (again IMO) poor composition and lighting. I cannot see any compelling reason for favouring what is, date aside, a very poor image over a very fine image. Nor am I persuaded that there is any particular merit in the item holding a more up-to-date image than an older image.

Consensus on image choice not having been reached on the talk page, this page, watched by many more eyes, becomes the favoured place for the discussion. To date there has been zero support for AmirahBreen's preferred solution, and several editors have objected to AmirahBreen's conduct in pursuing the matter. I expect AmirahBreen to respect that no-one has expressed any support for their passion for changing the image, and to understand that there is consensus against edit wars. Amending the image choice again, without evidence of consensus, will be a purely disruptive act essentially calling for a block; something I hope AmirahBreen will wish to avoid. Courtesy ping @AmirahBreen: --Tagishsimon (talk) 10:21, 5 November 2021 (UTC)

Aside from the case here, you're right that we don't have (and very well may need) a policy on selecting pictures for objects. As far as persons are concerned there is a "project" (well a big category, but unluckily lacking policy as well as here ... but at least with some sort of common understanding) on commons, which tags pictures to be cropped for Wikidata, so that the shown person is as good recognisable as possible (good picture, showing as much face as possible). This supports our object pages here (you see and recognise a person) as well as the infoboxes in most other wikis. If a special wiki/person wants another picture in his box/article the only important thing for is, is to provide it (ideally with a date and a description - so filtering is possible). We can not change our picture with every new wish of every new wiki - that has to be done there.
So it might be time to discuss a policy for a) which pictures do we want here (and how many), b) which one should be featured (new vs. recognisable vs. typical) and c) which additional informations should be provided (like year) if there is more than one.
I'm not sure if and where we should start that discussion. Mirer (talk) 12:00, 5 November 2021 (UTC)
I understand what you have explained about the resolution of the image. Beforehand I was simply being unfairly accused of vandalism by an editor who appeared to me to have article ownership issues. I also understand what Mirer is saying about policy and agree with it. If you are writing a policy, please also include in it some policy on issues about using images which have been edited in some way from the original. The original of this image (which can also be seen on Wikimedia Commons) had poor light quality, and the image has been edited so it does not show a natural appearance of the subject. This issue has already been discussed on Wikipedia, and consensus was eventually reached to use the image I uploaded here which is currently in use on Wikipedia too. Perhaps resolution does not matter so much there as it does here. IMO the image here could still be improved upon by replacing it with a high resolution image which meets the criteria you have described and hasn't been edited in such a way so as to alter the persons complexion. If I come across a better image I will discuss on the article talk page.
Both Mirer and yourself are still accusing me of 'indifference' to an edit war which she started by reverting my goodfaith edit with no other explanation than the tag of vandalism without proper explanation. Mirer has also accused me again in this discussion of conflict of interest, despite the fact that I have already explained that I do not have a COI. In the absence of a policy about COI here on Wikidata, then I suggest you see the policy on Wikipedia, which says that if you believe an editor has a COI you should raise the matter first on their talk page (not in a noticeboard discussion) and give them a chance to respond to it. Repeatedly accusing an editor of COI without good reason and on noticeboard discussions without having made any attempt to discuss the issue first on their talk page could also be viewed as trolling [6].
I would like to remind you again to Wikidata:Assume good faith. By chasing new editors away who have done nothing wrong but making a good faith edit you are doing nothing to improve Wikidata. - AmirahBreen (talk) 13:00, 5 November 2021 (UTC)
There is a point where the complaining and asking for AGF should stop and one himself should be able to come down, discuss about the issue and give others some AGF too. This isn't a one way street. Mirer (talk) 18:29, 5 November 2021 (UTC)

little help with merging

Please could someone have a look at Wikidata:Database reports/Descriptions with Q and search for the word "please" on that page? You'll find a couple of items that could (and should) easily be merged. Thanks a lot! 2003:E5:370C:CA00:D9A6:8EFC:9C93:673B 19:36, 5 November 2021 (UTC)

Pląskowski-Oksza (Q16588693) what can be done ?

This item Pląskowski-Oksza (Q16588693) has been incorrectly designated as a family name. In fact it is a Polish noble family as the corresponding article in the Polish Wikipedia states.

The family name is Pląskowski (Q28363589)/Pląskowska. To complicate matters even further there's also an item Pląskowski-Oksza (Q63531223) which seems to be more or less a duplicate of Pląskowski-Oksza (Q16588693).

What would be a proper course of action in this case ? Kpjas (talk) 10:38, 4 November 2021 (UTC)

  • About the family name: If the sitelink is on the wrong item, create a new item and move it there. If nobody uses the spelling of Q16588693 list it for deletion.
About the family: It's somewhat random for which families we have items and for which ones we don't. Generally this follows Wikipedia articles, but if this can be referenced somehow, that would even be better. The main problem is that we could end up with countless family items (e.g. one for each couple). --- Jura 09:01, 6 November 2021 (UTC)
@Jura1:. Thanks for your answer. BTW As far as the Polish Wikipedia goes, articles and corresponding Wikidata items are created for noble families or families that date back many generations, so not likely that they would make up a very large number of items.

Connecting house and street?

What are currently the best practices for

  • Indicating at the item of the building which street it is on (I am aware of Property:P6375, this is not what I am looking for; I am asking about the best property which I can use on Q2497365 to indicate it is located on Q109392385);
  • Reciprocal relation: Do we currently indicate which buildings are located on a certain street, and, if yes, which property do we use on Q109392385 to indicate that Q2497365 is located on this street?
Thanks--Ymblanter (talk) 12:22, 5 November 2021 (UTC)
Use simply located on street (P669). I am not aware of reciprocal relations, use queries for that (see User:Jklamo/Domy/Ulice for example). --Jklamo (talk) 12:29, 5 November 2021 (UTC)
I use located on street (P669) a lot. See i.e.
SELECT ?Nyhavn ?NyhavnLabel ?billede WHERE {
   SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
   ?Nyhavn wdt:P669 wd:Q943946.
   ?Nyhavn wdt:P31 wd:Q41176.
   OPTIONAL { ?Nyhavn wdt:P18 ?billede. }
}
Try it!

, which I personally find quite usefull.--Hjart (talk) 12:37, 5 November 2021 (UTC)

Great, thanks a lot.--Ymblanter (talk) 12:59, 5 November 2021 (UTC)


@Jura1: Sorry, but [9] is not the correct way of using street address (P6375). According to property description, you should indicate "building number, city/locality" and "post code" which is missing in your exemple. Ayack (talk) 16:32, 5 November 2021 (UTC)
I just restored what was there, but yes, please complete. --- Jura 09:47, 6 November 2021 (UTC)
Thanks, qualifier is a good idea as well.--Ymblanter (talk) 09:30, 6 November 2021 (UTC)

I added an allowed qualifiers constraint (Q21510851) to public key fingerprint (P3721). I'm trying out Wikidata:Be bold

I'm usually not very bold. I'm mostly unsure about a lot of things. This is the first time in my life that I've ever changed a Wikidata property in any form. I added the email address (P968) to allowed qualifiers constraint (Q21510851) for the property public key fingerprint (P3721) and I added my comments on the talk page for the property. Email data is included in many OpenPGP (Q2141493) public keys. Getting an error entering such data would make no sense to me considering that the data exists and is interrelated to the other data in the key. If the community thinks email data is too sensitive then maybe we should not enter such data but right now I have not seen anything about that we should not include such data in relation to public keys. If somebody has an opinion then please let me know. Oduci (talk) 16:19, 6 November 2021 (UTC)

seems reasonable. BrokenSegue (talk) 14:25, 7 November 2021 (UTC)

Elsevier Digital Commons

https://digitalcommons.augustana.edu/cgi/viewcontent.cgi?article=2523&context=swensonsag and https://www.elsevier.com/solutions/digital-commons/how-it-works Does anyone know what license this commons is released under. I have been following a half dozen links from one of the published papers, and still cannot figure out if these are published under some creative commons license, of if they are just using "commons" as a marketing term. Does anyone know more? I emailed them a week ago, no one responded. --RAN (talk) 23:04, 7 November 2021 (UTC)

I don't know, but at least previously it would seem Elsevier used licenses that were more restrictive than Creative Commons. The following is from enwiki:

Some publishers (e.g. Elsevier) use "author nominal copyright" for OA articles, where the author retains copyright in name only and all rights are transferred to the publisher.[1][2][3]

HTH. --Infrastruktur wdt:P31 wd:Q5 (T | C) 01:52, 8 November 2021 (UTC)
Referenzen
  1. ↑ Morrison, Heather (2017). "From the Field: Elsevier as an Open Access Publisher". The Charleston Advisor. 18 (3): 53–59. doi:10.5260/chara.18.3.53. hdl:10393/35779.
  2. ↑ Pablo Alperin, Juan; Rozemblum, Cecilia (2017). "The Reinterpretation of the Visibility and Quality of New Policies to Assess Scientific Publications". Revista Interamericana de Bibliotecología. 40: 231–241. doi:10.17533/udea.rib.v40n3a04.
  3. ↑ "Open Access Survey: Exploring the Views of Taylor & Francis and Routledge Authors". 47.

Will we eventually be able to see from Wikidata when an entry is used as Structured Data at Commons

I may have asked another time. Will we eventually be able to see from Wikidata what is used in Structured_Data (SD) at Commons? We can see at Commons what exists at Wikidata when we link it, will we eventually be able to see from Wikidata when an entry is in use at Commons? Will use at Commons prevent deletion at Wikidata in the future? --RAN (talk) 13:33, 4 November 2021 (UTC)

I don't think that this would scale well. There are currently 2677 Files in Commons that depicts (P180) house cat (Q146). Why and how would you want to make this visible in Wikidata? However, there's a Ticket for a Module that can display structured Data from Commons in Phabricator -- Dr.üsenfieber (talk) 21:14, 4 November 2021 (UTC)
@Dr.üsenfieber @Richard Arthur Norton (1958- ) A link to an appropriate Commons Query Service from each item would be a good start. Maybe we could deliver this link to advanced users with a user script? Vojtěch Dostál (talk) 20:26, 7 November 2021 (UTC)
  • It's something that in the pipeline of the SDC team, but apparently not a priority. So contributors at Commons have to live with occasional deletions for now. --- Jura 09:49, 6 November 2021 (UTC)
Thanks all! The ability to see use-at-Commons from Wikidata may not be useful for cat and dog images, but will be useful for obscure people appearing in images, or mentioned in news articles, or as authors of obscure books scanned at Commons. --RAN (talk) 20:33, 7 November 2021 (UTC)
At least for https://www.wikidata.org/w/index.php?title=Q146&action=info , it doesn't matter how many uses there. --- Jura 20:46, 7 November 2021 (UTC)
@Richard Arthur Norton (1958- ):, I'm not sure is that is what you want, but you can try my modification of User:Lockal/Gadget-relateditems.js (how to add). It is the same as relateditems (Q102435390), but it also adds now the second button to get inverse statements from Wikimedia Commons Query Service and gives queries like [10]. --Lockal (talk) 14:18, 8 November 2021 (UTC)

Issue with item used as catalog?

It seems Colección Patricia Phelps de Cisneros (Q27430435) is often incorrectly used for property catalog (P972). See here. It seems like this is a mess that someone should clean up, but I'm not sure of the best place to raise it. Thanks! Calliopejen1 (talk) 01:09, 6 November 2021 (UTC)

If the issue is that humans should be linked to Colección Patricia Phelps de Cisneros (Q27430435) via has works in the collection (P6379) rather than catalog (P972), that's an easy fix. But right now that's just an assumption on my part: Calliopejen1 - is that the case? --Tagishsimon (talk) 13:21, 6 November 2021 (UTC)
I'm not sure what was intended here. Tagging Staboo who set this on Marisol (Q438248), BrillLyle who set this on Gyula Kosice (Q183438), and VEOP2 who set this on Ferdinand Konrad Bellermann (Q317888). Could one of you explain what was intended? They all seemed to get set in 2017 so presumably a part of some sort of coordinated activity. Calliopejen1 (talk) 17:14, 8 November 2021 (UTC)
Actually it looks like these are still being set, see e.g. Yeni & Nan (Q42886590) from today that uses both catalog (P972)=Colección Patricia Phelps de Cisneros (Q27430435) and catalog (P972)=WIKIarte (Q36788585). Calliopejen1 (talk) 17:16, 8 November 2021 (UTC)

The Mexican Grid

There's a marked misdistribution of coordinates situated in (rural, in particular) Mexico. My current guess is that there was confusion between arc minutes (ranging from 0 through 59.999... in a degree) and hundredths of a degree; that seems both a likely error to have happened along the way, and according to my eyeballing it matches the actual distribution of points.

This query shows the pattern: Coordinate positions with [6789] as their first post-decimal-point digit seem much less likely.

San Francisco (Q50009072) would be an example of a human settlement that seems situated improbably. By my theory, it's actually meant to be located here.

My suggestion would be to figure out where those data items come from, fix the coordinates, then modify all the claims that are still current in their incorrect version. I've yet to figure out how to do any of that, but it seems a fun project... Streetmathematician (talk) 17:28, 6 November 2021 (UTC)

Where they come from is easy: GeoNames; in this case, by the looks of it, via Serbian or Serbo-Croat wikipedia. But the source is GeoNames. Your observation of a lack of decimal places with first digit in [6789] is well made. Likely each of the items will have an External ID pointing back to GeoNames, so there's a possibility that more spadework can be done comparing source and WD's P625 values. --Tagishsimon (talk) 18:54, 6 November 2021 (UTC)
Thanks! I'm a bit confused, because in the example above, we have three different sets of coordinates:
1. Wikidata currently says 20°14' N
2. GeoNames has 20°27' N
3. INEGI has 20°23' N
None of the three sources say they've been modified. GeoNames claims INEGI as a source. The Wikidata coordinates correspond to the INEGI coordinates with my hundredths-of-degrees-as-minutes theory, but not the GeoNames coordinates.
I'm currently querying the INEGI API to identify the incorrect points. So far (I'm using randomized order), just short of half the places (2337/4799) with INEGI coordinates appear to be incorrect.
Once that's done, we'll have a list of incorrect points, which I believe we can legally use. We'll also have a list of corrected coordinates from INEGI, but I don't read enough Spanish to establish whether we have a license to use that. Ideally, if someone can confirm it's okay to do so, we'd just correct the claims for those incorrect points; alternatively, we could correct them automatically by running them through the reverse transformation.
Points without an INEGI ID don't appear to be affected. Streetmathematician (talk) 21:24, 6 November 2021 (UTC)
Looks like INEGI is free to use, so long as the source is credited - https://www-inegi-org-mx.translate.goog/inegi/terminos.html?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=en-GB&_x_tr_pto=nui --Tagishsimon (talk) 21:35, 6 November 2021 (UTC)
Thanks again! With that out of the way, I'd like to pick an example and correct it, using the same script I intend to use on a larger batch (for confirmation), then all of the incorrect locations.
1. My intuition is to deprecate (using incorrect value (Q41755623)) rather than remove the old coordinate claims; they've been in Wikidata for a long period.
2. The correct coordinates would be added with reference URL (P854) pointing to [11] (with the obvious substitution), and retrieved (P813) set to the current date.
I'm open to any suggestions for how to do things differently, of course. Streetmathematician (talk) 10:39, 7 November 2021 (UTC)
Okay, I've now done this for La Cruz (Q61312352). I'm not sure I got it right, though, because I think retrieved (P813) should probably qualify reference URL (P854) rather than count as a reference of its own? Streetmathematician (talk) 15:28, 7 November 2021 (UTC)
@Streetmathematician: The garbage coordinates, being cited only to a particular Wikipedia (and thus of similar provenance to unsourced coordinates), can and should be removed once coordinates with better external sources come around; if the garbage coordinates were cited to an external source, they could be deprecated as you have done. As for the references, the P813 and P854 claims should be on the same reference, as I have just now adjusted on the item for La Cruz. Mahir256 (talk) 03:28, 8 November 2021 (UTC)
Like this (for Loma de Carretas (Q61283561))? Streetmathematician (talk) 06:35, 8 November 2021 (UTC)
@Streetmathematician: Yes. Mahir256 (talk) 14:39, 8 November 2021 (UTC)

Thank you! I've now made a list of places that haven't been modified recently, have a single set of incorrect coordinates sourced to the Serbian Wikipedia, and have a label (that last criterion is to get a usable initial batch size, and to deal with the most active items first). The list, with 74 entries, is at User:Streetmathematician/The Mexican Grid. So I'll go ahead and modify those, I think. (This would be an ideal time to stop me, though) Streetmathematician (talk) 15:05, 8 November 2021 (UTC)

I only skimmed the conversation but if you're gonna make a ton of changes might want to move this conversation to Wikidata:Requests_for_permissions/Bot. Though you can definitely do a batch of 74 no problem either way. BrokenSegue (talk) 15:34, 8 November 2021 (UTC)
Thanks! I wasn't sure how to proceed, actually. I've now done the 74-item batch of labelled items and I've opened Wikidata:Requests for permissions/Bot/StreetmathematicianBot. Streetmathematician (talk) 17:52, 8 November 2021 (UTC)

Geysers are classified as events. What exactly went wrong?

Right now Giant Geyser (Q1129264) and geyser (Q83471) in general are classified as all being following things:

  • en: geyser (hot spring characterized by intermittent discharge of water ejected turbulently and accompanied by steam) [12]
    • en: fountain (water) (discharge of water ejected turbulently (under pressure); (artificial or natural) water feature consisting of one or more streams of water originating from a water source) [13]
      • en: jet (stream of fluid projected into the surrounding medium) [14]
        • en: flow of fluid influenced by pressure (type of flow of fluid) [15]
          • en: flow of fluid influenced by force (type of flow of fluid) [16]
            • en: fluid flow (movement of fluid matter) [17]
              • en: motion (change in position of an object over time; a body is said to be in motion if it changes its position with respect to its immediate surroundings) [18]
                • en: move [19]
                  • en: behavior (way that one acts in different situations) [20]
                    • en: change (process, event or action that deviates from the present state) [21]
                      • en: occurrence (occurrence of a fact or object in space-time; instantiation of a property in an object) [22]

What went wrong here? It seems not utterly wrong to say that geyser is a fountain or that geyser is a jet. But saying that geyser is a behavior is definitely wrong, while current structure claims this. How it can be fixed?

To avoid XY problem: I am trying to distinguish events such as invention of telephone (Q13443401) from objects representing specific physical feature such as Giant Geyser (Q1129264) or Wawel Castle (Q18820)

Mateusz Konieczny (talk) 07:59, 8 November 2021 (UTC)

  WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

The description of water jet (Q41554881) seems incoherent, blending an event (discharge of water ejected turbulently) and a physical object (water feature). Bovlb (talk) 20:03, 8 November 2021 (UTC)

Wikidata weekly summary #493

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • Mismatch Finder: We finished adding “next steps” dialog and a “more info” dialog to tell users more about the import a mismatch was added in.
    • We had discussions about how to best make the SearchBox WVUI component for the new Vector skin work well for Wikidata (phab:T275251)
    • We have put a version of WBStack on wikibase.dev. We will start work to deploy to wikibase.cloud in the coming weeks.

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 labels of occupation always male labels?

In many languages, occupations have a female and a male label. For instance, actor (Q33999) is "actor" for male and "actress" for female. However, it seems that the Wikidata label of the item is often the male form. Then you have property female form of label (P2521) for female form and property male form of label (P3321) for the female form.

If you think about it, there is no reason to prefer the male form over the female one. This assumption that male form would be neutral is weird. Another point is that there is no label for non binary people.

Is there any discussion and decision about this topic in the history of Wikidata?

Do you think it would be better to use gender neutral labels or just propose male and female form in the item label ("actor or actress")? PAC2 (talk) 11:01, 27 October 2021 (UTC)

afaik, there is not consensus on whether establishing that e.g. 'actor' is unisex; or continuing with 'actor'/'actress' ... is the better way to go. Right now you're the one framing 'actor' as male. cf. https://www.theguardian.com/theobserver/2011/sep/25/readers-editor-actor-or-actress & many similar discussions. --Tagishsimon (talk) 11:18, 27 October 2021 (UTC)
Also, for concepts like chairperson (Q140686), firefighter (Q107711), police officer (Q384593), aircraft pilot (Q2095549) where there is a accepted gender-neutral formulation, that is indeed what is used as the label. Inductiveload (talk) 11:31, 27 October 2021 (UTC)
I guess the discussion may be of interest for   Notified participants of WikiProject Occupations and professions.
Inductiveload thanks for your examples. There are some examples of neutral labels in English but I guess that's just a few cases.
Tagishsimon Thanks fir the reference but I guess the problem is more general than just actor/actress in English. It concerns also French, German, and so on. It would be good to have a recommendation toward gender neutral labels. PAC2 (talk) 20:15, 27 October 2021 (UTC)
Don't disagree; with the caveat that it's probably an issue with its own nuances and complexities, the aim is highly desirable. Some additional words on Help:Label might be in order. There's also an issue of probable lack of diversity attaching to the image (P18) selected for occupations; something else to keep an eye on. --Tagishsimon (talk) 20:38, 27 October 2021 (UTC)

I've been investigating the issue further. According to this query there are few examples of male biased labels in English. By contrast, it seems to be a big issue in French (this query).

  Notified participants of WikiProject France. I have some SPARQL queries here User:PAC2/Gender neutral labels. PAC2 (talk) 20:04, 29 October 2021 (UTC)

I've developed a tool to count gendered neutral labels in different languages https://observablehq.com/@pac02/gendered-labels-in-wikidata. It shows that English labels are very often neutral (but we don't have much data). However labels in French, Spanish or German are much more biased toward male form. PAC2 (talk) 06:58, 2 November 2021 (UTC)
How could be go further on this topic ? Should we add a recommandation on Help:Label saying that if possible it is recommended to use gender inclusive labels (ie labels which are not male form as default and which include women and non-binary people) ? Should we create a specific Wikidata:Requests for comment on this topic to get some consensus ? Should we create a wikiproject to work on gender diversity issues in Wikidata ? PAC2 (talk) 19:36, 4 November 2021 (UTC)
I think using gender neutral labels is already pretty accepted practice. I would feel free to add that to Help:Label. I think things get tricky mainly in cases where such terms do not exist which gets especially problematic with certain languages and there isn't an obvious one size fit all solution. BrokenSegue (talk) 20:24, 4 November 2021 (UTC)
  Done Help:Label/general_principles#Use_gender_neutral_labels @BrokenSegue: PAC2 (talk) 07:58, 9 November 2021 (UTC)
@PAC2: Your tool is very useful, but it misses a common case: items where there is only a female form or only a male form. For example, it misses actor (Q33999) since that item only lists a female form ("actress") and not a male form. I would list items like that as "male generic" and items with only a male form as "female generic". Nosferattus (talk) 00:55, 9 November 2021 (UTC)
you're right. I will see how to include those cases. The point is that if you have p3321 but not P2521, you're not sure that the label is not neutral. PAC2 (talk) 08:01, 9 November 2021 (UTC)

@lokiDAcar Understandable, but in the case that we were to create genderfluid job titles, How would we get them to everyone, and how will we know that people would use them?

Oops

I created "Category:William Sinclair" here at Wikidata instead of at Commons, can someone delete it. Sorry! --RAN (talk) 19:13, 9 November 2021 (UTC)

  Done--Ymblanter (talk) 19:37, 9 November 2021 (UTC)

Content migration from one file format to another

Wikidata specifies the file formats a software tool can read (readable file format (P1072)) and write (writable file format (P1073)). But the digital preservation community would need a way to describe migration pathways (e.g., tool X can migrate content from format Y to format Z with no or little information loss in the process).

How can we express this?   WikiProject Informatics has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

Example: we made an experiment about converting (layered) PDFs to JPEG (whether this migration is relevant and what we lose in the operation are other questions!) and found that PDFBox gives pretty good results.

Proposal: use writable file format (P1073) to express the target format, and qualifiers readable file format (P1072) to express the source format and has use (P366) with value data conversion (Q1783551). See the proposal at Q17512395#P1073.

--Dipsode87 (talk) 14:38, 10 November 2021 (UTC)

It's an interesting question & modelling problem. I'm not totally sold on your Q17512395#P1073 example b/c from the SPARQL reporting perspective it requires the user to look for a P366 qualifier of P1073 to capture the set of data migration (Q1932543) items ... which isn't a very big problem, but is more complex than would be the case if you could instead have a wdt: query.
So by way of example, an alternative might be to have a main property has use (P366) taking the value data migration (Q1932543), and (perhaps optionally?) having the read and write formats as qualifiers. I say perhaps optionally b/c one might assume that the item can convert from any of the read file formats to any of the write formats, and these would all presumably be specified as main statements in the item.
There's a large sense in which it's six & two-threes; you could do one, or the other, or both.
In the absence of any consensus on or interest in the design of the model, I suggest you press ahead with whatever arrangements you think are best. So long as the model is consistently applied, it will be easy to amend details in member items later (e.g. via SPARQL & Quickstatements) should there ever be a decision to amend said model. There are all sorts of areas in WD where users are figuring out how to model niche stuff; and it's more than legitimate for domain experts such as the digital preservation community to carve out & make arrangements on WD which suit your use cases / interests. --Tagishsimon (talk) 18:51, 10 November 2021 (UTC)
Whilst more work, I suggest creation of a new item for conversion between any two file formats, with the new item a subclass of data conversion (Q1783551). data conversion (Q1783551) is a process (algorithm) applied to an object (data structure/file format) so is the correct subclass to use. For example "conversion between XYZ and IJK" would be a new item of subclass data conversion (Q1783551) and the software which uses this process/algorithm would refer to it using has use (P366). Dhx1 (talk) 10:26, 11 November 2021 (UTC)

Actor ID or Artist ID?

At Nationalmuseum Sweden actor ID, there is a debate over the name, 99% of the entries are for artists with an occasional artist's studio, or book publisher. The ratio is similar to National Gallery of Art artist ID or Museum of Modern Art artist ID and others we use. The word "actor" seems very confusing. What is your opinion? Also peek at Maria Elisabeth Coyet (Q62118567) where a value is being added to people depicted in images that isn't valid. I already corrected a few, but left this one as an example. I contacted the person adding them, but they have not responded. We need to figure out a search that would find invalid entries and delete them automatically. --RAN (talk) 13:20, 11 November 2021 (UTC)

Uniform Resource Name

Hello! Does anyone know if Wikimedia Foundation already has or are planing to register a URN (Q76497) for Wikidata? With a registered URN its is possible to refer to wikidata objects like Sweden (Q34) as urn:wikidata:Q34 without having to bother about the technicalities of a URL. /ℇsquilo 13:46, 11 November 2021 (UTC)

Inconsistencies in Q380123 and Q23779926

There are inconsistencies in their relationship in these two data objects. While the relationship start is set to 2019 in data object Kim Hee-chul (Q380123) ([23]), data object Momo Hirai (Q23779926) ([24]) shows that the relationship started in 2020. Could someone please correct these inconsistencies? --Gymnicus (talk) 13:24, 10 November 2021 (UTC)

Seems to be 2019 ... story is dated 2020.01.02 and states (in translation) "The two have been close friends in the entertainment industry, but it has been confirmed that they recently met with a good interest in each other." I'm presuming the date format is yyyy.mm.dd & recently means slightly longer than 'yesterday'. Shame it didn't last :( --Tagishsimon (talk) 14:05, 10 November 2021 (UTC)
@Tagishsimon: In the German-language article, 2018 is indicated with this Korean-language link ([25]). I can't say whether this is true because I don't speak Korean. But at least the year 2018 is not explicitly to be found in the article. --Gymnicus (talk) 14:18, 10 November 2021 (UTC)
I'm not seeing 2018 in the English language translation - https://n-news-naver-com.translate.goog/entertain/article/112/0003252613?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=en-GB&_x_tr_pto=nui so ... don't know. At least WD is internally consistent & points to a half-way reasonable source. --Tagishsimon (talk) 14:36, 10 November 2021 (UTC)
@Tagishsimon: I don't see any direct or indirect reference to the year 2018 in the English translation, too. So I fully agree with you and would like to thank you for editing the data object Momo Hirai (Q23779926). --Gymnicus (talk) 10:22, 12 November 2021 (UTC)

Wrong merge?

Could someone please take a look at Talk:Q732590#Takeda Izumo I and II? User:Matlin is unable to answer, but continues editing. --Kolja21 (talk) 17:41, 11 November 2021 (UTC)

@Kolja21 Takeda Izumo is the professional name (Q11415657). There are three persons who called Takeda Izumo. Takeda Izumo I(?-1747), Takeda Izumo II(1691-1756), and Takeda Izumo III(?-?). The jawiki article is a mixture of profiles of the three. I have no idea the best practices in this case. Afaz (talk) 13:54, 12 November 2021 (UTC)
In this case we need one item for the professional name (Q11415657) (= jawiki article) and a single item for each person. --Kolja21 (talk) 14:56, 12 November 2021 (UTC)

Request for processing of Q5604070 and Q97859038

The two data objects Greengrass (Q5604070) and Greengrass (Q97859038) were incorrectly merged together. The data object Greengrass (Q5604070) was the family name and the data object Greengrass (Q97859038) was a Wikimedia definition of terms. That would have to be reversed, please. In addition, the English name of the data object Greengrass (Q5604070) would have to be changed, as this currently incorrect name leads to errors when sorting in Wikimedia Commons. --Gymnicus (talk) 10:12, 12 November 2021 (UTC)

@Gymnicus: Unclear why you cannot do this? There are lots of things wrong on WD. It's probably not practical to bring tham all here; easier if users fix them. Reversing a merge generally involves looking at the histories of both items, and restoring the version immediately prior to the merge. Occasionally a sitelink needs to be temporarily removed to allow that; the restore dialogue will indicate if sitelinks are frustrating a restore. Labels can be edited. --Tagishsimon (talk) 10:19, 12 November 2021 (UTC)
@Tagishsimon: Thanks for explaining how a merge can be reversed. Of course, I know this and also how to edit labels. Normally I would do this work myself, but if you no longer have rights to edit data objects, lexemes and properties, then you have no choice but to go this way. It is clear to me myself that this way is quite impractical, but there is currently no other way, as no admin wants to decide on the unblock of my block. As long as the block is in place, this is the only way. Of course you could also use the discussion pages of the data objects. But when does someone look at this? --Gymnicus (talk) 10:30, 12 November 2021 (UTC)
@Gymnicus: Ah, I see; I was not aware of that. In which case, yes, I agree, this is the best forum. As to the unmerge; I had a quick look earlier and might come back to it. I came away with thoughts that 1) there were a couple of different external IDs 2) a suggestion that there was a single person known by two names 3) a suggestion that two people are conflated in the same item 4) a variety of sitelinks under one or other name 5) jp wiki article, at a glance, seeming to document 3 different people 6) de wiki not agreeing about the date of death for its article subject, when compared with en wiki and jp wiki. So. It is all as clear as mud and may take some effort to understand and remedy. --Tagishsimon (talk) 11:06, 12 November 2021 (UTC)
@Tagishsimon: I believe that your answer from “As to the unmerge ...” onwards does not belong to the discussion here but to the “Wrong merge” discussion a little further above. At least my two data objects are not about two people who may be just one person after all. --Gymnicus (talk) 11:23, 12 November 2021 (UTC)
@Gymnicus: Oh, you're right :) ... I've unmerged them! --Tagishsimon (talk) 14:00, 12 November 2021 (UTC)

Help at seamless data import

I think I have a way to make a truly user-friendly Wikidata quick import from Wikipedia:

1. Click on [add to Wikidata] (to example.org), the item and properties are included there. Example: 

{{+data|Q1|P31|?}} - Universe is an instance of ...

2. Go to a dialog that ask you to click on buttons, forms or a simple search box in Wikidata. Example:

What is the state of the building? (add to Wikidata sandbox via quickstatements URL)
In use
Abandoned

3. Click a button and never have to worry about Q-something again!

However, I cannot find a way to do so seamlessly, and I think I need some help on getting Wikidata to receive info from Wikipedia. Do you guys have any ideas? CactiStaccingCrane (talk) 09:59, 12 November 2021 (UTC)

Probably start by looking at https://ru.wikipedia.org/w/index.php?title=MediaWiki:WEF_AllEditors.js which is the most effective tool I know of for adding to WD from a WP article. It does not have the question / clickable button answer UI which you are contemplating, but, equally, I don't have much belief that that is a good way to go. --Tagishsimon (talk) 10:07, 12 November 2021 (UTC)
Hmm, if so, then nevermind. I gonna figure out another way. CactiStaccingCrane (talk) 12:00, 12 November 2021 (UTC)
@CactiStaccingCrane: you may be interested in the plugin I wrote Wikidata:Wwwyzzerdd. It isn't quite as smart as you describe but I would be interested in improving it. BrokenSegue (talk) 15:39, 12 November 2021 (UTC)
@BrokenSegue I love it! I gonna finish my template migration first though, since it is pretty useless to have info there without being able to display. CactiStaccingCrane (talk) 15:55, 12 November 2021 (UTC)

Ranking for YYYY dates when DD-MM-YYYY is known

There are two sourced true statements, 1956 and 16-07-1956, for date of birth (P569) of Lutz Eigendorf (Q523440). Should it have normal and preferred ranking or deprecated and normal respectively? Сидик из ПТУ (talk) 17:14, 12 November 2021 (UTC)

As both claims are referenced, neither should be deprecated. Instead the precise date should be marked as preferred rank with a reason for preferred rank (P7452) set to most precise value (Q71536040). From Hill To Shore (talk) 17:29, 12 November 2021 (UTC)
(ec) Preferred and Normal. Both are right, one is more precise. Deprecation is, in general, for wrong info which has a good source. Note the qualifier reason for preferred rank (P7452) with a value most precise value (Q71536040) now on Q523440#P569. Or, concur exactly with FHTS. --Tagishsimon (talk) 17:31, 12 November 2021 (UTC)
  • At one time we had a bot performing this automatically, but there are so many of these dual values to set, it got stopped by the one minute time limit for searches. The solution was to concatenate smaller batches of the search, so that when one ended the next started, but several error detection searches I monitor are still over the time limit, despite broken into smaller batches. Anyone know the current status of the bot and if there will be relief from the congestion in SPARQL searches? --RAN (talk) 23:54, 12 November 2021 (UTC)
    My bot does this still but as you said it's pretty expensive SPARQL-wise. I might rewrite it to use the data dumps directly at some point. BrokenSegue (talk) 03:00, 13 November 2021 (UTC)
    I can see only ~7K instances Ghuron (talk) 07:22, 13 November 2021 (UTC)
We need more processing power to get searches for instance_of=human to run. Last time I checked we had 9,370,675 humans. There are several projects proposed to upload more databases of humans that are on hold because we cannot handle what we have already. I am lucky if the error/duplicate detection queries run once a month, they attempt to run daily. We probably waste a lot of processing power on searches that are incomplete and don't give any output. --RAN (talk) 03:45, 13 November 2021 (UTC)

How to add YouTube channel ID correctly

I tried to add YouTube channel ID Property:P2397 Mahatch for the channel https://www.youtube.com/c/Mahatch

But it is interpreted as https://www.youtube.com/channel/Mahatch what is incorrect.

How to add the property correctly? --Perohanych (talk) 09:20, 13 November 2021 (UTC)

@Perohanych: "Mahatch" is the name of the channel. For this property, you need to find the ID number for the channel. Here are some instructions for checking the source code of the YouTube page. -- Oa01 (talk) 10:31, 13 November 2021 (UTC)
@Oa01: Thank you very much, it works! Nevertheless, probably new properties or qualifiers for Property:P2397 should be created to enable adding names of a channel in formats .../c/Name and .../channel/Name --Perohanych (talk) 10:57, 13 November 2021 (UTC)
Hiya @Perohanych:. You can also add a qualifier "named as" in the P2397 statement. Example: Q7809. -- Oa01 (talk) 11:06, 13 November 2021 (UTC)

New user here. What do the green and red numbers next to my controbutions mean? are they some kind of scoring system?

I looked all over the site to figure out what they are supposed to mean but came up empty handed. Sorry if it's obvious. Thanks in advance.  – The preceding unsigned comment was added by HyperSniper9744 (talk • contribs) at 03:11, 9 November 2021‎ (UTC).

Characters added or removed per edit. --Tagishsimon (talk) 03:36, 9 November 2021 (UTC)
@HyperSniper9744: Please see the above. They get more red the more characters are removed and more green the more that are added. It isn't tallied anywhere or a "score" though. --TheSandDoctor (talk) 16:12, 9 November 2021 (UTC)
Yes, sort of. They get emboldened if greater than about 500 characters have been added or removed. Same font colour, though, afaics. --Tagishsimon (talk) 22:02, 9 November 2021 (UTC)
These are actually bytes, not characters. Not that it matters much here. --Matěj Suchánek (talk) 14:58, 14 November 2021 (UTC)

Filter for absent tags

Tracked in Phabricator
Task T119072
Tracked in Phabricator
Task T174349

Good morning (day, evening, night) everyone,

whenever I scan IP ranges that are home to known vandals I'd like to look at changes which have not been reverted yet. Is it possible to filter for contributions that don't have a certain tag? In this particular case I'd like to only see revisions without the tag `mw-reverted`. I've searched mediawiki docs but I coudln't find a hint how to do this. Is anyone able to help? -- Dr.üsenfieber (talk) 09:44, 10 November 2021 (UTC)

@Dr.üsenfieber: No, there is apparently not. See phab:T119072 and phab:T174349. --Matěj Suchánek (talk) 12:59, 14 November 2021 (UTC)

Position held has a vacancy for a few years

When you are filling in Position_held and there is a vacancy for a few years, what is the best way to model it? If I skip the vacancy and go to the next person in succession, future readers may think I made a mistake with an omission. See: wikipedia:sv:Livregementet_till_häst_(det_äldre)#Regementschefer --RAN (talk) 21:44, 10 November 2021 (UTC)

Presuming we're talking about a set of person items with position held statements pointing to the position? The standard is to have qualifiers of start date, end date, replaces, replaced by, and, ideally, series ordinal. (And whatever other stuff you want to throw at it - end cause, &c). The gap is what it is; future readers should trust that the coherence of qualifier values across the set of items indicates a mistake has not been made. Presume you're also aware of {{PositionHolderHistory}} e.g. Talk:Q319145. I'd not be inclined to try to fill the gap in any way. @Oravrattas: may have a view. --Tagishsimon (talk) 22:25, 10 November 2021 (UTC)
  • It's a significant event (P793) for the position. You can add it there with position not filled (Q108437605) (or sede vacante (Q48017)). Sample at Q21561383#P793. P1308 isn't suitable as it isn't used for most positions. --- Jura 11:44, 11 November 2021 (UTC)
  • I went with position_not_filled, it will eliminate the quandary in the future of people thinking it was an error, thank you. It would still look nice if it appeared in the chart. See: Talk:Q109517943. We have something similar in lists of mayors, where a person appears in italics if they were an interim holder of the office and don't count in the official tally because they are not assigned an ordinal number. Perhaps we can have an entry called "vacancy in office" and treat it as if it was a virtual person, and it appears in italics. Any other ideas that would make an entry in the chart? --RAN (talk) 13:33, 11 November 2021 (UTC)
Sweet. Meanwhile, might work: https://github.com/tmtmtmtm/wmf-position-holder-history/issues/21 --Tagishsimon (talk) 15:36, 11 November 2021 (UTC)
  • P1308 is currently in use on over 6000 positions 
    SELECT (COUNT(DISTINCT ?position) AS ?count) WHERE { ?position p:P1308 ?ps
    Try it!
    A significant number of these appear to be attempting to enumerate every holder ever there: 
    SELECT ?holderclaims (COUNT(DISTINCT ?position) AS ?positions) 
WITH { 
  SELECT ?position (COUNT(DISTINCT ?ps) AS ?holderclaims) WHERE {
    ?position p:P1308 ?ps. 
  }
  GROUP BY ?position
} AS %counts
WHERE {
  INCLUDE %counts .
}
GROUP BY ?holderclaims
ORDER BY DESC(?holderclaims)
    Try it!
    Seems to me it that it makes more sense to actually continue to use the property that's explicitly for linking a position to its officeholders, rather than introducing significant event (P793) to the mix to do that indirectly, especially as there's already a reasonable number of cases of using P1308:novalue like that: 
    SELECT ?position ?positionLabel ?start ?end {
  ?position p:P1308 ?ps .
  ?ps a wdno:P1308 .
  OPTIONAL { ?ps pq:P580 ?start }
  OPTIONAL { ?ps pq:P582 ?end }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
} 
ORDER BY DESC(?start)
    Try it!
    --Oravrattas (talk) 16:25, 11 November 2021 (UTC)
That works; doesn't exclude the possibility of duplicating the info in significant event (P793), but you're right that it's a more apt property. I've added this to RAN's item - Q109517943#P1308. --Tagishsimon (talk) 17:45, 11 November 2021 (UTC)
Seems odd to use a property that isn't applicable just to show that there is no value.
Depending on the function, it's generally not just that there is a vacancy, but for some reason the function was suspended.
But I guess it's in line with someone's triplication project (add each position three times to Wikidata). --- Jura 19:39, 14 November 2021 (UTC)
IMHO simpler and more straightforward to use replaces (P1365)=<novalue> [wdno:P735] as a qualifier of the corresponding position held (P39) triplet itself, e.g. this change. Lαδδo chat ;) 22:27, 12 March 2022 (UTC)

Is there a good way to find and fix these?

Sample:

Now we have two items for the same person.

The pair was found with Wikidata:Database reports/identical birth and death dates/1 as both items have identical dates of birth and death, but not all new items have that.

Same with https://en.wikipedia.org/w/index.php?title=Prince_Pedro_Henrique_of_Orl%C3%A9ans-Braganza&action=history :

Also with https://en.wikipedia.org/w/index.php?title=Rani_Begum&action=history :

There are currently 500,000 sitelinks to redirects. --- Jura 06:56, 13 November 2021 (UTC)

From the bot side, I'm not sure how to improve this - the best case would be if the sitelink was auto-updated when the page moved, is there a reason why that didn't happen here? Thanks. Mike Peel (talk) 09:16, 13 November 2021 (UTC)
This happened because the enwiki User "Stretchrunner II" does not have a local account on Wikidata, see User:Stretchrunner II (it says “User account "Stretchrunner II" is not registered. Please check if you want to create/edit this page.”). There are a couple of other situations where the page move is also not reflected on Wikidata, such as when the page moving user is blocked here locally.
  • You can check for this situation before creating a new item (look for page redirects that are connected to Wikidata and figure out what to do).
  • For German Wikipedia, there is User:Krdbot who tries to update all dewiki-moved pages to the new sitelink. This tries to avoid forgotten sitelink updates in these situations.
—MisterSynergy (talk) 09:25, 13 November 2021 (UTC)
  • I don't think these are necessarily created by Pi bot. The samples yes, as Pi bot efficiently adds P560/P570. All three accounts that moved the pages on enwiki aren't registered at Wikidata. --- Jura 12:15, 13 November 2021 (UTC)
As for the mitigation, in my opinion we should consider requesting some feature enhacements, such as autocreating the account on Wikidata or using a dummy account for updates that cannot be completed due to missing account here. --Matěj Suchánek (talk) 15:03, 14 November 2021 (UTC)
A dummy account sounds more reasonable, since "local account does not exist" is only one of many reasons why this could fail. "User is locally blocked" or "user has exceeded the ratelimit" and potentially some others can happen as well. —MisterSynergy (talk) 15:07, 14 November 2021 (UTC)
BTW, I added to Wikidata:Bot_requests#request_to_mirror_Wikipedia_page_moves:_enwiki_and/or_others_(2021-11-14). --- Jura 19:44, 14 November 2021 (UTC)

Merge

Hi. Please merge Isla Timotéo Domínguez (Q32431450) Timoteo Domínguez Island (Q11683850) I will prefer that the last one (Q11683850) persists, as it is the older one. Muralituy (talk) 15:24, 14 November 2021 (UTC)

Please learn how (and when, when not) to merge things. help:merge. If you are sufficiently confident to be giving advice on which way the merge should work, then you should be able to do the whole process youself. I've merged this pair. --Tagishsimon (talk) 15:28, 14 November 2021 (UTC)

Please add "Lippuautomaatti" to Q657345

Please add fi:Lippuautomaatti to the protected item Q657345, thank you. 85.76.100.94 17:02, 14 November 2021 (UTC)

done --Tagishsimon (talk) 17:17, 14 November 2021 (UTC)

How to model when two islands have merged into one?

Hi. What would be the best solution to model the case of two islands, which have merged and now there are only one geographical element? The case is Martin Garcia Island (Q719903) and Timoteo Dominguez Island (Q11683850), which now had becomed the Martin Garcia-Timoteo Domiguez Island.

Muralituy (talk) 00:28, 14 November 2021 (UTC)

Create a new item for the Martin Garcia-Timoteo Domiguez Island. Use end time (P582) and start time (P580) respectively. Pmt (talk) 08:01, 14 November 2021 (UTC)

Redirects

Tracked in Phabricator
Task T278962

When we wil be finally able to add intentional sitelink to redirect (Q70894304) to existing item and link redirects in Wikidata item without need to edit Wikipedia page? It is like editing without one hand so you can imagine how long I'm waiting for this and not only me. Eurohunter (talk) 17:03, 10 November 2021 (UTC)

You can track the progress at phab:T278962. --Matěj Suchánek (talk) 13:15, 14 November 2021 (UTC)
Is there any mitigation strategy for linkrot in sitelinks planned?
See Wikidata:Project_chat#Duplicates_on_Wikidata_(sitelink_and_sitelink_to_redirect) for part of the problem below. --- Jura 13:30, 14 November 2021 (UTC)
@Eurohunter: probably you already know the unfortunate truth, if issue exists for 10 years, it may continue to exist for the next 10 years. But it is not difficult to automate the workaround with userscript. It is just a balance between suffering and spending time for writing userscripts. --Lockal (talk) 12:31, 15 November 2021 (UTC)

Image of the person versus a decorative image

How can we stop the importation of decorative images as opposed to images of an actual person being loaded from Wkipedia? See for example: Johan Cronman II (Q5613525) --RAN (talk) 23:54, 13 November 2021 (UTC)

This issue is not limited to persons. For literary works, images are imported from Wikipedia that treat the same source subject, but are not actually tied to the literary work itself. for example, a painting of a famous mythological story, where the myth was also the inspiration for a play, book, opera, etc., yet the painting does not depict any event in the literary work or play. --EncycloPetey (talk) 00:28, 14 November 2021 (UTC)
These pictures can be moved to related image (P6802). Thierry Caro (talk) 09:11, 14 November 2021 (UTC)
Yes, but will that stop the bot from importing it again? I will delete it today, and monitor the entry to see if gets imported again. As with humans and literary_works, perhaps we need a human eye to make sure we are not importing decorations stored in the infobox. Maybe the solution is to not have decorations in the infoboxes, but that would involve the cooperation of every language Wikipedia. --RAN (talk) 02:21, 15 November 2021 (UTC)
Speak to the bot's operator and, if that doesn't help, the bot department. --- Jura 11:05, 15 November 2021 (UTC)

Wikidata weekly summary #494

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • The language codes agq (Aghem, T288335) and mcn (Massa, T293884) are now supported.
    • Mismatch Finder: Added various dialogs and help texts to make it easier to understand what reviewers need to do and what information they are seeing in the tool
    • Mismatch Finder: started polishing and bug fixing for release of the first version
    • Making the order of Lexeme's grammatical features consistent (phab:T232557)
    • Investigating how to share complex SPARQL queries in Wikidata Query Service via short URL (phab:T295560)


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.

An update on IP Masking implementation

Hello friends!

We have new information on IP Masking for you. Thank you for being patient as the project unfolds.

IP masking hides the IP addresses of unregistered editors on Wikimedia projects, fully or partially, from everyone except those who need access to fight spam, vandalism, harassment and disinformation.

So far, we have had conversations on why we are masking IPs and the tools you will need to continue fighting abuse. What is up next is, we want to share with you details about the implementation itself.

This update answers some likely questions you may have about who gets to view IP addresses and the various IP Masking implementation approaches and how each of them will impact the communities.

Please see this section for the latest information.

If you need a background on IP Masking, there’s a summary here for you.

–––

Best regards,

STei (WMF) (talk) 14:42, 15 November 2021 (UTC), on behalf of the Anti-Harassment Tools team.

Is there are Wikidata property to allow for a longer description than the 'description'?

Hi all

I'm working with a large natural history collection and they have a lot of paragraph long descriptions of species which could be used to create Wikipedia articles in the long run. What I'm wondering is if there is a Wikidata property for 'longer description' or something which could be used as a place for these descriptions to improve the Wikidata items and also act as a place for the descriptions to be indexed before they're used in Wikipedia.

Thanks

John Cummings (talk) 11:24, 15 November 2021 (UTC)

@John Cummings: We have a few properties that allow for long string "paragraphs" if from some standardized source - scope and content (P7535) for example. If you can't find an existing property that works for this then you could try proposing a new one. There is a length limit on Wikidata strings though so you might want to check on that first. Also if the paragraph-long-descriptions are on a page somewhere then they can presumably be retrieved via identifiers rather than being stored directly here, if that works? ArthurPSmith (talk) 20:39, 15 November 2021 (UTC)

Question for newyorkers

Could any NY help distinguish / disambiguate between Fulton Center(Q1020690) and Fulton Street(Q2982552) ? Should train informations be moved to Fulton Street(Q2982552) ? Bouzinac 💬●✒️●💛 20:54, 15 November 2021 (UTC)

How can I know the Freebase ID from a Google search?

Hello. In many items I find the "Freebase ID" attribute, followed by something like "/m/09qgk1". Apparently it is something important for Google, but where do you find this? Thanks in advance --El Mono Español (talk) 19:33, 12 November 2021 (UTC)

You may find this useful:
- Fuzheado (talk) 22:43, 12 November 2021 (UTC)
Thank You! El Mono Español (talk) 05:50, 17 November 2021 (UTC)

Merge

Hi. Q56246559 and Q2281694 should be merged. Can anyone help? Pirhayati (talk) 07:50, 14 November 2021 (UTC)

They must not be merged. Disambiguation items must not be merged with family name items. From Hill To Shore (talk) 10:20, 14 November 2021 (UTC)
All the pages in different Wikipedias refer to the family name not disambiguation. I merged them manually. Pirhayati (talk) 08:21, 17 November 2021 (UTC)

number of contributors to source code?

Is there a property to record the number of people who contributed to a source code? Similar to number of registered users/contributors (P1833). (I tried this but it does not seem quite right). Thanks. -- Oa01 (talk) 10:25, 13 November 2021 (UTC)

i do not think so. maybe number of participants (P1132). not all facts are worth representing here though. BrokenSegue (talk) 18:25, 13 November 2021 (UTC)
@BrokenSegue: thanks! -- Oa01 (talk) 19:24, 13 November 2021 (UTC)
@BrokenSegue: @Oa01: note that contributors to source code is subset of participants (bug reporters, testers, translators). On typical large project translators will outnumber code contributors. Mateusz Konieczny (talk) 11:49, 17 November 2021 (UTC)
@Mateusz Konieczny: Good point. Thanks. -- Oa01 (talk) 12:41, 17 November 2021 (UTC)

Does anyone know what data has been imported from Encyclopedia of Life?

Hi all

I'm speaking to someone who runs Encyclopedia of Life, I wondered if anyone had any information on wether their database of species had been fully imported into Wikidata or not. I'm not clear on how to run a query to count the number of statements of Encyclopedia of Life ID (P830) across Wikidata to check without getting a timeout.

If not would anyone be interested in working with me to import the database? I think it could really help with a lot of other projects across Wikimedia. I'm not a technical person but have a working knowledge of taxonomy.

Thanks

--John Cummings (talk) 11:45, 15 November 2021 (UTC)

Property_talk:P830 has some info. --- Jura 11:51, 15 November 2021 (UTC)
Thanks Jura1 it looks like the property is used about 1 million times, am I reading this correctly? If so it looks like about half of EOL has a corresponding Wikidata item. --John Cummings (talk) 12:13, 15 November 2021 (UTC)
@John Cummings This would a query for statements referenced with Encyclopedia of Life (Q82486): https://w.wiki/4PsY (grouped by property) - it might be of interest to you. Vojtěch Dostál (talk) 18:20, 15 November 2021 (UTC)
@John Cummings: The only data that we currently use from EoL is their ID numbers (which are in use on over a million items). As a taxonomic database, EoL is rather poor quality as its data is typically 3rd-hand and I don't think we would get any benefit from it. My understanding is that most EoL data comes from Catalog of Life data exports (which happen every few years), which in turn are sourced from smaller databases of varying quality. Thus by the time it gets to EoL, much of the data is missing (like most synonym data) and it's already outdated. The Catalog of Life is slightly better, but only slightly. Some of the smaller databases that CoL pulls from are not reliable. As far as importing taxonomic data into Wikidata in general, the best approach is to identify high-quality specialized databases that are kept up to date, for example the World Spider Catalog or World Register of Marine Species. Hope that helps! Nosferattus (talk) 19:13, 15 November 2021 (UTC)
Hi Nosferattus this is very helpful to know, thanks so much. A couple of follow up questions:
  1. Which databases it pulls from are not reliable?
  2. Would it be useful to import just the species names and IDs to identify species missing from Wikipedia etc?
  3. How would someone identify which databases would be reliable sources?
Thanks very much
John Cummings (talk) 12:04, 16 November 2021 (UTC)
@John Cummings: I should mention that all of the information I have about EoL is based on investigations I did two or three years ago, so it's possible my information is outdated. To answer your questions...
  1. CoL imports taxonomic data from 165 other databases. Last time I checked (which was a few years ago), some of those databases were self-published by a single person (rather than a scientific organization or university) and were not appropriate for importing. Unfortunately, I can't recall which ones specifically.
  2. Maybe, but due to #1, I wouldn't have confidence that 100% of those species names were valid.
  3. Reliable taxonomic databases are generally published by an established scientific organization and are commonly cited by literature in that field. They also typically provide primary source references for their data.
To be honest with you, I think what is more needed than taxonomy data importing is clean-up and maintenance of the existing data, which is a mess. Right now Wikidata has over a million items for species names, but there is very little information on which of those are currently accepted species, which are just outdated synonyms, and what names are synonyms of what other names. In order for Wikidata's taxonomic data to be useful to sister projects or 3rd party projects we need to tackle some of those problems. Of course that's a lot harder than just importing species names, and some have argued that it is impractical so long as there are competing taxonomies. Until this becomes a reality, we may just be stuck with the mess. In the meantime, I wouldn't oppose importing data from reliable specialized databases (vetted by discussion with editors specializing in those fields, for example, fish editors, spider editors, etc.). Nosferattus (talk) 02:56, 17 November 2021 (UTC)
 
Taxonomic data aggregation
P.S. - The graphic to the right might be helpful for this discussion. I think ideally we should be importing data from reliable "multi-class aggregations". Nosferattus (talk) 03:10, 17 November 2021 (UTC)
Hi Nosferattus thank you very much indeed for the very clear explanation. I'll speak to EOL to understand their data sources more and see if they have any kind of classification for their sources that would fit this description. --John Cummings (talk) 10:22, 17 November 2021 (UTC)
IIRC, long time ago I created most of the EOL statements, based on ITIS/NCBI/GBIF matching. EOL changed a lot since then. They are more like Wikidata right now - with about 1000 of sources - https://eol.org/resources - reconciled with OpenRefine. I don't remember mass imports from EOL, because, why? It is an aggregator of trees, Wikidata users can import from primary catalogues, if needed. --Lockal (talk) 16:06, 16 November 2021 (UTC)
Thanks very much for the explanation Lockal, when you say Wikidata users can import from primary catalogues, if needed can you explain a little bit more about what you mean? Why would someone collate and import all the sources individually when EOL is collating them all already? Is there some issue with how EOL imports sources? --John Cummings (talk) 18:13, 16 November 2021 (UTC)
@John Cummings:, these are just generic concerns, not directly related to EOL. All aggregated databases have a specific error rate (e. g. 0.1% of wrong statements). When you use such data, you get error rate of 0.1% + delta. As described in documentation, EOL uses OpenRefine, so you can achieve better error rate by using OpenRefine directly on primary sources and Wikidata. --Lockal (talk) 08:04, 17 November 2021 (UTC)
Hi Lockal I don't think I could do a better job than people who aggregate the data professionally, but thank you for the explanation. --John Cummings (talk) 09:47, 17 November 2021 (UTC)

New user question: removing old logo images (that were clearly marked with "end time")

A user is trying to have an old logo deleted from Commons. There is no dispute that it is a valid logo.

This logo is legitimately in use on other projects, which automatically makes it within Commons' scope. Therefore, the user is trying to have uses of the logo removed. They removed it from Wikidata themselves; before this, it was clearly marked with "end time". Is this behaviour considered acceptable on Wikidata? Brianjd (talk) 08:30, 17 November 2021 (UTC)

The article en:Théâtre Feydeau is primarily about the theater company and links to Wikidata item Théâtre Feydeau (Q825805, an instance of a theater company), whereas the highly related article fr:Théâtre Feydeau is primarily about the theater building and links to Wikidata item Salle Feydeau (Q61557606), an instance of a theater building. Could the lack of a link between the English and French articles somehow be solved by, for example, linking the English redirect page en:Salle Feydeau to Wikidata item Salle Feydeau (Q61557606)? If so, how does one do this? --Robert.Allen (talk) 04:10, 15 November 2021 (UTC)

@Robert.Allen: Yes this is possible but clunky. You need to remove the redirect temporarily in order to add the site link. You can do a similar action on fr:Théâtre Feydeau if you like. — Martin (MSGJ · talk) 15:50, 17 November 2021 (UTC)
@MSGJ: Thank you! This worked great. I followed your example and created a link to the French redirect fr:Théâtre de Monsieur, which is nearly ideal, since that was the name of the theatre company before it moved to the Salle Feydeau. --Robert.Allen (talk) 21:09, 17 November 2021 (UTC)

Feedback requested: Use of P642 "of"

  WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

JakobVoss (talk) ClaudiaMuellerBirn (talk) Criscod (talk) Daniel Mietchen (talk) Ettorerizza (talk) Ls1g (talk) Pasleim (talk) Hjfocs (talk) 17:24, 21 January 2019 (UTC) PKM (talk) 2le2im-bdc (talk) 20:30, 24 January 2019 (UTC) Vladimir Alexiev (talk) 16:37, 21 March 2019 (UTC) ElanHR (talk) User:Epìdosis (talk) Tris T7 TT me UJung (talk) 11:43, 24 August 2019 (UTC) Envlh (talk) SixTwoEight (talk) User:SCIdude (talk) Will (Wiki Ed) (talk) Mathieu Kappler (talk) So9q (talk) 19:33, 8 September 2021 (UTC) Zwolfz (talk) عُثمان (talk) 16:31, 5 April 2023 (UTC) M2k~dewiki (talk) 12:28, 24 September 2023 (UTC) —Ismael Olea (talk) 18:18, 2 December 2023 (UTC) Andrea Westerinen (talk) 23:33, 2 December 2023 (UTC) Peter Patel-Schneider

  Notified participants of WikiProject Data Quality

There have been concerns about property of (P642) going back to 2014 (see the property's Talk page), relating to translation problems and conceptual fuzziness. Lucas Werkmeister wrote an extended essay on issues with this property, and I have raised it as a data quality issue. I believe we would need strong consensus to change the best practices or constraints on a property that is used 11 million times (and a team of volunteers to make any agreed changes). If you have an opinion on this issue, please weigh in! - PKM (talk) 22:26, 16 November 2021 (UTC)

My opinion is that where it is often used as a qualifier of instance of (P31) or subclass of (P279) is not helpful. The relationship that is trying to be captured there would be better handled by some sort of direct statement; if we are missing a property for this, let's propose one to be created. ArthurPSmith (talk) 18:21, 17 November 2021 (UTC)

Update logo for British Library (Q23308)

Hi! I would like to update the the file format for the logo for the British Library (Q23308). From File:Britishlibrary.png to File:BritishLibrary.svg. The svg format displays bette at all sizes. However the page is locked. Could someone please help me? P. S. Burton (talk) 11:14, 23 November 2021 (UTC)

@P. S. Burton:   Done --SilentSpike (talk) 15:09, 23 November 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. SilentSpike (talk) 15:09, 23 November 2021 (UTC)

How were the items that got low-digit QIDs decided?

I cannot for the life of me figure out how the low-digit QIDs were allocated to various items. The allocation seems somewhat intentional at first but then also cheeky as shown with Q13 and Q24. How were these items that got QIDs in the single or double digits decided? While beginning with giant concepts like the universe (Q1) and life (Q3), QIDs 6, 7, 9, 10, 11, etc. are currently empty, so sequential order clearly was not the decider, but is the history of this allocation explained anywhere? I suppose that I am deeply curious why Africa and then Canada and then Japan and then a seemingly arbitrary collection of European countries with the United States tossed in the middle were decided to have QIDs sequentially upwards from Q15 but then why the concept of cheating was randomly included to interrupt this list of places by taking Q19? And why were those countries picked and in that order? I am just completely fascinated by this seemingly senseless but somewhat intentional allocation and would love insight to this project’s history. Thank you for the knowledge!

-- Nicolás Macri (talk) 00:52, 17 November 2021 (UTC)


There was a twelve month research projects that lead to that order ;)

Or maybe just Wikidata:Humour. --- Jura 09:14, 17 November 2021 (UTC)

Note Q42 Mateusz Konieczny (talk) 11:51, 17 November 2021 (UTC)

I'm just happy that Berlin (Q64) comes long before Germany (Q183). I'd venture a guess who ad WMDE would do such a thing, but its probably a Murder on the Orient Express (Q3241699)-like situation. (Any hints, Lucas Werkmeister?) Karl Oblique (talk) 18:17, 17 November 2021 (UTC)

I only started working at WMDE in 2017, I don’t have any special knowledge about the early days :P Lucas Werkmeister (talk) 18:36, 17 November 2021 (UTC)
It's all in the item history. Reedy (talk • contribs • logs) created Berlin (Q64), and Jeroen De Dauw (talk • contribs • logs) created Never Gonna Give You Up (Q57). Presumably they had their reasons. Ghouston (talk) 03:10, 18 November 2021 (UTC)
Here is an interesting report - User:Jeblad/first_3130_items. Apparently, Wikidata ("Wikidata in English") started with Africa (Q15), the universe was created 1 hour later. --Lockal (talk) 07:09, 18 November 2021 (UTC)
@Nicolás Macri, Lockal: I wonder why Canada was second? Anyway, from that you can see the group created by "‎127.0.0.1" were the ones that had been pre-selected for their numerical relevance - a bit of an odd collection. ArthurPSmith (talk) 15:49, 18 November 2021 (UTC)
I believe Q133 (created by ‎Yair rand) was the first item created when Wikidata was already freely editable by everyone (unless I am mistaken, this was in the pre-SUL time, so one would still need to register and then to start editing). Everything which was created before was created by the WM-DE team, and they could do (and did) whatever they wanted.--Ymblanter (talk) 17:39, 18 November 2021 (UTC)
It was literally first come, first served (bar a few reserved IDs). Reedy (talk) 18:57, 18 November 2021 (UTC)

property for books plot summary and a book series summary

Each book listed on Wikipedia has a summary of the book. I would like to be able to access this data via property. Ultimately using sparql to query for containing text in that property. Is there a property available that contains this information. Same thing for a book series?

Thanks

 – The preceding unsigned comment was added by Bradyguyc (talk • contribs).

I don't believe there is or likely will be as this is not structured data. Perhaps the case could be made for the official blurb being a data feature to capture, but that would likely not apply to many books for copyright reasons. --SilentSpike (talk) 17:09, 18 November 2021 (UTC)
We do have first line (P1922) which is sort of related. ArthurPSmith (talk) 18:26, 18 November 2021 (UTC)

"Investigate anomalies in wikidata language groups" on Unicode CLDR's Jira

See https://unicode-org.atlassian.net/browse/CLDR-15020, are there anyone familiar with items listed here? Note that this is unrelated to any "merge or not" questions as listed items are at least 90% under different topics' umbrellas. Liuxinyu970226 (talk) 11:46, 18 November 2021 (UTC)

To merge: Science articles

Hi all. Huge amount of science article items were created last time. Many duplicates were created unfortunately. More than 47000 duplicates were merged by bot. But there are many items that are similar, but have some conflicting properties:

User:Ivan A. Krestinin/To merge/Scholarly articles

The items requires manual merge. Or some more intellectual algorithm than bot has currently. Maybe somebody want to play merge game. Or fix some errors that we made during import. Be wary, some cases in the list are non-trivial. For example Q95759802 and Q93929232 are looked as equal, but represent different articles most probably. — Ivan A. Krestinin (talk) 20:59, 16 November 2021 (UTC)

@Ivan A. Krestinin: Thanks, looks like a useful list. I've previously merged many articles based on identical DOI's if their titles were "close" but that's perhaps not entirely safe. And it looks like some of these are missing titles/labels? ArthurPSmith (talk) 18:15, 17 November 2021 (UTC)
I've manually gone through a bunch at the start of your list. Authors (and author name strings) are very helpful in distinguishing duplicates from non-duplicates, so you might want to add that to your check criteria, although checking that author names match can be tricky. Many cases with differing authors seem to be where a journal (NEJM is one example) has assigned the same DOI to several different comments on a paper, while pubmed (and hence Wikidata) has a separate entry for each comment. Comments often use exactly the same title string, and they can be quite short, so it's not surprising to have identical title, volume, issue, page, etc. data for a case like that. On the ones I've checked where that is the situation I've added a "different from" statement to prevent merging these items. ArthurPSmith (talk) 17:07, 18 November 2021 (UTC)
Hello, thank you for your work! Some items does not have titles at all. Bot did not merge the items automatically because it is a bit strange case. I tried to analyze and use author (P50)/author name string (P2093) but it is non-trivial. Bot need to analyze and merge values of two properties and compare values using non-trivial algorithm, for example the next values should be detected as equal: Alexander Varshavsky (Q442466), "Варшавский, Александр Яковлевич", "Александр Варшавский", "Varshavsky Alexander", "Varshavsky AY", "Varshavskij Aleksandr", "Alex Varshavsky" and etc. "different from" is very useful for bot and other users. Such pairs will be excluded from the report on the next update. — Ivan A. Krestinin (talk) 22:26, 18 November 2021 (UTC)

Sculpture is classified as an event. What exactly went wrong?

Right now Four Seasons (Q5475472) is classified as being following things:

  • en: putto (figure in a work of art depicted as a chubby male child, usually nude and sometimes winged) [26]
    • en: artistic figure (type representation in artistic domain) [27]
      • en: depiction (presentation of facts or concepts through images or other media) [28]
        • en: communication (act of conveying intended meanings from one entity or group to another through the use of mutually understood signs and rules) [29]
          • en: information exchange (Ministry of information of Cambodia) [30]
            • en: interaction (kind of action that occurs as two or more objects have an effect upon one another) [31]
              • en: action (something an agent can do or perform) [32]

What went wrong here? It seems not utterly wrong to say that sculpture is a putto or that putto is a artistic figure. But saying that sculpture is an action is definitely wrong, while current structure claims this. How it can be fixed?

To avoid XY problem: I am trying to distinguish events such as invention of telephone (Q13443401) from objects representing specific physical feature such as Giant Geyser (Q1129264) or Wawel Castle (Q18820)

  WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

Mateusz Konieczny (talk) 11:45, 17 November 2021 (UTC)

seems like we need to split communication (Q11024) the verb apart from the noun (that which is communicated). unclear what it would be a subclass of though. not information (Q11028) since that is abstract. maybe work (Q386724) ? BrokenSegue (talk) 15:57, 17 November 2021 (UTC)
Four Seasons (Q5475472) should not be instance of (P31) putto (Q284865), rather Four Seasons (Q5475472) depicts (P180) putto (Q284865) is the correct relationship. Dhx1 (talk) 12:27, 19 November 2021 (UTC)


What's all the available public information about a property?

I know that I can see the information of a property by visiting the corresponding page in Wikidata. For example: instance of (P31)

I also know that I can retrieve "Label", "Description" and "AltLabel" from properties using "SERVICE wikibase:label". See https://w.wiki/4R3x.

I just learned that I can also retrieve the propertyType of a property using "wikibase:propertyType". See: https://w.wiki/4R3y . Note that this information is not presented in the Wikidata entity.

My question is: Just as propertyType is information of a property and it is not listed in the page of the corresponding Wikidata entity. Is there any more public information that can be obtained from properties (and is sort of hidden)?

Rdrg109 (talk) 23:57, 18 November 2021 (UTC)

Have you checked out https://www.mediawiki.org/wiki/Wikibase/Indexing/RDF_Dump_Format which specifies in great detail entity representation and other data model fu in wikidata. (Although that said, there doesn't seem to be any more public information that can be obtained from properties). --Tagishsimon (talk) 00:04, 19 November 2021 (UTC)
@Rdrg109 the property type is listed on the property page, see the “Data type” section between the labels/descriptions/aliases and the statements. Lucas Werkmeister (WMDE) (talk) 10:45, 19 November 2021 (UTC)

Superior hotel classification

Simple hotel star classification (number of stars) can be noted using award received (P166), hotel rating (Q2976556) and quantity (P1114). But it is maybe impossible to specify classes like 4 star superior (★★★★S). Which qualifier can be used to store the superior class (of course, a superior star class item has to create, too)? --RolandUnger (talk) 07:22, 19 November 2021 (UTC)

I'd advise setting up a new item for this if you are wanting to apply it. "4 star (Superior)" "a ranking for hotels awarded by some rating organisations" with an instance of "award." On your hotel item you then set the award received to point to your new rating item. You'll probably want to set the conferred by qualifier to clarify who granted the award as well as start qnd end times (as few hotels keep the same rating forever). From Hill To Shore (talk) 08:49, 19 November 2021 (UTC)

merge

please merge Q5961102 to Q1331140. thanks.🙏  – The preceding unsigned comment was added by M.Nadian (talk • contribs) at 19:47, 29 October 2021‎ (UTC).

How to get the v1 commands from an existing wikidata item?

I want to copycat the commands from an existing wikidata item to create a new one just with different values. So how to get the v1 commands from an existing wikidata item?--CreateAccou4343nt555 (talk) 23:36, 19 November 2021 (UTC)

This is one option: User:Magnus Manske/duplicate item.js ... instructions in rows2-7 of the code. In other news, I think conventional WD terminology would be 'statement' or 'claim', not 'command'. --Tagishsimon (talk) 23:52, 19 November 2021 (UTC)
@Tagishsimon I will try that. I said v1 commands because I will use quickstatements for creating new items (or at least that was my plan). CreateAccou4343nt555 (talk) 23:56, 19 November 2021 (UTC)
That's another valid route; but if so, in my experience, one hand-knits a spreadsheet which produces QS commands covering the statements one wishes to see in the new item(s), based on experience and/or observation of the sorts of statements used on items of that sort. It is possible to use the query system to provide a breakdown of the predicates used on an item, though it's probably rare that it's worth taking this route rather than inspecting the item thru the normal UI: example query dealing with truthy statements for Douglas Adams (Q42): https://w.wiki/4RRn --Tagishsimon (talk) 00:17, 20 November 2021 (UTC)
@Tagishsimon that is pretty useful. Thank you so much! CreateAccou4343nt555 (talk) 01:19, 20 November 2021 (UTC)

Can we add a warning?

When you add a person as their own father or mother, we should have a warning. I have come across this a few times when looking at the familytree function. It happens when multiple generations have the same name, so people have added the person from the wrong generation. --RAN (talk) 14:54, 18 November 2021 (UTC)

  • @Richard Arthur Norton (1958- ): Maybe there should be a warning, but I believe all self-referencing statements like this have the edit tagged with "self-referencing", so that's one way to track them down. ArthurPSmith (talk) 15:56, 18 November 2021 (UTC)
    I still think it would be nice to have a general "this property cannot form cycles" class/constraint, even if we can't verify it automatically just yet. More complicated because there are two properties for "father" and "mother", but I suppose it's too late to switch to a single "parent" property? Streetmathematician (talk) 16:48, 18 November 2021 (UTC)
    • Not sure how that last point would help in general and/or for the issue mentioned. --- Jura 16:59, 18 November 2021 (UTC)
      For the issue mentioned, it would help because we could more easily detect cycles of the form "A is B's father, B is C's mother, C is A's father" and so on.
      In general, there's the whole sex vs gender issue, weird genetics (particularly in non-human species), and fictional scenarios.
      Can you name a way in which having two (or three) properties helps? Streetmathematician (talk) 17:27, 18 November 2021 (UTC)

If I have a website with some information about writer Jules Verne, I can hint my readers that more information is found in Wikipedia: 

<a href="https://fr.wikipedia.org/wiki/Jules_Verne" >Wikipedia</a>

but that is only the French Wikipedia, not English Wikipedia, not Wikisource, etc. I could add links to all of these sites, which would result in very many links about the same author. Instead, perhaps I could link to Wikidata? 

<a href="https://www.wikidata.org/wiki/Q33977" >Wikidata</a>

The problem with this is only that the link leads to a very ugly page (Q33977) that doesn't look useful to anybody, and from where it is not obvious that more information can be found in various languages of Wikipedia or whole books in Wikisource. So how should this be done? What is the right way, that other websites use for linking to Wikidata? Please avoid guessing and speculation. Show me real examples of real websites that has links to Wikidata. My real website that does link to Wikipedia and Wikisource is here. --LA2 (talk) 16:24, 20 November 2021 (UTC)

Maybe link to the view of WD provided by Reasonator, which is a very fine view indeed? --Tagishsimon (talk) 16:31, 20 November 2021 (UTC)
It is indeed less ugly than Q33977, but it does not serve as a landing page for people who want to visit Wikipedia or Wikisource. Links to Wikipedia and Wikisource are hidden in the bottom right corner of that long page. Perhaps the solution is to create a different tool for my purposes at the Toolforge? Or is it possible to do something only with CSS? --LA2 (talk) 20:19, 20 November 2021 (UTC)

Swedish photographs

I want to show the status of copyrights for a photographer from Sweden. In Sweden ordinary portrait photographs were exempt from copyright prior to 1971. Excepted were "art photos", that are protected 70 years pma. How would we show that at Kjell Sandström (Q109620143)?. It affects his US copyrights, because the images were in the public domain in Sweden prior to 1996. They would still be under copyright in strict 70 year pma countries. We made need to create a new "determination method" to cover the Swedish copyright exemption, but what would it be called? --RAN (talk) 22:24, 20 November 2021 (UTC)

We have paragraph 49a, chapter 5, Swedish copyright act (Q65007158), I think that's enough. Ainali (talk) 22:56, 20 November 2021 (UTC)
Just what I need, thanks! Can you look at Kjell Sandström (Q109620143) now and see if you think it is clear as written. --RAN (talk) 00:52, 21 November 2021 (UTC)
I wasn't the one being asked, but... It's excellent except that, if I'm not mistaken, it should be start time (P580), not end time (P582) as a qualifier. The qualifier signifies when the statement became valid/invalid. Since the statement is a negation, it started when the copyright ended.
If you want to go totally overboard, you can further add applies to part (P518)portrait photography (Q182956)excluding (P1011)fine-art photography (Q1066582). However, that can be deduced from the existing data, so I can see arguments for adding it, not adding it, or adding/removing it to/from all items currently using that copyright status. Karl Oblique (talk) 01:50, 21 November 2021 (UTC)
I like that, very clever. I think if I was looking at the entry with no prior knowledge I would be confused, your additions make it very clear. --RAN (talk) 02:38, 21 November 2021 (UTC)

constraint error

Why do we get a constraint_error at 70 years after creation by anonymous or unknown author (Q65016080) for the value of "70"? --RAN (talk) 23:29, 20 November 2021 (UTC)

We appear to get the same error on other items that have a unit of year but not for items with a larger number but a smaller unit. I am unclear on the system mechanics here but I am guessing that the value range in the constraint is also referring to the scale set in the unit item (possibly one of the conversion statements). Because 70 x a big unit is larger than 1,000 x a small unit, the year values are exceeding the range when smaller unit values are not. Howvever, even if I am correct, I have no idea on how to resolve this. From Hill To Shore (talk) 00:28, 21 November 2021 (UTC)
Because duration (P2047) is constrained to be at most 1,000,000,000 seconds. According to the documentation:
WikibaseQualityConstraints normalizes all quantities (applies unit conversion) before checking this constraint.
For example, a mass of 500 g will be considered to lie within the range of  , although 500 is not within the range  , because the range endpoints are converted to  ; on the other hand, a duration of 100 days will be considered to lie outside the range of   (note the missing units) because it is interpreted as 8640000 seconds. Streetmathematician (talk) 00:36, 21 November 2021 (UTC)

Nationality for ancient people

Hi, c:Creator:Johann Gabriel Schleich: This German painter lived from 1710 to 1743. So obviously linking to the modern state of Germany doesn't make sense. Johann Gabriel Schleich (Q15821241) says country of citizenship (P27)=Duchy of Württemberg (Q2227570), but the nationality doesn't show in the Commons template. Idem for people who lived in Italy before 1860 (Kingdom of Italy (Q172579)). How to do that properly? Thanks, Yann (talk) 20:28, 20 November 2021 (UTC)

@Yann: Wikidata:Nationality. Multichill (talk) 11:28, 21 November 2021 (UTC)
@Multichill: I mean by using demonym (P1549). It seems to work in some cases, but not here. Yann (talk) 11:52, 21 November 2021 (UTC)

I am working with the Yale Center for British Art to load much of their content into Wikidata and Commons. Yale have asked that all of these works be marked as public domain. This is straightforward if the creator of the artwork is known and died over 100 years ago, or the work was created a long time ago. But what kind of qualifier should be used with copyright status (P6216) public domain (Q19652) if the work’s creator is unknown/anonymous and the inception/creation date of the work is also unknown? Rob Lowe - Smartify (talk) 17:41, 19 November 2021 (UTC)

That will depend on the particulars of each case and the reason why they are considered to be public domain. While the precise dates are unknown, is there evidence that they belong to particular periods? For example, are there archive records saying the works existed for at least 100 years? Also, is there evidence of country of origin or publication? From Hill To Shore (talk) 19:15, 19 November 2021 (UTC)

Merging of Q108881946 and Q109669844

The two data objects Arndís Anna K. Gunnarsdóttir (Q108881946) and Arndís Anna K. Gunnarsdóttir (Q109669844) treat the same woman, so I ask for them to be merged. --Gymnicus (talk) 12:17, 22 November 2021 (UTC)

Done. From Hill To Shore (talk) 13:02, 22 November 2021 (UTC)

Disambiguating authors using ORCIDs from Crossref Data

(I'd be happy to move this to a more specific discussion page if one exists).

In this situation:

  • We already have a Wikidata item for an author
  • That item has an ORCID.
  • We already have a Wikidata item for an article
  • That item has a DOI
  • Crossref knows about the DOI and identifies the author as having the right ORCID
  • All names and numerals match: the article currently has a author name string (P2093) with the right series ordinal (P1545), Crossref has an author with the right name in the right position, and the label of the author's Wikidata item contains the same name.

I would like to replace the author name string (P2093) with a author (P50) specifying the Crossref API as a reference URL (P854).

The Crossref API states:

The data is not subject to copyright, and you may use it for any purpose.
Crossref generally provides metadata without restriction; however, some abstracts contained in the metadata may be subject to copyright by publishers or authors.

As no abstracts are used in the process, I think it's legally okay.

I've semi-manually prepared an edit, and would like to proceed by automating the rest of the process, creating an WD:RFBOT and, provided no objections are raised, performing the test edits before waiting for the RFBOT's resolution.

I can't give the precise number of affected items since my preliminary numbers are based on an old Crossref public data file dump rather than live API requests, but I'd estimate it's 400,000±200,000. Streetmathematician (talk) 13:53, 20 November 2021 (UTC)

this seems fine to me. One thing though. Would it be better not to replace the author name string (P2093) but merely add author (P50)? Is it wrong to have both? BrokenSegue (talk) 17:41, 20 November 2021 (UTC)
See here if you want to fetch all author item of an article (but you need to create the author items first). Alternatively, Wikidata:Tools/Author Disambiguator can convert author name strings to authors (via a semi-manual process not involving ORCID). @BrokenSegue: qualifier object named as (P1932) should be used if we already have author (P50).--GZWDer (talk) 18:01, 20 November 2021 (UTC)
  • @Streetmathematician: As long as you are matching both on name and series ordinal (P1545) this should be fine. Note that articles may have several authors with the same "name" (eg. "A. Smith") in which case other information is necessary to facilitate matching. Also note that ORCID itself, if it provides ORCID-DOI relationships, does not include information on position in the author list, so another source is needed for that (Crossref should be fine). Do you have a particular name matching algorithm in mind or will you be only looking for exact string matches (perhaps ignoring punctuation)? WD:RFBOT would be one good place to discuss this in more detail; there's also Wikidata:WikiProject Source MetaData. ArthurPSmith (talk) 20:11, 22 November 2021 (UTC)

Properties for court verdicts?

  Notified participants of WikiProject Law

We currently have trial-related properties such as defendant (P1591), charge (P1595), plea (P1437), convicted of (P1399), and exonerated of (P7781), and a host of other law-related properties at Wikidata:WikiProject Law, but I can't find a property for "verdict" or "court decision", which seems like an important property to have in items about trials and/or defendants. For instance, charge (P1595) → negligent homicide (Q641449); Verdict → acquittal (Q1454723) (or perhaps new trial-specific items for "guilty", "not guilty", etc. as warranted). I also can't find the opposite property of convicted of (P1399), which would be "acquitted of" or "found not guilty of". Such properties could be also be qualifiers for properties such as charge (P1595). Have such properties been proposed before? Are there compelling reasons (e.g. redundant) why they should not exist? -Animalparty (talk) 03:09, 21 November 2021 (UTC)

Yeah I'm really shocked we don't have this yet.*Treker (talk) 07:24, 21 November 2021 (UTC)
One issue I can think of is the interaction with living people. In some parts of the world the official record of some low-level crimes is expunged after a sentence is served. This is to allow petty criminals to reintegrate into society after they are punished. For example, a teenager who steals an item from a shop is punished for theft but then has the slate wiped clean; this allows them to obtain a job and move on rather than be ostracised for life over a single mistake. If our structured data captures information on all crimes, we are likely to retain details of convictions after the official record is expunged. This could lead to potential legal action against us. Is there an existing guideline on recording crimes for living people? If not, I'd recommend setting one up. From Hill To Shore (talk) 09:29, 21 November 2021 (UTC)
I imagine these properties (including ones already existing) are meant to be used for people who are notable for their crimes.*Treker (talk) 11:25, 21 November 2021 (UTC)
There is Wikidata:BLP. Karl Oblique (talk) 16:31, 21 November 2021 (UTC)
The property convicted of (P1399) would seem to be a much more sensitive privacy issue, and it's been in place since 2014! Also, FWIW I don't really believe Wikidata cares about privacy from an institutional level, even though individuals rightfully have concerns. Wikidata:Living people is basically lip service without substance, saying that sensitive/privacy violating material just needs verification, and putting the onus on the subject of the item (who may not know anything about Wikidata) to complain. Wikidata as a collective beast cares only about data. Sooner or later, dead or living, we will likely all be described in Wikidata. -Animalparty (talk) 17:39, 22 November 2021 (UTC)

The name of an entity changes - what to do?

Hello! I have only recently begun to use Wikidata, and so I am not completely familiar with it. I noticed that several new counties have been created in Iran in the recent years, usually by elevating them from their former status of a district within another county. To be specific: Lashar County, which used to be Lashar District of Nikshahr County. Now, as I started to update the higher entity (Sistan and Baluchestan province), I noticed that no entity exists under the name of Lashar County, so I created it. However, it was only after this that I noticed that Lashar District is actually an extant entity.

My question: what should I do? Should the new entity be deleted and the old one renamed and updated with the new realities? Or is there a way to somehow merge them? Or should Lashar District be left as a separate entity, being a now defunct district? Thank you in advance. --Mathae (talk) 11:55, 22 November 2021 (UTC)

@Mathae: I would leave both entries so that we have a structural data record of before and after the changes. However, I would insert replaces (P1365) and replaced by (P1366) on the two items to show the connection. I'd also set inception (P571) and dissolved, abolished or demolished date (P576) on the two items to show that one is a current record and the other is a historic record. From Hill To Shore (talk) 13:09, 22 November 2021 (UTC)

@From Hill To Shore: Thank you for the quick answer. This seems to be a reasonable and precise solution. --Mathae (talk) 14:42, 22 November 2021 (UTC)

If its only a rename, then you can keep the old instance, and add official name (P1448) with the two names and their start time (P580)/end time (P582). But since in your case it also changed its status in the hierarchy, you have to create new instances. There might even be the case that a newly created county contains just one single district with the same name. Ahoerstemeier (talk) 15:45, 22 November 2021 (UTC)
@Ahoerstemeier What about Dijon University (Q97073477)? Does it make sense to have it separate from University of Burgundy (Q287072)? I think it's a candidate for merging. Vojtěch Dostál (talk) 19:58, 22 November 2021 (UTC)

Wikidata weekly summary #495

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week
    • WikiCite Search is a bibliographic search engine for Wikidata that finds articles either by searching for keywords, or by string matching.
  • Development
    • Working on displaying the grammatical features of Lexemes in a particular order in the UI (phab:T232557)
    • Mismatch Finder: continuing polishing before first release. Focusing on making API documentation available and adding a footer to the site
    • The ongoing work on MediaWiki skin improvements especially for Wikipedia will break the search box for Wikidata. We're working on addressing this. (phab:T275251)
    • Migrating a number of components to vue 3 to keep up with the rest of MediaWiki (phab:T294465)

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.

Search function not working

I'm having no success searching in Special:Search or in the web interface when looking for items to add as values. Anyone having a similar problem? I tried logging out and back in, and have used two separate browsers (Chrome and Firefox). AleatoryPonderings (talk) 20:16, 22 November 2021 (UTC)

Yes, seems to be kaput. --Tagishsimon (talk) 20:21, 22 November 2021 (UTC)

Property proposal help

I think it would be useful to create a property for commercium book used by student society. For example Woltje uses Codex Woltiensis, Commin@ uses Codex Studiosorum Bruxellensis. Somebody who has experience in this sort of property who would like to help me make a proposal? Jhowie Nitnek 22:29, 22 November 2021 (UTC)

Page renames at Wikipedia/Wikisource, etc

At Help:Sitelinks#Page_renames, I added a new section about what happens. Please correct/complete as needed. --- Jura 10:59, 23 November 2021 (UTC)

Compact view of items as gadget in preferences

Hi all! I have been using for some months User:Jon Harald Søby/compact items.css by @Jon Harald Søby:, which I find greatly useful for sparing time in scrolling items; it is used by nearly 60 users as of now. I have sometimes one little problem: when I have to make presentations about Wikidata, I need to turn it out editing my common.css (to have an interface more similar to the one of my public) and then turn it in again after the presentation, in both cases editing my common.css (I know the possibility of appending "?safemode=1", but doing it for each item I open can be annoying in long presentations); managing it in my preferences would be faster. In conclusion, given that the gadget is used by a relevant number of users and could probably gain other users being activatable in Special:Preferences (many users are unfamiliar with common.js, still more probably are unfamiliar with common.css), I would propose to add it to Special:Gadgets (which should be done by an interface administrators, if there will be consensus here), as it was done in this and other cases. --Epìdosis 22:01, 18 November 2021 (UTC)

  Support as a user of User:Jon Harald Søby/compact items.css (can a CSS change be a gadget? I have no idea...) - PKM (talk) 22:38, 18 November 2021 (UTC)
  Support--Alessandra Boccone (talk) 11:49, 19 November 2021 (UTC)
  Support Pure quality of WD life. Moebeus (talk) 23:46, 20 November 2021 (UTC)
  Support--moz (talk) 09:09, 22 November 2021 (UTC)
Oh, that's a nice stylesheet, definitely going to adopt that one. Note that if you load a stylesheet via javascript (gadget), there will be a noticeable delay. Whereas stylesheets loaded the usual way will be fast and most likely cached as well. Edit: Seems external stylesheets are not cached, and so they are noticeably slower than the user's own stylesheet. --Infrastruktur (T | C) 12:29, 19 November 2021 (UTC)
In my humble opinion, what the script displays should be the default appearance of item and property pages. Thierry Caro (talk) 15:32, 19 November 2021 (UTC)
+ 1 Ayack (talk) 09:32, 23 November 2021 (UTC)
Thanks for the suggestion, Epìdosis. Glad you like it! Thanks to Nikki, it has now been made into a gadget! 🥳 If anyone has any feedback or notices any bugs, feel free to contact me so I can sort them out! Jon Harald Søby (talk) 17:57, 23 November 2021 (UTC)
Great thanks again, @Jon Harald Søby: and @Nikki:! Good evening, --Epìdosis 21:26, 23 November 2021 (UTC)

Adding Template:Graph:Pie chart to cywiki

@Jura1: I'd like to know if there are any technical reasons for not placing the Wikidata generated graphs such as Template:Graph:Pie chart onto a Wikipedia article. Can someone please advise? I've copied the code over here, but looks more like a template containing Lua script] and doesn't work, at present. Thanks. Llywelyn2000 (talk) 17:08, 22 November 2021 (UTC)

  • I'm not sure, probably not. Some wikis have editorial policies requiring contributors to re-type data, but it's not a technical requirement.
@Bouzinac: uses a similar template for airport statistics on frwiki (and possibly others). Some time ago, the feature broke and it took months to get working again.
Practically it's a query embedded in a template. What is inserted in an article can be limited to {{some template}}.
It's rather tricky to set up and can fail for obscure reasons. Also, there are a few similarly named templates that work differently and some wikis have those under the same name set up.
Wikidata:Report_a_technical_problem might be able to provide more support. --- Jura 14:39, 23 November 2021 (UTC)
Brilliant! Thanks once again Jura1! You never fail! Llywelyn2000 (talk) 16:39, 23 November 2021 (UTC)

Merge two Items

I want to merge the items Q97400325 and Q7132478. Do I really have to change all the descriptions manually to make Special:MergeItems working? Or is there another way to do this? --Redrobsche (talk) 20:13, 26 November 2021 (UTC)

Thank you for spotting those. I went ahead and merged the two items since this one was a bit tricky. We wanted to merge into Q7132478, since it was the original of the two. However, the descriptions at Q97400325 were preferable. The solution was to merge Q7132478 into Q97400325, then undo the creation of the redirect and merge Q97400325 back down to Q7132478.
--Quesotiotyo (talk) 23:00, 26 November 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. SilentSpike (talk) 14:08, 29 November 2021 (UTC)

Notability point 2, what is a "serious" source?

I've been around for a while now and a point of uncertainty remains for me around notability. At WD:N point 2 currently states "The entity must be notable, in the sense that it can be described using serious and publicly available references."

The word that stands out to me is "serious". Because depending on your interpretation of that one word many possible entities land on one side of the line or the other. I get that the notability guidelines are intentionally vague, but this strikes me as a point where they shouldn't be. Is any recognisable news website a "serious" source even though articles are pumped out en mass? Are company websites "serious" sources even though 1st party? There is no agreed community definition that I've seen, and I think having some clarity there would be valuable. --SilentSpike (talk) 22:29, 19 November 2021 (UTC)

@SilentSpike: should be like w:WP:SOURCE. Multichill (talk) 11:33, 21 November 2021 (UTC)
Thanks for linking that @Multichill: I tend to try and follow similar policy as enwiki, but it would be nice to see something more concrete in our own policy to reflect an agreed understanding. Wouldn't want to just refer to enwiki though since this is a multilingual project. --SilentSpike (talk) 11:58, 24 November 2021 (UTC)
The question hasn't been quite fleshed out, mostly because there are far fewer conflicts here than on WP, which means that people are less likely to engage in elaborate schemes with manipulated sourcing, and also that there is less temptation to argue about a source to get some statement thrown out. Practically speaking, it's awfully close to "I know it when I see it". Or, slightly more specific: do I trust this information at the Google-Maps-level, to plan my route etc? Note how, at that level, I have no serious doubts in its accuracy, but wouldn't use it on enwiki. That's also were first-party sources fall, as long as they have some sort of track record. Part of the reasoning is that parallel, contradictory statements are perfectly possible as long as they are sourced in such a way that data consumers have a choice. Although I fear that idea hasn't really been tested yet.
Specifically to notability, I believe the intent (and definition of notability you quote) is close to "we have no doubt it exists", with the exception of items people create for themselves, which are delete because it's just uncouth.
More generally, the elephant in the room is that 90% of statements have no sources, or just imported from Wikimedia project (P143), and how that doesn't bother you much looking at the item, but makes it completely useless for any actual use. Karl Oblique (talk) 16:27, 21 November 2021 (UTC)
Thanks @Karl Oblique: you touch on some really insightful points (the Google Maps comparison being quite apt) that definitely cross my subconscious while editing, but which I haven't considered problematic enough yet to discuss in the community. Regarding the lack of existing sourced statements, part of my motivation (as someone who envisions a future where all statements are sourced) for questioning this policy is that it would be good to solidify it ahead of time to encourage use of meaningful sources by clarifying what we consider those to be. I think even simply clarifying types of sources that don't meet our definition of "serious" would be a great (and easier) starting point. --SilentSpike (talk) 11:58, 24 November 2021 (UTC)
Company websites can’t be enough to prove the notability of a company; it might be very different for products of said company. I’ve started some personal thoughts on that matter on my backyard, but it’s neither official nor really sophisticated. Emu (talk) 16:54, 21 November 2021 (UTC)
I tend to agree @Emu: though I think even if the company is notable, if their product is only documented on their own site then it is not by association notable. Nice personal thoughts too, while not official they're more fleshed out than my own currently. --SilentSpike (talk) 11:58, 24 November 2021 (UTC)

Property with no item using it

I've found a property that has zero items using it: identifier shared with lexeme (P9531). I can confirm this because of this simple query that returns 0 rows.

I'm now wondering: Am I missing something or no item use this property? If the latter is true, why does this property exist?

Rdrg109 (talk) 15:36, 23 November 2021 (UTC)

It happens that we create properties and the user(s) who requested it lost interest in it ..
Many properties aren't used that much.
For identifier shared with lexeme (P9531), you need to query qualifier uses: [33]
BTW Property_talk:P9531 has those stats too. --- Jura 15:41, 23 November 2021 (UTC)
Thanks for the info! It turns out that this is indeed used as a qualifier. I should have checked that. Rdrg109 (talk) 21:49, 23 November 2021 (UTC)
Special:WhatLinksHere might also be helpful for a quick overview, as in Special:WhatLinksHere/Property:P9531 for instance. On that special page, it does not matter how exactly the property is used, it just links all pages that somehow use the property. —MisterSynergy (talk) 10:19, 24 November 2021 (UTC)

Please undo a lot of nonsense edits

Could somebody (automatically) undo these edits: https://www.wikidata.org/w/index.php?title=Special:Contributions/Jobu0101&offset=&limit=500&target=Jobu0101 . A user added the German titles of a German series as English titles on thousands of pages. They are complete nonsense, because this series is not even shown in English countries and if so, the episodes would not have German titles in English countries. --92.213.11.38 11:19, 23 November 2021 (UTC)

What does User:Jobu0101 say about it? --HarryNº2 (talk) 11:44, 23 November 2021 (UTC)

I think that what I did is perfectly fine. I didn't add titles, I added only labels. See my full answer here. --Jobu0101 (talk) 12:36, 23 November 2021 (UTC)

By the way: Having Enlgish labels has advantages: When you see the items linked in some other items and your language settings are set to English you only see the wikidata id. This tells you nothing. Now you see the German title (as English label). --Jobu0101 (talk) 12:40, 23 November 2021 (UTC)
It has disadvantages, too, not least of obscuring the absence of an English language label (i.e. a label that is actually in English, rather than a German language string); and so it is not 'perfectly fine' - that would be a complacent view. --Tagishsimon (talk) 12:49, 23 November 2021 (UTC)
Labels are for the names/titles/terms most commonly used for a topic in that language. There's absolutely no requirement or expectation that the string used as a label must only use that language. Since that's not how it works in reality --2A02:810B:580:11D4:2F4:6FFF:FE72:D65F 04:40, 24 November 2021 (UTC)
You could add it as alias and with title (P1476) instead. This way it can easily be found also in English and other languages. Please don't fill the enlabel without some source. --- Jura 12:57, 23 November 2021 (UTC)

I thought a label is distinct from title (P1476) or other actual properties just a string we use to label the item to help us see faster what the item is about. So it's no official data of the item, it's just a string helping us. In the case here where we talk about German television episodes the best possible string helping us would be a string containing the name of the series and the title (which is in most cases of course German). I'm using Wikidata in English by the way and it is quite annoying for me if items don't have English labels. For example when I view a Tatort item and wanna see which episode was the next one (which is included with the qualifier follows (P155)) I rather see a German label instead of a non-telling wikidata id. --Jobu0101 (talk) 14:31, 23 November 2021 (UTC)

Given the settings on User:Jobu0101, the German label should be visible even if your GUI is set to English. If not, it's a bug to fix. I think setting title (P1476) is more important than copying an possibly incorrect label to 300 languages. --- Jura 14:45, 23 November 2021 (UTC)
I have this bug then for years. Of course I also reviewed the title (P1476) of all Tatort episodes and corrected it where it was wrong. I also corrected wrong episode numbering and added publication date (P577) where it was missing. See for example [34], [35], [36], [37], [38] and many, many more! --Jobu0101 (talk) 14:53, 23 November 2021 (UTC)
 
If you don't get something like the box on the right side with English/German/French, ask at Wikidata:Report_a_technical_problem. --- Jura 16:00, 23 November 2021 (UTC)
That box I get. But if in one item another item is linked it shows the wikidata id instead of a (possibly German) label if no English label exists. --Jobu0101 (talk) 20:41, 23 November 2021 (UTC)
Maybe the fall-back chain needs fixing. --- Jura 09:40, 24 November 2021 (UTC)
It's perfectly okay to use the original title/name as a label if a work or any other concept doesn't have a localized/translated name or title. If an American book doesn't have a French translation, that English title is used as the French label. If a French television series uses an English-language title and is never released in Germany, the English title will be used for the French, English and German labels (as well as all other languages where that show was either released or referenced using that title or never released/referenced). Really no different than, say, organizations, streets, ship names or other stuff.
Labels are very important for both users on our site as well as reusers of our data. There's no point in keeping labels intentionally blank just because a work wasn't released/referenced in those languages, if we know how that work would be referenced in those languages (namely by it's original title). --2A02:810B:580:11D4:2F4:6FFF:FE72:D65F 04:40, 24 November 2021 (UTC)
Thank you very much. You explained it better than I could have done it but this are precisely my thoughts behind it. Let me add one more thought about it. Imagine some German Tatort episode would have an article in the English Wikipedia without being released in any English speaking country. What lemma would we choose for this episode? I guess the lemma would be the original title or at least include it. That's the way we should think about labels in Wikidata. --Jobu0101 (talk) 09:18, 24 November 2021 (UTC)
If only we had been here before and settled on & documented the radical idea that EN labels should be in EN. https://www.wikidata.org/wiki/Help:Label#Items_without_pages_on_Wikimedia_sites points 4 & 5. --Tagishsimon (talk) 09:33, 24 November 2021 (UTC)
  • No, it's not perfectly ok. The comparison with streets is obviously nonsense, as streets rarely have localized names. The consensus for films and related items is not to fill them. If you want to change it, please bring it up on WikiProject Movies. --- Jura 09:40, 24 November 2021 (UTC)
What I did is perfectly in accordance with point 4 of your linked page. Since we are talking about art work here the right enlabel would be the transliteration of the original title. That's basically what I did.
Once again, nobody would complain if the episodes had English Wikipedia articles. It's ridiculous that the absence of those articles forbid to enter an English label. --Jobu0101 (talk) 09:50, 24 November 2021 (UTC)
Transliteration is something entirely different. You may assume that there are three languages to fill, but Wikidata has hundreds and no, we don't want to sort through all of them between those filled in Jobu0101 and actual titles. You still haven't explained what would be the benefit of doing so (besides that there are problems with the fall-back chain). --- Jura 10:04, 24 November 2021 (UTC)
  • I've seen this with scientific article items a lot, but in those cases usually English titles are being put as labels in languages other than English. Since we apparently talk about work titles, I do not really see an issue here. —MisterSynergy (talk) 10:17, 24 November 2021 (UTC)
I believe Jobu0101 is right in identifying Help:Label#Items_without_pages_on_Wikimedia_sites point (4) as the closest to what they are doing. Note that the policy is flawed in that its 5-step process doesn't allow for leaving the label blank except in cases where the user is not familiar with the source language, which is clearly not the case. It does, however, start with mentioning that some items should have English labels, implying the existence of others, but not giving criteria for that distinction.
The reason why point (4) might not apply in Jobu0101's cases is that titles in non-latin alphabets have a stronger need for an English label. It is quite possible to work with latin-character names (when going through lists etc) even if it is in some language you don't speak. If you cannot read the alphabet it uses, it becomes (in my experience) extremely hard ("long line from the 9 o'clock position upwards to the right...")
Similarly, the logic of pointing to situations where one episode of a TV series has a original-language enwiki sitelink as precedent is so obvious that I cannot quite believe people's blanket dismissal of these ("...but we are not talking about..."), leaving the argument standing at this point. As an attempt to take a whack at it: enwiki has a stronger need to come with a label, any label, because it is also the page title and the URL, i. e. it is required for an article. That is not case here on wd. When we use enwiki titles for the label here, we do not rely on enwiki as the authority to decide such labels, but, in turn, on their policy and ability to have identified some trusted outside authority to make that decision.
Reading through this thread I am getting the impression that there may be possible common ground in adding the original-language title not as a label, but as an alias? (Apologies if I missed something). This would be justified on real-world grounds because even in cases *with* a localized title, original-language titles are sometimes used (because it makes you seem worldly and smart). A side effect would be that it might move up a few spots in some implementations' fallback chain and might beat out the QID in some cases where it currently does not (although not the one for the main search function here, as I just noticed). Karl Oblique (talk) 20:41, 24 November 2021 (UTC)

Does Wikidata have a NPOV policy

Does Wikidata have a NPOV policy similar to w:WP:NPOV and how is it enforced? I cite the description of Q61791960 as an example that wouldn't pass muster on Wikipedia. Cheers, --SVTCobra (talk) 05:02, 24 November 2021 (UTC)

We have the rather toothless Wikidata:Living people, which is full of what noble-minded Wikidatians "should" do, and quite mum on what they must do (besides encouraging Shamima Begum to personally email Wikidata to challenge or remove certain elements). Wikidata is still largely the wild west, where almost anything goes until someone shoots you and takes your horse. Contentious labels and descriptions should be scrutinized and amended on sight IMHO, but Wikidata, sadly, fundamentally cares only about verifiable data nuggets, and as an always incomplete project full of never completable items, a certain POV can easily be introduced, even inadvertently, via selective inclusion, selective omission, selective sourcing, etc. -Animalparty (talk) 05:25, 24 November 2021 (UTC)
@SVTCobra: Since most of Wikidata is about structured data (hopefully obtained from third party sources) NPOV shouldn't generally be a question. However you are right that descriptions (and other free-form text) may have neutrality issues. Help:Description specifically states the policy on this here: "Avoid opinionated, biased or promotional wording - Neutrality is a core value of Wikidata, Wikipedia and other Wikimedia projects. Try to keep descriptions neutral by avoiding opinionated or biased terms. In addition, avoid terminology that could be interpreted as promotional. "American novelist" is a neutral description while "the best novelist in America" is both opinionated and promotional." ArthurPSmith (talk) 18:14, 24 November 2021 (UTC)

Duplicate or potential reference?

This page is about an artist, Q2778104

This other page is about an entry in a biographic reference book regarding the artist: Q21139051

Just curious whether the pages are supposed to be merged into one or if the person page should have a reference to the biliographical reference page. It would be excessive to have an entry for every single person referenced in every single biographic book so I feel like only the person page (Q2778104) existing would be the better option.  – The preceding unsigned comment was added by Joeri.Steegmans (talk • contribs) at 09:38, 17 November 2021 (UTC).

THe main reason there is that second item is that the page exists on Wikisource. Otherwise it wouldn't be needed. One could imagine that items for Wikisource pages be in a separate Wikibase installation. --- Jura 09:46, 17 November 2021 (UTC)
  • I switched described_by_source to Wierix, Hieronymus (Q21139051) in the person entry. It used to link to the entry for the entire book, now links to the specific entry with the associated text. --RAN (talk) 21:07, 20 November 2021 (UTC)
    @Richard Arthur Norton (1958- ): I have reverted your edit as it was already set correctly. described by source (P1343) is meant to point to the source publication, which in this case is the book, and then have a qualifier of statement is subject of (P805) to point to the item representing the individual Wikisource page. The original commenter was in error as the link was already used as a reference for months or years. From Hill To Shore (talk) 23:02, 20 November 2021 (UTC)
    I have updated the unsigned template to point to the correct original commenter. From Hill To Shore (talk) 23:05, 20 November 2021 (UTC)
  • Sorry! I see it now, I mistook it for "named as=Wierix, Hieronymus" per the first comment, I should have seen it was a clickable link, good catch of my error, thanks. hould all links be modelled that way? I see more that are linking directly to the Wikidata entry for the individual biography first, rather than the entry for the book first, then the individual biography. Maybe a query can be devised to detect them, and bulk change them, so we have harmonization. --RAN (talk) 23:21, 20 November 2021 (UTC)
    No problem. It partially depends on how Wikisource has formatted the book in its system. Some encyclopedia-style books have a primary page for the work (which is linked from our item on the work or our item on the edition) and then a separate page for each topical entry (which we have a corresponding item page to match Wikisource). However, other books on Wikisource (usually smaller documents) have a single page to cover both the work and the contents; in these cases we link the single page from our item on the edition. As with everything on Wikidata though there are probably many examples where we have not used the properties consistently. From Hill To Shore (talk) 00:10, 21 November 2021 (UTC)
    Thank you for the explanations. I'll be removing this question later today. Joeri.Steegmans (talk) 07:55, 25 November 2021 (UTC)

Changing property

Why can’t I change property without changing value? 217.117.125.83 16:15, 19 November 2021 (UTC)

You may be interested in Wikidata:Tools/Edit items#Move claims 2 — Martin (MSGJ · talk) 17:02, 19 November 2021 (UTC)
Note that Wikidata:Tools/Edit items#Move claims already does what #2 made possible, so no need to use that. --Matěj Suchánek (talk) 10:52, 20 November 2021 (UTC)
Thank you! 217.117.125.83 16:49, 25 November 2021 (UTC)

mixture of Dutch and Swedish

Hi, some objects have a weird label, which seems to have been copied from the Swedish label. See e.g. Q21871428 or Rauchkogel (Q21878466). In Austria alone I found > 1300 such labels: Query. Notified the bot that presumingly caused that problem, for the bot operator it seems to be fine. Is there a possibility other than doing it myself to fix such problems? best --Herzi Pinki (talk) 17:51, 25 November 2021 (UTC)

In my opinion, it's not fine. The bot is not only propagating already bad labels, it is making them worse at the same time. @Edoderoo: you should pause your bot until you fix this. Ainali (talk) 20:10, 25 November 2021 (UTC)
Huh? I already answered on this last afternoon, and got mail about it that I made an entry here. This edit mentioned above was done in 2017, which is almost 5 years ago. This has been fixed long ago too. Edoderoo (talk) 20:27, 25 November 2021 (UTC)

Regex for software version identifier (P348) needs some fixes regarding the format constraint system?

ie. the property for software version identifier (P348) does not give an error for the value 4.15-rc8 though at the same time for this value "1.0.0.4-rc4" there is a format constraint message saying that this does not match the format constraint. Is the problem that the format constraint can't handle more than 2 values separated by a radix character (Q6575377) / "." character? Oduci (talk) 16:14, 25 November 2021 (UTC) - fixed an error in my link Oduci (talk) 16:21, 25 November 2021 (UTC)

Yeah, no, I'm not touching that thing. It's rather impressive as a work of art, and asa proof that regular expressions are easier to write than read. But what's the purpose, really? Everyone makes up their own format, and if you think you have accounted for all possibilities, someone names their next version junior. Just ignore the warningL you have double-checked that the value is true, and the constraint has sort-of fulfilled its purpose. If it causes you sleepless nights, add it as an exception (or add |(\d+\.)*\d+-rc\d+ to the end of the regex). Karl Oblique (talk) 23:16, 25 November 2021 (UTC)

Closing the comment period for the Universal Code of Conduct Enforcement Draft Guidelines

Thank you for your continued comments and ideas on the Universal Code of Conduct enforcement guidelines. Your responses have helped to build a stronger Universal Code of Conduct.

If you have not already provided your comments, now is the time as the drafting committee has been meeting to update the enforcement guidelines. The drafting committee wants to consider all comments as they make their updates. Please submit any comments by the end of November. The Committee hopes to finish its revisions before the end of the year, and the revised guidelines will be published as soon as they have been completed.

The next steps for the Universal Code of Conduct include conversations about ratification of the enforcement guidelines. There will be a conversation about ratification on Nov 29.

The Wikimedia Foundation will make recommendations to the Board of Trustees about the ratification of the guidelines in December. The recommendations will inform the next steps in the Universal Code of Conduct process.--YKo (WMF) (talk) 02:37, 26 November 2021 (UTC)

Blais

Hello fellow wikipedians, I was trying to link the french (fr) page Blais to the english (en) one. It is already linked to the german (de) and italian (it) pages for Blais however it does not work, because "the same name linked to pagename Blais and in a property p1889 is mentioned as different" as reported to me by phabricator. Thank you for your attention --Ricercastorica (talk) 14:13, 25 November 2021 (UTC)

@Ricercastorica: I've moved the en wiki sitelink from Blais (Q21482871) to Blais (Q15989234), which sorts out the identified problem. different from (P1889) is doing what it says on the tin: specifying that an item for a disambiguation page is not the same as the item for a family name. It would be good, btw, to provide links when reporting issues here or elsewhere. --Tagishsimon (talk) 09:08, 26 November 2021 (UTC)

Talk to the Community Tech: The future of the Community Wishlist Survey

 

Hello!

We, the team working on the Community Wishlist Survey, would like to invite you to an online meeting with us. It will take place on 30 November (Tuesday), 17:00 UTC on Zoom, and will last an hour. Click here to join.

Agenda

  • Changes to the Community Wishlist Survey 2022. Help us decide.
  • Become a Community Wishlist Survey Ambassador. Help us spread the word about the CWS in your community.
  • Questions and answers

Format

The meeting will not be recorded or streamed. Notes without attribution will be taken and published on Meta-Wiki. The presentation (all points in the agenda except for the questions and answers) will be given in English.

We can answer questions asked in English, French, Polish, Spanish, German, and Italian. If you would like to ask questions in advance, add them on the Community Wishlist Survey talk page or send to sgrabarczuk@wikimedia.org.

Natalia Rodriguez (the Community Tech manager) will be hosting this meeting.

Invitation link

We hope to see you! SGrabarczuk (WMF) (talk) 20:03, 26 November 2021 (UTC)

"Cameroon" as unit

It seems to be a notorious problem with Wikidata's GUI. When trying to enter units as cm one gets Q1009 as unit.

"cm" is a valid alias for Cameroon (Q1009), but that it isn't useful as a unit.

Maybe I'm just the only one who still finds it odd. --- Jura 08:10, 17 November 2021 (UTC)

One could imagine various improvement to the unit search box:
1) It could prefer instance of unit of measurement (Q47574).
2) It could search inside unit symbol (P5061).
3) Long term, it could use lexemes for language-specific, grammatical forms of units.
Point 2) would reduce the need to include unit symbols as item aliases.
Point 3) would reduce the need to include plural (in some languages) or other forms as item aliases. Toni 001 (talk) 15:57, 17 November 2021 (UTC)
It could work like values for P31: while many are possible, one can easily pick "human" as value.
Of 608000 statements with width (P2049), 85% use Q174728. A further 11% m or mm. --- Jura 11:49, 18 November 2021 (UTC)
phab:T244856.--GZWDer (talk) 18:04, 20 November 2021 (UTC)
That seems to be a ticket for SDC, isn't it? Is there anything being done for Wikidata? --- Jura 12:11, 27 November 2021 (UTC)
A similar issue needing improvement is to prioritize suggestions that are instances or subclasses of the property. Thus, for family name (P734), instances of family name (Q101352) (e.g. Jackson (Q2732758)) should be prioritized over non-family name items like Jackson (Q3446260), Jackson (Q28198), or Jackson (Q959843). And for geographical properties, the inverse should be true, such that human names, disambiguation pages, etc. are not suggested for place names. -Animalparty (talk) 20:32, 26 November 2021 (UTC)

Property for dates of talks of a conference?

Let's suppose there's a conference focused on engineering. In this conference, talks will be presented on

  • Monday 6 December 2021(10:00-15:00 UTC)
  • Wednesday 8 December 2021 (10:00-20:00 UTC)
  • Thursday 9 December 2021 (08:00-13:00 UTC)
  • Friday 10 December 2021 (15:00-18:00 UTC)

Which property should I use for storing those days and hours?

I would use start time (P580) for storing the date in which the first talk is presented on Monday and end time (P582) for storing the date in which the last talk is presented on Friday. However, I don't know a property for storing the information on Wednesday or Thursday. Any help is appreciated.

Rdrg109 (talk) 05:53, 25 November 2021 (UTC)

See Help:Dates#Precision, unfortunately at the moment maximum precision of time is day. Jklamo (talk) 10:15, 26 November 2021 (UTC)
I don't think Wikidata is suitable for this type of info. --- Jura 09:25, 27 November 2021 (UTC)

Property for nationality

Is there a property representing national identity, or nationality? This is not citizenship, ethnic group, or a legal status. It is what a living person typically answers when asked their nationality. It is the nation that a historical persons identified as, is claimed by, or is defined as in sources. In the Soviet Union, nationality was noted in a person’s government ID (internal passport). —Michael Z. 20:50, 7 November 2021 (UTC)

No, but there have been several failed attempts to create such a property. Emu (talk) 23:27, 7 November 2021 (UTC)
Such Wikidata:Property proposal/Nationality in 2017 and Wikidata:Property proposal/nationality (cultural identity) in 2020. 2A01:CB14:D52:1200:E4F5:D635:3E0C:5849 18:40, 8 November 2021 (UTC)
Thank you. —Michael Z. 22:13, 13 November 2021 (UTC)

The absence of such a property is a serious problem.

It’s a problem of data representation. We have kludges like René Lévesque (Q381007) having ethnicity Quebeckers (Q245507), which is not defined as an ethnic group, but a residency status. But Quebec is officially a nation of Canada, and Quebeckers can live anywhere. It is a national identity, not necessarily a residency nor an ethnic group (in one poll, over half of Quebec residents named their ethnicity as Canadian, about a quarter French, and only 2.3% Québécois).

But much more importantly, this lack creates a systemic bias against members of stateless nations and formerly stateless nations, Indigenous nations, and nations and individuals whose identities are not defined strictly by a citizenship, a country, a faith, or an ethnicity (see w:WP:BIAS). It privileges, for example, historical empires, by forcing us to use citizenship as a proxy for national identity. It prevents us from respecting living persons’ own self-identification, and also many nations’ claims to historical figures important to their own identities.

See also national identity (Q1880695).

Would anyone support a new proposal, reframed to represent this? I think I would call it national identity to reduce confusion with nation as state. —Michael Z. 00:24, 14 November 2021 (UTC)

Oh god, no! Starting to tag people with their supposed "national identity" is a fever dream of the alt-right. Their most serious foray into philosophy is identitarism (Q105081226), after all, which inspired them to start the Identitarian movement (Q654863) in so many organizations, some of them started to get creative with the alphabet (Identity Evropa (Q30635109)) to claim their individval identity. They obviously like the concept because they don't agree with today's more liberal rules on citizenship and because the term was at some point less fraught with peril as race or ethnicity.
The concept of national identity is a mixture of heritage, culture, language, citizenship, and religion, which explains why it might be contentious, but also shows a way out, for most cases: we can already record native language (P103), languages spoken, written or signed (P1412), religion or worldview (P140), permanent resident of (P5389), country of citizenship (P27), ethnic group (P172) and permanent resident of (P5389). Record all that for someone, maybe add two or three generations of their ancestors, maybe top it of with it possessed by spirit (P4292), and see if anyone needs more information. The request, in fact, is not entirely unlike the situation that children of immigrants often complain about: that even after telling someone you were born in Billings, Montana, they'll keep asking "yes, but where are you really from".
There may be an issue with the specific cases of Quebec and the British nations. I'd suggest to wait for the latter problem to solve itself in upcoming referendums, at which point the remainder could be solved with a boolean property, Quebecois? Karl Oblique (talk) 09:03, 16 November 2021 (UTC)
None of those things is national identity unless a property indicates which one is national identity. Just picking which one you want is not a “way out” of something? It is a way to impose or deny a national identity. Choosing your own proxy property for someone is precisely saying “you’re really from where I decide you are from.”
Specific problems go back to the eighteenth century. There is no waiting for them to solve themselves.
If there’s no property, then this information will continue to be back-channeled in the description field anyway. —Michael Z. 20:25, 18 November 2021 (UTC)
not only "back-channelled" but also argued and holy-warred there. Nationality (non-legal) is a very sensitive and political topic. It's also often a private information whose de-anonimization can be considered dangerous. --Infovarius (talk) 16:06, 26 November 2021 (UTC)

@Mzajac, Emu, Karl Oblique: As this is a recurring topic I created Wikidata:Nationality. Multichill (talk) 11:27, 21 November 2021 (UTC)

Citizenship and ethnicity would cover a lot of modern people, but what about feudal and tribal societies where group membership isn't really the same as "citizen"? I recently stumbled on Tecumseh (Q257808)country of citizenship (P27)Shawnee (Q253436) permalink ⁓ Pelagic ( messages ) 03:09, 28 November 2021 (UTC)

@Pelagic: Yes, country of citizenship (P27) doesn’t really make sense before the notion of the modern nation state. Even Roman citizenship (Q213810) isn’t really a citizenship in the modern sense. Yet we do use this property in a very liberal sense and even tolerate that many values are unsourced and highly dubious. (On a sidenote: More than 13.000 items claim a citizenship of Austria-Hungary (Q28513) that never existed. We accept that too because there is little harm. The potential for harm and Wikidrama is much higher for “nationality”.) Emu (talk) 18:03, 28 November 2021 (UTC)

Bot work for Cleveland Museum of Art

Hi, Thousands of files from the Cleveland Museum of Art were uploaded by Madreiling (not active since August 2019), and Wikidata items created for them. However,

Could it be possible to fix that with a bot? i.e.

  1. Add the WD Q number in the Artwork template;
  2. Complete the WD item with information on Commons (at least the creator);
  3. Upload a JPEG version with the highest possible resolution;
  4. Add the JPEG version to the P18 property;
  5. Mark the JPEG version as preferred rank in WD;
  6. Use a Creator template for the Artist in Commons.

The list of item can be accessed via [39], [40], etc. See also related discussion at c:Commons talk:Structured data#CC-0 conflicts with PD statement. Thanks, Yann (talk) 09:30, 27 November 2021 (UTC)

Points #1, #3, #5 seems to be a matter for Commons.
Maybe it's worth creating a status of the available SDC data and complete it from the template on the files.
SDC then allows easy sync with Wikidata items. A dedicated property could be used instead of "described by url".
Here is what's available on Wikidata: https://w.wiki/4TYk
Depending on the type of work, Wikidata's model is more or less developed and maintained.
Some of the items still have Q18593264 that should eventually be removed. --- Jura 10:01, 27 November 2021 (UTC)
Yes, all images from this category. Yann (talk) 12:48, 28 November 2021 (UTC)
When fixing a few painting items, I noticed that they were mostly created by @Dominic: part of some project.
Maybe ask them before adding unfinished Wikidata parts to Wikidata:Bot request. --- Jura 08:37, 29 November 2021 (UTC)

Catalonian republics

Did all these partially/fully unrecognized states have Barcelona (Q1492) as their capital? If so, should Barcelona (Q1492) have inverse claims (as Property:P1376#P2302 implies)? Please judge whose version is correct in our conflict with User:Lopezsuarez. --Infovarius (talk) 21:16, 27 November 2021 (UTC)

There is no conflict. It is totally absurd to say that Barcelona was the "capital of the Catalan Republic" in 2017, when Catalonia has never been a republic, not in 2017 or ever. Lopezsuarez (talk) 21:57, 27 November 2021 (UTC)
Is the claim ok? I am talking about property constraints too. --Infovarius (talk) 11:43, 28 November 2021 (UTC)
If there is a reference supporting one claim, it's applicable to the other as well. One can then set appropriate ranks. --- Jura 08:48, 29 November 2021 (UTC)

Latin language Label

Hello,

Currently Pons Aelius (Q7227963) lack a Latin language Label. Could somebody correct this? Also do the same for every Roman military base, item with Digital Atlas of the Roman Empire ID (P1936), item with vici.org ID (P1481). 2A01:CB14:D52:1200:3950:38ED:E8B2:4846 20:58, 28 November 2021 (UTC)

You can edit the label at Special:SetLabel/Q7227963/la. --- Jura 08:45, 29 November 2021 (UTC)

P1476 (title)

Many movies have their original title in English language, even when it's, let's say, German produced movies. For example Knockin' on Heaven's Door (Q166355). Should the property title (P1476) then show as language German or English? And if English is correct, would it be the same for Korean movies and series. Many of them have English titles, but they are almost always written in Korean alphabet Hangul (Q8222), e.g. Glitch (Q107646135). --Christian140 (talk) 07:26, 27 November 2021 (UTC)

Are these situations where the film is redubbed in a new language? If so, create one item to hold the original version details and then a second item with the details of the translation. The two items can be linked by has edition or translation (P747). See Snow White and the Seven Dwarfs (Q134430) as an example. From Hill To Shore (talk) 09:15, 27 November 2021 (UTC)
In my view title (P1476)Knockin’ on Heaven’s Door@en is correct; the language code denotes the language of the title value string and has no real relationship with the country of production nor the country in which the film is marketed under that title. If you wish to show the location in which a particular title was used, there is presumably a qualifier available. If you want to show the country of production, there is a property for that. And obvs, the item can have more than one title value.
In other news, I've changed the heading of this question. Christian140, if you use a template for a heading, then it is not possible to navigate to the thread directly from a watchlist, and indeed the watchlist does not show the resolved header, and so one needs to remember or lookup P1476 in order to figure out which index entry to select to get to the thread; or use a diff. It's hideous & wrong. --Tagishsimon (talk) 08:13, 29 November 2021 (UTC)
And then with Korean films? Many Korean films (or books or TV series) have English titles, but are spelled in Korean letters. Should they also be shown as English. I sometimes did it like that. Like, the original title of Glitch (Q107646135) is just the English word "glitch", but spelled in Korean instead of Latin letters. --Christian140 (talk) 19:43, 29 November 2021 (UTC)

need alternative for "parent organization" property

Civil Air Patrol (Q781210) currently contains the parent organization (P749) property with the value Air Combat Command (Q407015). But it's not that simple. CAP is a not-for-profit organization chartered by Congress, and is able to conduct many operations on its own authority. Certain operations and policies, such as searching for downed aircraft and setting a policy for wearing uniforms, require approval from Q108444184, which is indeed part of Air Combat Command (Q407015). Is there any property that covers this kind of relationship? Jc3s5h (talk) 16:53, 27 November 2021 (UTC)

regulated by (P3719) maybe? Karl Oblique (talk) 19:51, 27 November 2021 (UTC)
So far "regulated by" seems better than "parent organization", but I'll give some time for others to respond. Jc3s5h (talk) 20:10, 27 November 2021 (UTC)
General alternatives might be owned by (P127) or operator (P137) but they don't seem applicable here. ArthurPSmith (talk) 17:56, 29 November 2021 (UTC)

Wikidata weekly summary #496

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week
  • Development
    • Mismatch Finder: Continued working on last remaining tickets for the first version. Added a footer to the site, improved documentation and added ability to delete a batch of mismatches.
    • Made good progress on migrating our on-wiki Vue apps to support the new Vue.createMwApp compatibility layer in MediaWiki core (phab:T294465)
    • Continued work on making it possible to define a custom ordering of grammatical features on Lexemes (phab:T232557)
    • More research and discussion on mul language code (phab:T285156)
    • Discussing with data re-users about their views on the ontology issue classification we worked on earlier this year to get their input (slides from Data Quality Days session)

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.

Nominating MoveClaim to become a Special:Gadget

Following recent proposal and enaction of Compact View becoming a member of Special:Gadgets, I thought I'd also nominate User:Matěj_Suchánek/moveClaim.js which is currently used by many (~492) users. It seems sensible as a very commonly convenient operation and since there already exists a special gadget for moving sitelinks. --SilentSpike (talk) 00:46, 24 November 2021 (UTC)

+1 Salgo60 (talk) 01:10, 24 November 2021 (UTC)

Help! My Mullvad (Q47008412) VPN ip got through(I did not expect this), you need to block my IP. Please replace it with Oduci, please scrub my IP and replace it with Oduci. Thank you.

I was in the process of creating a new item for Wikiversity lesson (Q109794522) when I realized I was allowed by the system to create the item using an IP address by Mullvad (Q47008412) VPN service provider. The reason this mistake happened was because I was confused and thought I was using a browser I use for Clearnet (Q20706905) browsing, but instead I was using the browser that is connected through my Mullvad (Q47008412) VPN. The edit was allowed and the rest is history. I am not a researcher, I am just reporting my IP in knowledge that your policies when I last read them regarding proxies and VPNs that they are not allowed to make edits with.

Is it possible to replace the Mullvad (Q47008412) IP with my username that was used in this edit at https://www.wikidata.org/w/index.php?title=Q109794522&oldid=1535729837 ? There is no privacy issue and I feel no need for any urgency on a personal level as I am not ashamed that I am using a VPN provider to browse the net with(not make edits with! How was I to know that an edit would be allowed, but I assume this happens rarely anyway...).

To my knowledge Wikimedia has policies against proxies and VPNs and if one wants to make real change maybe contribute to Wikiversity projects and research projects which do research on finding solutions to combat spam so that one day maybe VPNs can go through a filter that is strong enough that most bad edits will be impractical to perform. VPN service providers could also potentially find technical solutions(ie. allowing editors with possession of accounts with a good history of edits to use a special IP that is not given to most VPN users for Wikimedia edits) and VPN customers can demand from their VPN providers that this is a need they have = to make edits on Wikimedia projects and asking for solutions. I am also aware that users can apply for special permissions from Wikimedia personnel but I assumed that is only in special cases where one might be in danger from a government in case one contributes to Wikimedia projects and I live in Sweden and I consider Sweden to not be such a country where I need to fear for my life when I make edits here so I do not think things like check-user exemption applies to my situation.

P.S. I am making this edit using Mullvad (Q47008412) for it to be easier for you to block. If something has changed in the policy that now allows edits from VPNs that will be news to me but I think I would know if a change had happened because I have much interest in any developments in this area.

If there is an easier and maybe better way to report such occurrences in the future please enlighten me as most words I wrote as part of this text was more about my self conscious shock that this could happen.(an ip "slipping through"). I hope this of help and not a bother. Thanks again! Oduci (talk) 20:27, 28 November 2021 (UTC)

@Oduci: I've hidden the IP address but we cannot reattribute the edit to your account. In the future please follow the directions at Wikidata:Oversight rather than posting about it in public. If you have a compelling need to have Wikidata:IP block exemption, then please file a request at WD:RFP accordingly.--Jasper Deng (talk) 19:45, 30 November 2021 (UTC)

When you have been around Wikidata for a while, you learn what linked datasource each type of entry should have for example a podcast needs links out to the major players and directory as well as social media links. A politician made needs links to various political listing sites. An actor should be linked to IMDB etc etc. Should we (or do we have) wiki pages that suggest the various places to look for links for different types of wikidata entrys? Back ache (talk) 10:28, 30 November 2021 (UTC)

Yes. Wikiprojects often have a list of properties that pertain to types of items they work with. For an example, see Wikiproject Music. There is also Wikidata:Schemas, which is a long-term in-progress solution that offers significant benefits over these lists of properties. Lectrician1 (talk) 14:12, 30 November 2021 (UTC)
Once you’ve set a instance of (P31), suggestions for further properties adapt, although I am not sure if it’s specific enough to then further adapt to, say, musicians vs. politicians.
Theres also recoin, which you can enable in your gadget settings which gives you similar information in a list above the item. Karl Oblique (talk) 21:08, 30 November 2021 (UTC)

Q1758677 (oeuve)

Q1758677 is described as a disambiguation page but that's only true for the English Wikipedia page; the other five are no dps. The item should be split but I am not sure about the best way to do this. Is removing the link to the enwiki, blanking the remaining descriptions and deleting the disambiguation statement an acceptable solution?

Btw, WikiData has no Helpdesk? Not sure where to post questions. A redirect from 'Helpdesk' to 'Project chat' could be, well… helpful.  →bertux 14:27, 30 November 2021 (UTC)

No, one shouldn't repurpose Wikidata items. Just move stuff that belongs elsewhere to another item, new or existing. --- Jura 17:25, 30 November 2021 (UTC)
So which part should be moved, enwiki or the other ones?  →bertux 17:44, 30 November 2021 (UTC)
Check the P31 value to determine. --- Jura 17:49, 30 November 2021 (UTC)
(EC) Sitelinks to any pages that are not disambiguation pages should be moved elsewhere. Perusing the history of the item, its first (& only?) P31 value was disambiguation (even though the item may have started with only a sitelink to a non-dab page). So, it's a dab item and as Jura rightly says, such things should never be changed. --Tagishsimon (talk) 17:50, 30 November 2021 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2021/11&oldid=1592683005"