Wikidata

Wikidata:Project chat/Archive/2020/07

< Wikidata:Project chat‎ | Archive‎ | 2020
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

part and whole: merge or not?

This edit by an IP brought these two items to my attention: trick-taking game (Q1191150) and trick (Q11755381). Upon translating the Polish article at Q11755381, I was able to identify that it deals with the concept of tricks, while the English article at Q1191150 is about games that employ tricks, so they are conceptually different, though very similar. I tend to think that they should not be merged, but would like to hear other opinions. If we keep the items separate, a couple of articles will have to be moved from Q1191150 to Q11755381: de, et, nl, no, pt and maybe stq. —capmo (talk) 23:47, 29 June 2020 (UTC)

@Capmo: If you added part of (P361)-has part(s) (P527), it looks like you are just vetoed merging possibility by yourself. --Liuxinyu970226 (talk) 00:09, 30 June 2020 (UTC)
Liuxinyu970226, my edits can be easily reverted if necessary. I just wanted a confirmation that I'm doing the right thing. —capmo (talk) 02:26, 1 July 2020 (UTC)

Entity classification with P31

I've been wondering about this for a while and am not sure if there are other situations where this comes up.

In the UK hills are often given somewhat "arbitrary" classifications based on various features (height, drop and region typically) as decided by various hillwalkers/organisations (all well documented - such as Munro (Q1320721)). Currently in Wikidata this is reflected using instance of (P31), see Ben Nevis (Q104674) for example. You can find some basic documentation of the most well known in Scotland on my user page.

Does this seem like the correct way to capture these classifications? It feels a little iffy to me, but I struggle to distinguish why exactly. It means that if classifications exist which aren't related through a class hierarchy (most of them aren't) then an item can arbitrarily be an "instance of" many - none of which feel the same as saying it's an instance of a mountain. Any thoughts? --SilentSpike (talk) 18:57, 27 June 2020 (UTC)

  • As a follow up point, I would add that I've considered these could be modelled inversely (a property on the classification item specifying a query to return all items to which it applies). However, this seems non-trivial for cases (e.g. Wainwright (Q62082131) or Munros which must be regarded by the Scottish Mountaineering Club as distinct and separate mountains) where the classification includes more arbitrary points and not factual information about the entity which would be stored in Wikidata. --SilentSpike (talk) 19:07, 27 June 2020 (UTC)
  • I'd like to see these treated as a classification scheme (Q5962346). We would need:
    1. a new item for "United Kingdom hill classification system" <has parts of the class> "United Kingdom hill class"
    2. a new item for "United Kingdom hill class" <part of> "United Kingdom hill classification system"
    3. Munro (Q1320721), Simm (Q26709966), etc., would then become <instance of > "United Kingdom hill class" rather than just class (Q16889133)
    4. Once everything is in place, propose a new property "United Kingdom hill class" and use that property rather than P31.
We have a great many properties for fairly specific classification systems already (see search results). - PKM (talk) 19:21, 28 June 2020 (UTC)
  • I like the idea of using this as a classification scheme in a specific property (as @PKM: suggests) rather than using instance of (P31) - in general we try and keep this sort of thing in properties other than P31. Maybe a more generic "classification by size" property would be good, rather than one limited to hill classes, though? It feels like it might be useful in general; eg we could consider using it for the various classifications of cities by size, which are also sometimes challenged as not quite right in P31. Andrew Gray (talk) 18:24, 30 June 2020 (UTC)
    • @PKM, Andrew Gray: Yes I also like this idea and am considering whether it can be generalised further (or whether we want to - because you could then end up in a situation where there can be statements grouped together mixing multiple systems). The other thing is I'm not entirely sure there's a formalised "UK hill classification system" so much as just informal designations which have become popular and don't really fall under one system. --SilentSpike (talk) 09:53, 1 July 2020 (UTC)

Removing from the watchlist

How do I remove Wikidata:Requests for deletions‎ from my watchlist? Pmt (talk) 20:32, 30 June 2020 (UTC)

For any page you wish to remove from your watchlist, visit the page. Near the top of the window, a little right of center, you will see a solid blue star; it's right next to the box to search Wikidata. Click on the star and the star will be filled with white instead of blue. This indicates it is no longer on your watchlist.
Conversely, for any page you want to add to your watchlist, click the white star; it will turn blue and be on your watchlist. Jc3s5h (talk) 21:20, 30 June 2020 (UTC)
Is there a way to edit the whole watchlist? Levana Taylor (talk) 03:49, 1 July 2020 (UTC)
Yes, Special:EditWatchlist. Ghouston (talk) 04:07, 1 July 2020 (UTC)

Merging entries about the African diaspora: " Project:Make Wikipedia blacker (Q63870968)" and " WikiProject African diaspora (Q15304953) "

I am trying to merge " Project:Make Wikipedia blacker (Q63870968)" Q63870968 (French, Spanish, and Catalan wikipedias) and " WikiProject African diaspora (Q15304953)" Q15304953 (English and Portuguese wikipedias) as all of the projects refer to the African diaspora, but the conflicts prevent a merge.

Would it be alright if steps were taken to combine these?

Thank you WhisperToMe (talk) 03:43, 30 June 2020 (UTC)

What do you plan to do to Wikidata:WikiProject Ennegreciendo Noircir Wikimedia and Wikidata:WikiProject African Diaspora? Visite fortuitement prolongée (talk) 16:05, 30 June 2020 (UTC)
It seems like someone should suggest a merge -- they seem like they cover the same ground but in different languages (like the respective WP projects). Calliopejen1 (talk) 00:12, 2 July 2020 (UTC)
I am not sure this should be merged. I see the point about covering the same ground (so useful for links), but people in the project might object being seen as 1 single project, especially when it come to governance and WMF representation (since the rule is usually 1 representative per project). Have you discussed with the projects about that ? --Misc (talk) 09:18, 2 July 2020 (UTC)

Problems with item submission

I've been trying to create a page and it was disallowed because of the "specific spam items". Where can I learn what exactly is considered a spam item?  – The preceding unsigned comment was added by 93.100.66.124 (talk • contribs) at 17:28, 2 July 2020 (UTC).

Looks like you tripped Special:AbuseFilter/132, which is a non-public filter (I don't have access to it, either). User:Jasper Deng, is there any advice you can give without compromising the effectiveness of the filter? Vahurzpu (talk) 18:32, 2 July 2020 (UTC)
@Vahurzpu: The filter worked as designed; the item they tried to create was really spammy: (English) label "Changelly", (English) description "Instant cryptocurrency exchange platform that acts as an intermediary between exchanges and users.".--Jasper Deng (talk) 18:56, 2 July 2020 (UTC)
@Jasper Deng: Okay, thanks. Good to know that the filter's working properly. Vahurzpu (talk) 19:00, 2 July 2020 (UTC)
@Vahurzpu:
@Jasper Deng:

Thank you, however I was wondering if there is anything the mentioned company can do to get this item published

interwiki links not appearing on Wikimedia Commons category

Tracked in Phabricator
Task T257266

I'm not sure whether this issue is related to Wikidata or Wikimedia Commons, but the interwiki links are not showing up on Commons:Category:Kyburz, California even though the corresponding Wikidata item links to the correct pages elsewhere. I tried to re-link the category to a Wikipedia article but got this message: "The page you wanted to link with is already attached to an item on the central data repository which links to Category:Kyburz, California on this site." I removed the links on Wikidata and re-added them, but this did not solve the problem.

Any idea what's going on? --Ixfd64 (talk) 21:21, 6 July 2020 (UTC)

@Ixfd64: thanks for the report, I opened T257266 to track this issue. --Lucas Werkmeister (WMDE) (talk) 22:17, 6 July 2020 (UTC)
Update: the issue has been resolved for now by reverting some config changes. (The task isn’t marked resolved yet, but the remaining work probably doesn’t matter to this discussion, as long as we don’t break sitelinks again when switching from the reverted config to the real solution.) --Lucas Werkmeister (WMDE) (talk) 07:43, 7 July 2020 (UTC)
I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Lucas Werkmeister (WMDE) (talk) 07:43, 7 July 2020 (UTC)
I can not see the Wikidata link on Commons in Q5364187, does it mean the fix has not been implemented, or is this another issue?--Ymblanter (talk) 13:01, 7 July 2020 (UTC)
I don't see it there either, so it's not a caching issue specific to Ymblanter. - Jmabel (talk) 15:04, 7 July 2020 (UTC)
Try [1] - it works after purging for me. Thanks. Mike Peel (talk) 15:10, 7 July 2020 (UTC)
Sorry, I forgot to mention this here – yes, affected pages need to be purged. --Lucas Werkmeister (WMDE) (talk) 15:22, 7 July 2020 (UTC)
Yes, it works for me now, thanks a lot.--Ymblanter (talk) 17:19, 7 July 2020 (UTC)
I can confirm that the issue is resolved for me. Thanks! --Ixfd64 (talk) 18:58, 7 July 2020 (UTC)

Guam

Anybody interested in helping sorting up the administrative division of Guam (Q16635)? Do we have a Wikiproject page for Guam?

Guam is divided into 19 "villages". But it does not look like we have any item describing this administrative division, even though we have an item for every county in all of the states of the US. We have Category:Villages in Guam (Q7293178), but it does not look like we have any "main namespace-item" for them. 62 etc (talk) 10:08, 1 July 2020 (UTC)

There is w:Villages of Guam linked to a list item. --- Jura 10:12, 1 July 2020 (UTC)
They're not all modelled in exactly the same way (e.g. some have instance of (P31):village (Q532) and others instance of (P31):village (Q751708), but we can get the list of all 19 via 
SELECT ?item ?itemLabel WHERE {
  ?item wdt:P31/wdt:P279* wd:Q532; wdt:P131 wd:Q16635 .
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
ORDER BY ?itemLabel
Try it!
I suspect we want a new item for "village of Guam" (which is a subclass of municipality (Q15284) or something in that hierarchy), which each of them can be set to, and then have list of Villages of Guam (Q1258753) as a list of that, etc. That can all be automated fairly easily, and I'm happy to help out with it if that would be useful. --Oravrattas (talk) 10:29, 1 July 2020 (UTC)
Good! Be aware of ceb:Kategoriya:Mga subdibisyon sa Guam which seem to be a category of all 19 administrative villages while ceb:Kategoriya:Mga lungsod sa Guam is about populated places within these villages, often with the same name. 62 etc (talk) 11:48, 1 July 2020 (UTC)

Software change

Hello, all,

The mw:New requirements for user signatures will begin on Monday, 6 July 2020. This is a change to MediaWiki software that will prevent editors from accidentally setting certain types of custom signatures, such as a custom signature that creates Special:LintErrors (such as <span>...<span> instead of <span>...</span>) or a signature that does not link to the local account.

Few editors will be affected. If you want to know whether your signature (or any individual editor) is okay, you can check your signature at https://signatures.toolforge.org/check You are not required to fix an invalid custom signature immediately. Starting Monday, editors will not be able to create new invalid signatures to Special:Preferences. Later, we will contact affected editors. Eventually, invalid custom signatures will stop working. There will be an announcement in m:Tech/News then. You can subscribe to m:Tech/News. You can also put mw:New requirements for user signatures on your watchlist.

I don't know if you have a help page or policy related to custom signatures; if you do, it's possible that it might need some small changes. If you have questions, then please ping me or ask questions at mw:Talk:New requirements for user signatures. Whatamidoing (WMF) (talk) 03:46, 2 July 2020 (UTC)

To check your directly: https://signatures.toolforge.org/check/www.wikidata.org/Jmabel
--- Jura 06:41, 2 July 2020 (UTC)
Thank you both. I have just fixed the URL. Whatamidoing (WMF) (talk) 02:51, 3 July 2020 (UTC)
An overall report for Wikidata is also available, as well as overall reports for some other wikis (but not all). --Lucas Werkmeister (WMDE) (talk) 12:38, 3 July 2020 (UTC)

Position held / occupation

Hi all! Has the difference between the position held (P39) / occupation (P106) changed? An editor regularly deletes position held (P39) properties ([2], [3], [4], [5], [6] etc) and rewrites them to the occupation (P106) property - in a lot of items (for example: [7] -> [8]). Is this correct or can it be undone?

(I corrected some earlier, but he withdrew it) See also: Wikidata:Database reports/Constraint violations/P1047 Palotabarát (talk) 08:20, 2 July 2020 (UTC)

  • @Royalty & Nassau Expert, Looperz: --- Jura 10:27, 2 July 2020 (UTC)
  • That definitely looks wrong to me - the specific "Bishop of X" is definitely a position. There's nothing wrong with saying "position held: Bishop of X" and "occupation: bishop" (or clergyman, or priest, etc), but removing position entirely is a problem. This is a very confusing way to handle it and it's not how we normally deal with positions. Andrew Gray (talk) 11:22, 2 July 2020 (UTC)
  • It's possible that some of this is down to the long-running issue as to the scope of P39. The labels and descriptions for the property still differ quite dramatically across different languages, with some restricting it specifically to political positions; some saying political and ecclesiastical; some saying public office; some basically any position at all (to presumably include things like Chief Executive of companies, etc.) --Oravrattas (talk) 12:24, 2 July 2020 (UTC)
I don't care anymore about what you think is right or wrong. I discovered that this database is corrupted with more fake information than I can correct. So go ahead, do whatever you like. I know when a case is a hopeless one. I won't contribute to Wikidata or Wikipedia anymore. So you won't see edits from me anymore. My watchlist goes empty, my e-mail address will be deleted. So I won't read what you will do. Royalty & Nassau Expert (talk) 17:01, 2 July 2020 (UTC)

Hello dear fellow contributors. and thank you for pinging me, Jura. I am sad to read about a conflict that caused somebody to be disappointed about or even quit his voluntary workload. But that's only one aspect and this was my first impression. Royalty & Nassau Expert did not want to accept any definition of the difference between those two things "I'm sorry but I'm not responsible for the fact that the database is wrongfully making distinction between occupation and position held." was something that you were able to read on his Discussion-Page before he deleted everything there only few minutes ago. Another cite is this "I don't care anymore about what you think is right or wrong." just one paragraph above. There are many ways to react on things you do not understand or do not agree with;

  1. keep away
  2. read the discussions and start new ones even trying to convince others
  3. try to find a coexistence of your solution and the other one (thats what often happens to special extra declared Bishop positions, some of them have extra entries, most of them dont)
  4. try to erase all other work because you are convinced your way of seeing things is the only correct one

Nobody has problems with methods 1 and 2. and in my brave moments i sometimes take even method 3, because using SparQL for statistical things i depend on certain entries. But taking option °3 is stressful enough for me, because this does not end the discussions that pop up from time to time. Taking option °4 on purpose and repeatedly is the no-go-area, and imho reason enough to ban a user temporarily, because this is the way we start to work against each other. That is not the way this beautiful project is able to flourish.

This all seems to be also a little bit about terminology and that could possibly differ from language to language. (Btw.: Look for wp-articles "Bishop of XYZ" that are mixed up with. "List of Bishops of XYZ" of another language.) Getting that international differences under control is our challenge.

In my opinion the difference between position held (P39) and occupation (P106) is already defined precise enough. Roughly and shortened: For a occupation (P106) you have "only" to be qualified enough (Including some times high level degrees like PhD and regional Limits) For a position held (P39) you have to be appointed/elected or in religious leading positions sometimes even to be consecrated. What happened here was the beginning of an edit-war. And thus I am glad, that somebody surrendered "nearly" at the right time. --Looperz (talk) 18:04, 2 July 2020 (UTC)

If you think you know better and make such a stupid distinction between occupation and position held for medieaval persons, be my guest. You apparently haven't understood much about history then. But don't tell me what I should do or don't do. I did read the "instructions". But I'm not going to uphold stupid instructions made by people that have no clue.
You don't like my contributions? Fine with me. Be so kind to undo them ALL. I don't care. Like I said, Wikidata is a hopeless case. I'm not going to put anymore time and energy in it. I would welcome it very much if all my edits will be erased from here. Then we both will be happy, you can carry on with more utter rubbish in the database, and I don't see my name as contributor to it. Royalty & Nassau Expert (talk) 00:32, 3 July 2020 (UTC)

street name sign

Hello. It useful (or do we already have/using a property) to have an image with the street name sign (Q1969455) that have the name of a person, to person's item?

The only item I have found is Simone Rapin (Q3484599), that uses image (P18) with qualifier depicts (P180) -> street name sign (Q1969455). Is that the right way to add the street name sign to person's item?

Xaris333 (talk) 10:28, 2 July 2020 (UTC)

For described by source (P1343), allow section, verse, paragraph, or clause (P958) under statement is subject of (P805)

Suppose, for described by source (P1343) of BookA in a Wikidata item, we want to link s:BookA/volume 1 and s:BookA/volume 2#section III. A user have to type, under described by source (P1343) of BookA: in statement is subject of (P805), [volume 1, volume 2] and, in section, verse, paragraph, or clause (P958), [<empty>, section III]. Few user know they must type an empty value. If we make a section, verse, paragraph, or clause (P958) option under statement is subject of (P805), then a user would type, in statement is subject of (P805) under described by source (P1343) of BookA, volume 1 and [volume 2, section, verse, paragraph, or clause (P958), section III]. It will make data structure clearer.--The Master (talk) 02:06, 3 July 2020 (UTC)

Q about web.archive urls

I just added some references to a painting from the Christie's auction record which was already archived to the wayback machine, so I used that as a ref here. I just noticed though that you can't expand the web.archive url to get the rest of the text. Is there a way to save the whole url to the web.archive? Q is Archduke Leopold Wilhelm in his Gallery in Brussels (Q19960949) and Christie's record with expandable text I would like to cite is here. Jane023 (talk) 08:24, 3 July 2020 (UTC)

Merge « reference url » and « url »

Why do we really have url Search both with reference URL (P854)   ? This caused at least one time confusion for newcomers, and I see no real good reason to keep both : When used as a reference url Search cannot be something else as the url of the reference.

I’d propose to merge them. I know this has been like that for a very long time, but it’s never too late to do good. author  TomT0m / talk page 11:21, 2 July 2020 (UTC)

OK with that. Thierry Caro (talk) 12:38, 2 July 2020 (UTC)
@TomT0m, Thierry Caro: The correct way to propose a merge is WD:PFD.--GZWDer (talk) 15:22, 2 July 2020 (UTC)
@GZWDer: I wanted to check wether there could be a chance community would agree before taking it a step further. These are widely used properties. author  TomT0m / talk page 15:39, 2 July 2020 (UTC)
Sounds like a good idea. Cheers, VIGNERON (talk) 17:50, 2 July 2020 (UTC)
  Support merging these. - PKM (talk) 22:51, 2 July 2020 (UTC)

@Ghouston, Dipsacus fullonum: Actually I think atm. the model has some unatural feature. I think for example that « official website » is a sane statement to make about a company or a person, but for other kind of entities like, say « website », it’s pretty unnatural to me to use « official website », but « url » is natural. For a press or academic article as well. author  TomT0m / talk page 07:39, 3 July 2020 (UTC)

Does Wikidata still import from Wikipedia?

Hi, noob here. Apologies if this is in the archives, I've done some searching.

My understanding is that Wikidata was in the past populated with data (such as place coordinates) from Wikipedia. Today, Wikipedia users are saying Wikidata should never be used in Wikipedia because supposedly the following can happen:

  1. someone incompetently creates a place article in Wikipedia with wrong, say, coordinates
  2. this place article is imported and becomes a Wikidata item
  3. the wrong data then is propagated to other Wikipedias (i.e. other languages).

Could this have happened in the past and can it still happen ? --Cornellier (talk) 23:16, 2 July 2020 (UTC).

There are also cases where a statement have the wrong value imported because the correct value doesn't have it's own WP article. --Trade (talk) 23:44, 2 July 2020 (UTC)
@Cornellier: If we are dealing with something not otherwise cited: the issue is exactly the same as if Wikidata isn't an intermediary. If material in Wikipedia in some language is uncited and possibly erroneous, and it is copied to another Wikipedia, then that other Wikipedia gets uncited and possibly erroneous information. The situation is literally the same if Wikidata is an intermediary. So it all comes down to what "should never be used" means. Are the same users saying articles should never be translated from one Wikipedia to another? Are they making any distinction between material that is cited from a reliable source and material that is not? - Jmabel (talk) 23:55, 2 July 2020 (UTC)
Thanks. Sure anything can be replicated, good or bad. But my question is this: is the exchange of data between Wikipedia and Wikidata synchronous or asynchronous? If one goal is to create a centralized db, then one would hope for the latter, surely? --Cornellier (talk) 00:08, 3 July 2020 (UTC)
There is no single answer. Wikidata is used to support different Wikimedia projects depending on the consensus of those Wikiprojects. Commons, for example, makes use of a lot of Wikidata to populate entries on paintings, creators and category infoboxes (see this Commons image as an example; almost all the text is imported from Wikidata). Some language versions of Wikipedia use Wikidata to populate their infoboxes, including automatically importing the image file used on the corresponding Wikidata item. For many languages with a low number of editors, the priority is to harness the power of Wikidata to generate content faster than they can do so on their own; the quality of Wikidata sourcing is also usually either on par with those projects or sometimes better than them. In other more developed and popular projects though, such as English Wikipedia, there is less trust in the quality of data compared to the quality of information already in their own project. Automatically importing Wikidata's still developing content into the more developed articles of a more developed Wikipedia would reduce the quality of that Wikipedia's content for a time, until the source data at Wikidata is corrected. Despite the lack of trust in automatic updates between Wikidata and English Wikipedia, enwiki does use Wikidata to generate non-article content, such as en:Wikipedia:WikiProject Women in Red/Missing articles by nationality/United Kingdom. From Hill To Shore (talk) 00:34, 3 July 2020 (UTC)
Agree, and I could give many examples of enwiki using wikidata. But my question is the other way around. Does Wikidata still source info from enwiki? --Cornellier (talk) 01:49, 3 July 2020 (UTC)
It happens that Wikidata contributor import information from Wikipedia, with or without its references. That Wikipedia has incorrect information happens, just like studies in journals can be retracted. The advantage of Wikidata is that it keeps track of correct and incorrect information. --- Jura 06:01, 3 July 2020 (UTC)
Yeah, there is some js widget that help with automated creation. While I can't speak for all items creation, based on the ongoing effort to revert the mass deletion of sexual orientation (P91) at the end of May 2020, the majority of those imports were correct. Based on the current work page on Wikidata:WikiProject_LGBT/Unsourced_sexual_orientation#Hard_to_source_and/or_false, we had around 8 where we couldn't find any sources at all, except if we count the bunch of folks listed as heterosexual imported from russian WP by 1 single user, which, given the climate in the country and stats around the topic, would be likely correct. 2 months after, we are still working on fixing the unannounced mass deletion, but clearly, the error rate was quite low (8 out of ~2000, we are in the 0.5% range, and only because we assume things to be false unless proven otherwise). Now there was some smaller errors (people listed as lesbian when they were bi, etc), but we didn't track those, and the modeling is a bit messy because that's a messy topic. --Misc (talk) 12:11, 3 July 2020 (UTC)

Export API-Commands from OpenRefine

Is it possible to export the API commands from OpenRefine, who are generated when the edits are uploaded to Wikidata. And then use a bot to run these commands instead of using OpenRefine or is there a batch mode in OpenRefine. --Hogü-456 (talk) 17:51, 3 July 2020 (UTC)

@Hogü-456: are you familiar with the undo-redo tab, where you can export a series of operations as a JSON file? You can use this to re-apply a workflow on a new version of a dataset. This requires that the new dataset has the same structure as the original one (same column names, for instance). We are working on improving this function. There are also third-party tools to run OpenRefine workflows from the command line: https://github.com/opencultureconsulting/openrefine-client − Pintoch (talk) 09:47, 4 July 2020 (UTC)

Open for signatures - Community open letter on renaming

Dear all,

There is an open letter that requests a pause to renaming activities being pursued by the Wikimedia Foundation 2030 Brand Project.

Individual editors and affiliates can sign with their logged-in account to show support.

The letter focuses on concerns about the process, and not about specific naming choices. With 50 major chapters and affiliates and 600+ individuals signing the statement, we are seeing great interest in this issue.

Related to this: the branding team is conducting a survey that runs until July 7. There is concern that the consultation process and options on the survey do not adequately reflect community sentiment, given the effect name changes for the foundation and movement would have. This served as a motivation for the open letter. Useful links are below:

  • Brand survey for individuals - Qualtrics survey. If there are options you would like to highlight outside of the three provided, it is possible to write in your own options and views at the end of the survey.

There will be a WMF board meeting scheduled in July to discuss the branding issue, so it is important to express your views now.

Thanks - Fuzheado (talk) 13:27, 3 July 2020 (UTC)

The unexpected consequences of a consensus

A tool was deemed to be problematic because of the number of duplicate ORCiD identifiers. As a consequence the tool is no longer available and as a consequence nobody has a commitment anymore to maintain duplicate ORCiD identifiers.. I had committed myself to maintain this as the prize for having the tool. We all lost because of the consequences of this regrettable consensus. Thanks, GerardM (talk) 09:23, 4 July 2020 (UTC)

Merged.--GZWDer (talk) 14:32, 4 July 2020 (UTC)
@GZWDer: What was merged to what? - Jmabel (talk) 15:58, 4 July 2020 (UTC)
When it is that simple, what was all the fuss about ? Thanks, GerardM (talk) 16:59, 4 July 2020 (UTC)

Creation of many duplicates

It seems that indefinitely blocked user User:MrProperLawAndOrder has created a large number of duplicates of persons. I looked at this batch and noticed dozens of duplicates (at least as far as Polish actors go). Is this batch worth keeping ? Kpjas (talk) 06:42, 3 July 2020 (UTC)

Seems to be the case here. 5000 items in the batch, 5420 created. --- Jura 08:20, 3 July 2020 (UTC)
I checked items with duplicated Deutsche Biographie (GND) ID (P7902) and found only one.--GZWDer (talk) 11:29, 3 July 2020 (UTC)
@GZWDer: Tadeusz Teodorczyk (Q95860697) vs Tadeusz Teodorczyk (Q9354938), Zbigniew Bielski (Q95876750) vs Zbigniew Bielski (Q9387771), Janusz Bukowski (Q95827880) vs Janusz Bukowski (Q7944766) and the list goes on and on. I have no idea how many there are new items by MrProperLawAndOrder but I presume there'd be some hundreds of duplicates. Kpjas (talk) 18:06, 3 July 2020 (UTC)
It's a known problem with their data dump. Epidosis and Bargioni did some cleanup, but there are still thousands left. Maybe someone assumed that Wikidata has only entries for people with GNDs defined as that identifier seems to be somewhat mandatory on dewiki. --- Jura 09:52, 5 July 2020 (UTC)
  • The duplicates I mentioned above, sometimes cause constraint violations like VIAF or ISNI. Would it be practicable for QS to stop the offending batch after some threshold is reached ? Kpjas (talk) 12:08, 3 July 2020 (UTC)

Item update request

Please add https://www.thecollectionbook.info/ on Microsoft Windows (Q1406) as reference site. Thanks!!! --151.49.98.40 18:54, 4 July 2020 (UTC)

Publishing Updated List of DTH Channel List-India

Dear Community,

I have done the review research on the DTH Channels currently being broadcasted in India along with their Owning Network name, Genre, Type, Language and current tarrif which reciprocal to their TRP, and also their channel number (may be holding good only for SES-7, Airtel Digital TV in India).

This review data collected and up-to-date as on 5th July 2020.

Kindly help me in publishing this for community benifit.

Thanks and Regards, Sowmya T.S. Sowmya T S (talk) 11:06, 5 July 2020 (UTC)

Q1381844

Q1381844 (Montoya) is currently a mix of interwikis for disambiguation pages and surname articles. Can these two be split into a page for diambiguation page interwikis and a page for surname article interwikis? (some of the language pages need adjustment as well to properly indicate the content of their pages)

ie.

-- 65.94.169.71 09:53, 5 July 2020 (UTC)

Contemporary constraint where father dies during pregnancy

Hi. I have worked through hundreds/thousands of biographical items and I occasionally stumble across a case where the father dies during the mother's pregnancy. The birth date of the child after the death of the father then produces a contemporary constraint warning. See Thomas Humphrey Sneyd (Q75921838) and Averil Marion Anne Sneyd (Q76237432) as an example. Is there a way that we can adjust the constraints on father (P22) and child (P40) to include a 10 month leeway? I suggest 10 months to account for late births; if there are some rare human births that go beyond 10 months and the father has already died, the relevant item could be listed as an exception to the constraint. From Hill To Shore (talk) 14:04, 5 July 2020 (UTC)

That is a good idea; 10 months is plenty but calculation is complicated because dates are often only precise to a year. It is possible to add the qualifier "sourcing circumstances: posthumous" but this doesn't stop the constraint being triggered either. Levana Taylor (talk) 14:41, 5 July 2020 (UTC)
what about fathers from donated sperm though? BrokenSegue (talk) 23:50, 5 July 2020 (UTC)
OK then, we should simply make it so that the constraint is satisfied if there's a qualifier "posthumous." Levana Taylor (talk) 00:01, 6 July 2020 (UTC)

I have a question, please help.

I have amassed a pretty serious body of sources that back a particular topic (newspaper articles, documentaries, etc.) Unfortunately, the rules that we have in Russian Wikipedia demand that "only references that are absolutely necessary for the article must be kept", so like 200+ sources that we painstakingly collected are about to be dumped into trash. I have a suspicion that that's the type information juuuuuust right for Wikidata, can you please fill me in on the nodes that I could use? -- Wesha (talk) 05:30, 6 July 2020 (UTC)

I don't think anyone would be upset about a ton of references here but I think you'll find the hard part will be encoding all the information in a way that wikidata can represent. Not sure what you mean by "nodes". Also adding certain kinds of citations to wikidata is more laborious because you have to make items for all the things (e.g. newspaper article) you want to source from. But if you want to put in the effort I doubt anyone will stop you. BrokenSegue (talk) 06:22, 6 July 2020 (UTC)
One thing to consider here is that unless the Russian Wikipedia purges that article's history, all of your work will remain in previous versions and can be retrieved later. The results of your hard work may be more difficult to locate, but it is unlikely to be destroyed. From Hill To Shore (talk) 06:55, 6 July 2020 (UTC)

Property:P599 format change

(Follows up Property talk:P599#URL is not working and w:cs:Diskuse k šabloně:ITF female profile.)

  Notified participants of WikiProject Tennis

Apparently, the URL format of ITF player ID before 2020 (archived) (P599) was changed (as reported, not for the first time). While old links still work (they are redirected), entries for new tennis players are not reachable.

For example, Serena Williams (Q11459) has "20007765", for which formatter URL (P1630) generates https://www.itftennis.com/procircuit/players/player/profile.aspx?playerid=20007765, which redirects to https://www.itftennis.com/en/players/serena-williams/800205424/usa/wt/s/overview/. The new value ($1 variable for P1630) is serena-williams/800205424/usa/wt, not a trivial change.

However, entries for new players get values starting with 8 (observation) and the old system doesn't work for them. For example, Linda Fruhvirtová (Q95568176) has "800485856", for which P1630 generates https://www.itftennis.com/procircuit/players/player/profile.aspx?playerid=800485856, which yields 404. The correct link is https://www.itftennis.com/en/players/linda-fruhvirtova/800485856/cze/jt/s/overview/ and I can't see any trick that would generate a working (redirected) link.

It is not possible to make the property "smart" and always generate correct links (and I like that because it keeps Wikidata simple), so we can only fix this by changing the property scheme and update all values (by a bot). Or create a new property.

What do you see as the best and most considerate process which ultimately results in working links for everyone in the universe? --Matěj Suchánek (talk) 11:48, 6 July 2020 (UTC)

  • There are really just two approaches:
(1) either we come up with a way to versionize property definitions (an approach suggested at project chat earlier, but didn't find any support).
(2) we create a new property when the definition changes. That we continue to support or delete the old property is a separate question.
I think we had too many cases where some thought that merely "updating" some values was a good idea (or might have been) and turned out problematic. --- Jura 12:10, 6 July 2020 (UTC)
    • In my opinion as this introduced a new type of ID which is a breaking and incompatible change upstream, a new property is deserved.--GZWDer (talk) 16:03, 6 July 2020 (UTC)

Wikidata weekly summary #423

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week
    • Ptable displays the periodic table automatically extracted from information provided by Wikidata; it also provides a check that all the elements are there with some basic properties. Additional pages provide charts of the nuclides under different criteria such as half-life. Each element or nuclide is linked to its Wikidata item for more information or to edit if necessary.
  • Development
    • More work on the consistent design system (Phabricator board)
    • More work on decoupling the different Wikibase extensions from each other
    • Polishing the first step of Federation (using Wikidata's Properties in another Wikibase installation) (incl. preventing users from selecting a federated property with a non-supported data type (phab:T252012) and preventing users from accessing Special:NewProperty when federation is enabled (phab:T255576) and viewing a list of all properties when federation is enabled (phab:T246339))
    • Continuing research and interviews around the topic of making it easier to access Wikidata's data for programmers
    • Doing first testing of mockups and prototypes of the first version of the Query Builder - coding can start soon
    • Convert a few properties from string to external identifier: Linguasphere code (P1396), KOATUU identifier (P1077) and ISIN (P946)

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.

How do I indicate that a photograph has been made aboard a (particular) ship?

[9] Right now it complains that "location" is of a wrong type. I tried searching for "aboard", "on board of" and the like byt to no avail. Help? -- Wesha (talk) 05:33, 6 July 2020 (UTC)

I'm unsure what the right solution here is. We could just modify location (P276) to allow the location to be a mode of transport (Q334166). "Aboard" seems a bit niche. BrokenSegue (talk) 06:27, 6 July 2020 (UTC)
Try location of the point of view (P7108). Thanks. Mike Peel (talk) 06:57, 6 July 2020 (UTC)
How about using location (P276) with qualifier statement is subject of (P805) [location is a mode of transport]. That could also be used as a qualifier for properties giving other locations like filming location (P915), place of marriage (P2842), place of birth (P19) etc. --Dipsacus fullonum (talk) 07:37, 6 July 2020 (UTC)
Doesn't location of the point of view (P7108) have the same problem (I ask because I'm trying to better understand how property constraints work). BrokenSegue (talk) 08:46, 6 July 2020 (UTC)
Property constraints are set separately for each property, constraints on location (P276) aren't necessarily the same as those on location of the point of view (P7108). I think in the latter case, it makes sense to include non-geographical items. However, it may need discussion on the talk page. Thanks. Mike Peel (talk) 16:43, 6 July 2020 (UTC)
I think the constraints for Commons shouldn't be defined at the same time as those for Wikidata. See Wikidata:Property_proposal/property_constraint_for_Commons for ways to solve this. --- Jura 16:57, 6 July 2020 (UTC)
At sea Quakewoody (talk) 13:25, 7 July 2020 (UTC)

Label menu

Hello. Is there a way to see only the English label and the label of preferred language in each item? Now, I can see two more language (because of my location I guess). Data Gamer play 13:10, 7 July 2020 (UTC)

Use {{#babel:en|de-2|...etc...}} somewhere in your user page. Paweł Ziemian (talk) 15:02, 7 July 2020 (UTC)
Thanks. Data Gamer play 15:07, 7 July 2020 (UTC)

Property proposal: Dowker-Thistlethwaite name

Hello! I'm still trying to get approval for my proposal at Wikidata:Property proposal/Dowker-Thistlethwaite name. If granted, this would be an extremely useful property for items about w:prime knots, as it would give each an unambiguous identifier, something which is not possible with any other knot naming system. I've also got a bot run ready to add this identifier to some 2900+ items. This would also provide a unique key for database comparisons for these objects, and also a key into an authoritative external database.

I've put a lot of work into getting all this ready -- data complitation and cleaning, manual merging, and bot writing -- and it would be a shame for the project to stall half way through.

At the risk of being a bore, I'd very much like to get this property through the approvals process, so I can finish the project. Would anyone like to take another look at it? -- The Anome (talk) 13:23, 7 July 2020 (UTC)

P1423 on module items?

I found some module items e.g. Module:WikidataCheck (Q15089555) do have template has topic (P1423), currently P1423's description says "primary topic of the subject template or infobox", and subject type constraint (Q21503250) says class (P2308)Wikimedia template (Q11266439) and even constraint status (P2316)mandatory constraint (Q21502408). So what should I do? Request another property "module's main topic"? Or apply Wikimedia module (Q15184295) to subject type constraint (Q21503250) qualifiers? --Liuxinyu970226 (talk) 00:07, 8 July 2020 (UTC)

@Avatar6: That example was added by you, am I wrong? --Liuxinyu970226 (talk) 00:09, 8 July 2020 (UTC)

Feedback on movement names

Hello. Apologies if you are not reading this message in your native language. Please help translate to your language if necessary. Thank you!

There are a lot of conversations happening about the future of our movement names. We hope that you are part of these discussions and that your community is represented.

Since 16 June, the Foundation Brand Team has been running a survey in 7 languages about 3 naming options. There are also community members sharing concerns about renaming in a Community Open Letter.

Our goal in this call for feedback is to hear from across the community, so we encourage you to participate in the survey, the open letter, or both. The survey will go through 7 July in all timezones. Input from the survey and discussions will be analyzed and published on Meta-Wiki.

Thanks for thinking about the future of the movement, --The Brand Project team, 20:33, 2 July 2020 (UTC)

Note: The survey is conducted via a third-party service, which may subject it to additional terms. For more information on privacy and data-handling, see the survey privacy statement.

  • I think it would have been hard to make the "consulation" in a way that's more insulting to the idea of consultation. There's no reason why the status quo isn't on the list despite the WMF thinking that their proposal is so unpopular that they know that nobody wants it and that it's necessary to circumvent democratic consultation for it. There's also no reason why all the options aren't shown to the user before voting on individual options. ChristianKl ❪✉❫ 07:44, 3 July 2020 (UTC)

Awards which have been revoked

award received (P166) allows us to record when people receive particular awards, but what happens when an award is removed or revoked? In particular David Starkey (Q1176720) was awarded the Medlicott Medal (Q17126380) in 1980 and it was withdrawn on 3 July 2020. Richard Nevell (talk) 16:01, 5 July 2020 (UTC)

You could use end time (P582) to indicate the date when the statement stops being correct (the date the award is cancelled or withdrawn). From Hill To Shore (talk) 16:12, 5 July 2020 (UTC)
@Richard Nevell: Also end cause (P1534) to indicate that it was revoked. --SilentSpike (talk) 14:55, 8 July 2020 (UTC)

Current highlights

How do you add a item to Current highlights? --Trade (talk) 07:26, 8 July 2020 (UTC)

Unsourced family

Somebody created a batch of items about some family members:

They does not have any sources (other than Wikia, which is also unsourced), and can not be related to any notable people (or at least, someone who is described by serious source). I want comment from community about how to deal with them.--GZWDer (talk) 18:49, 8 July 2020 (UTC)

Q82781158 seems to have a large number of id's and is claimed as founder of Q89033283, so presumably there's a structural need and validity there - although that could be disputed on notability grounds too I think. The others seem to have been entered just because of the familial relationship and I think should be completely removed based on our policies regarding living people. ArthurPSmith (talk) 19:00, 8 July 2020 (UTC)
While Q82781158 does currently have 16 external identifiers, pretty much all of them except Google Knowledge Graph are their own social media/website accounts. Crunchbase and Linkedin entries can be created by anyone for themselves. In addition, the Internet Game Database person ID (P5796) one leads to a 404 page. And the notability of Q89033283 is also questionable, since all external IDs were likely created by themselves and the only references are their own website, LinkedIn and Crunchase (which again were likely created by them). --Kam Solusar (talk) 23:59, 8 July 2020 (UTC)

format for ISNI

Please change the format constraint for ISNI ID to accept two different formats--it currently only accepts the ID number broken into 4 groups of 4 with spaces between, but VIAF for one gives the number without spaces. Levana Taylor (talk) 13:07, 8 July 2020 (UTC)

What I meant was, it should be like dates where you can enter it in multiple formats and have it stored and displayed in unified form. Why in the heck should someone who's pulling an ISNI number from VIAF, without spaces, have to hand-insert spaces? Anyhow, thanks for pointing me to the discussion page, I guess further comments should go there. — Levana Taylor (talk) 23:17, 8 July 2020 (UTC)
@Levana Taylor: currently you can use either format for input. A bot eventually converts it to the standard one. --- Jura 07:51, 9 July 2020 (UTC)

Change of "located in the administrative territorial entity" (P131) to accept any populated place as value

There is some discussion about this at Property talk:P131. Please comment there. --- Jura 07:58, 9 July 2020 (UTC)

Announcing a new wiki project! Welcome, Abstract Wikipedia

Hi all,

It is my honor to introduce Abstract Wikipedia, a new project that has been unanimously approved by the Wikimedia Foundation Board of Trustees. Abstract Wikipedia proposes a new way to generate baseline encyclopedic content in a multilingual fashion, allowing more contributors and more readers to share more knowledge in more languages. It is an approach that aims to make cross-lingual cooperation easier on our projects, increase the sustainability of our movement through expanding access to participation, improve the user experience for readers of all languages, and innovate in free knowledge by connecting some of the strengths of our movement to create something new.

This is our first new project in over seven years. Abstract Wikipedia was submitted as a project proposal by Denny Vrandečić in May 2020 [1] after years of preparation and research, leading to a detailed plan and lively discussions in the Wikimedia communities. We know that the energy and the creativity of the community often runs up against language barriers, and information that is available in one language may not make it to other language Wikipedias. Abstract Wikipedia intends to look and feel like a Wikipedia, but build on the powerful, language-independent conceptual models of Wikidata, with the goal of letting volunteers create and maintain Wikipedia articles across our polyglot Wikimedia world.

The project will allow volunteers to assemble the fundamentals of an article using words and entities from Wikidata. Because Wikidata uses conceptual models that are meant to be universal across languages, it should be possible to use and extend these building blocks of knowledge to create models for articles that also have universal value. Using code, volunteers will be able to translate these abstract “articles” into their own languages. If successful, this could eventually allow everyone to read about any topic in Wikidata in their own language.

As you can imagine, this work will require a lot of software development, and a lot of cooperation among Wikimedians. In order to make this effort possible, Denny will join the Foundation as a staff member in July and lead this initiative. You may know Denny as the creator of Wikidata, a long-time community member, a former staff member at Wikimedia Deutschland, and a former Trustee at the Wikimedia Foundation [2]. We are very excited that Denny will bring his skills and expertise to work on this project alongside the Foundation’s product, technology, and community liaison teams.

It is important to acknowledge that this is an experimental project, and that every Wikipedia community has different needs. This project may offer some communities great advantages. Other communities may engage less. Every language Wikipedia community will be free to choose and moderate whether or how they would use content from this project.

We are excited that this new wiki-project has the possibility to advance knowledge equity through increased access to knowledge. It also invites us to consider and engage with critical questions about how and by whom knowledge is constructed. We look forward to working in cooperation with the communities to think through these important questions.

There is much to do as we begin designing a plan for Abstract Wikipedia in close collaboration with our communities. I encourage you to get involved by going to the project page and joining the new mailing list [3]. We recognize that Abstract Wikipedia is ambitious, but we also recognize its potential. We invite you all to join us on a new, unexplored path.

Yours,

Katherine Maher (Executive Director, Wikimedia Foundation)

  1. ↑ Abstract Wikipedia
  2. ↑ User:Denny
  3. ↑ Abstract Wikipedia mailing list

Sent by m:User:Elitre (WMF) 19:56, 9 July 2020 (UTC) - m:Special:MyLanguage/Abstract Wikipedia/July 2020 announcement

Missing query result

SELECT ?AncientCivi ?AncientCiviLabel WHERE {
  ?AncientCivi wdt:P31 wd:Q28171280.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!

It returns 23 results. However non include Ancient Rome (Q1747689) which should be a query result by my understanding. Am I missing something or is it some form of bug? --Daanvr (talk) 18:39, 7 July 2020 (UTC)

@Daanvr: The reason Ancient Rome (Q1747689) is not in the result set is because it has a preferred-rank P31 value of historical country (Q3024240). I don't understand why that was set though. ArthurPSmith (talk) 19:32, 7 July 2020 (UTC)
This sort of question is becoming more common, and each time it comes up the consensus seems to nudge another little bit closer to "In most cases, no P31 statement should be marked as preferred". I don't if we've quite reached the point where we should actively remove lots of those preferreds, but this one doesn't seem like it has any strong reason to be so, so I've restored 'historic country' to normal rank. --Oravrattas (talk) 15:00, 8 July 2020 (UTC)
Is it possible to edit the query to include all results regardless of the rank? (maybe there are more results "missing" the same way as Ancient Rome (Q1747689) was "missing")--Daanvr (talk) 20:05, 9 July 2020 (UTC)
Of course:
SELECT ?AncientCivi ?AncientCiviLabel ?rank WHERE {
  ?AncientCivi p:P31 ?stmt.
  ?stmt ps:P31 wd:Q28171280.
  ?stmt wikibase:rank ?rank.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!

--SCIdude (talk) 07:02, 10 July 2020 (UTC)

Linking wikipedia categories to wikidata

Hi, I am new to editing wikidata directly and have been discussing wikidata items for ice hockey venue categories with J 1982 at User_talk:TSventon#Ice hockey venues without reaching agreement. I have two questions and hope this is the right venue:

  • is there any policy on which wikidata item to link a wikipedia category to (and where could I find it)?
  • are there any processes for resolving disagreements in Wikidata? TSventon (talk) 14:29, 9 July 2020 (UTC)
    • @TSventon: Wikipedia categories should be sitelinks under a Wikidata category item, see for example Category:Danish poets (Q7020217). Disagreements should be raised on item talk pages, user talk pages, or if two people cannot come to agreement possibly here on Project Chat. If it's a behavioral issue bring it to the Administrator's Noticeboard. ArthurPSmith (talk) 18:03, 9 July 2020 (UTC)
    • Ah, and on your specific question, in general Wikidata items should have a single conceptual meaning - see Help:Items - which allows us to unambiguously add statements about them. While interwiki linking is a side-benefit of this through the attached sitelinks, linking wikipedia pages about different things is to be avoided; the meaning of the item is most fundamentally defined by those sitelinks (if there are any). Things may be a bit more ambiguous for Category pages though, I haven't seen that discussed here before. ArthurPSmith (talk) 18:12, 9 July 2020 (UTC)
@ArthurPSmith: thanks for the link, it suggests that the other user is probably right. TSventon (talk) 08:37, 10 July 2020 (UTC)

Smaller RDF representations of entities

When requesting a resource in such a way that RDF is returned, e.g. https://www.wikidata.org/wiki/Special:EntityData/Q11442.ttl, the response contains a lot more data than we need. Notably, all the properties of the entity come with all labels in all languages. All in all, the file has 1.5MB and takes a long time to parse for us. I know we could use the SPARQL endpoint but plain http/linked data seems easier and I guess would consume less server resources, so:

Is there a way to request a more concise description, e.g. restrict the label language(s) or fetch no labels or even no property info at all?  – The preceding unsigned comment was added by Fkleedorfer (talk • contribs) at 13:39, 10 July 2020‎ (UTC).

@Fkleedorfer: You can add ?flavor=dump to exclude the “stub” data of other entities. This is used by the query service updater, for example. --Lucas Werkmeister (WMDE) (talk) 13:59, 10 July 2020 (UTC)
@Lucas Werkmeister (WMDE): Thanks! Your example is about 10% the size of the document we were fetching before, so that will definitely help. This does not seem to work for representations requested via 'Accept' header content negotiation, though. (still, this might be a good solution for us) Fkleedorfer (talk) 14:16, 10 July 2020 (UTC)

Interlanguage link alphabetical sort

Tracked in Phabricator
Task T257625

Hello all- Can anyone tell me why the interlanguage links for Q92062 are not listed in alphabetical order at en:Motte-and-bailey castle? I don't think I've noticed this issue elsewhere, and don't see any way to address it in the settings. Didn't find anything in a search of Wikidata guidance (searching on strings such as sort, alphabetic, order, etc). Thanks in advance for any ideas. Eric talk 14:34, 10 July 2020 (UTC)

There have been some other reports of this since yesterday, but it’s not yet clear where the issue is coming from. --Lucas Werkmeister (WMDE) (talk) 15:22, 10 July 2020 (UTC)
Danke, Lucas! Eric talk 16:28, 10 July 2020 (UTC)

allergist (Q2837942)

I have added the above occupation (P106)to Nathan Lass (Q30526904)and created Category:Allergists in Commons. But then noticed that Category:Allergologists already exsits in Commons. The label says allergist just as the article Allergist in en.wiki but wasn't sure what is correct her. Geagea (talk) 02:03, 11 July 2020 (UTC)

Do islands need to have separate items for its administrative territory?

I have been thinking about this problem for several weeks while editing articles on islands in Indonesia on Wikipedia. I realized that a lot of districts and administrative villages have their whole territory comprised of only an island or an archipelago often with the same name. For its usage in an article, I don't think there has to be two separate articles for an island and its administrative unit since they would contain very much the same information most of the time. In Wikidata, there are a lot of items already created for islands in Indonesia, mostly related to Lsjbot activity in the Swedish and Cebuano Wikipedia importing of Geonames toponym database. I have found this annoying many times because there are a lot of islands which are also an administrative units, with their own separate items already created. From here, I looked for existing examples but found that it's not that uncomplicated either. The island of Nauru for example has items for the country (Q697) and the island (Q30151675) even though both occupy the same area and the latter doesn't really have much info to add. In contrast, other items are for an island and its territorry like Saint Helena (Q34497) and Isle of Wight (Q9679). What should I do? Is there already an established guidelines about this? To me, I don't think there should be separate items because both are the exact same thing and I have found these instances a lot in Indonesia and it would only make things more complicated. RXerself (talk) 13:05, 5 July 2020 (UTC)

  • The established way of doing things is to have separate items. They are not the same thing as for example inception (P571) of both is different. ChristianKl ❪✉❫ 13:31, 5 July 2020 (UTC)
    • We can use qualifiers for it. Either for statement instance of (P31) or inception (P571). Estonia (Q191) as the modern sovereign state uses start time (P580) as qualifier while it uses its first republic establishment date in 1918 for inception. Meanwhile, Japan (Q17) uses multiple inception date, two for its contitutions and one for the traditional foundation date. RXerself (talk) 05:27, 6 July 2020 (UTC)
      • That approach makes it very difficult to query the data, however, particularly when it's not done consistently. For example, getting a list of all countries at a given historical date is effectively impossible from Wikidata at the moment, in large part due to this sort of modelling. --Oravrattas (talk) 05:32, 6 July 2020 (UTC)
        • Is the problem more about using qualifiers or the inconsistent use of different qualifiers among the items? RXerself (talk) 11:46, 6 July 2020 (UTC)
          • At the moment it's a little hard to tell. It's perhaps possible that we could come up with, and apply, a consistent way of doing all this with qualifiers, and document all the gotchas to avoid when querying. However, as at least some Wikipedias will often split these sorts of things across different items in at least some cases, and thus we will end up with separate items anyway, the general direction of travel is almost always going to be towards keeping distinct things distinct. --Oravrattas (talk) 21:15, 6 July 2020 (UTC)
          • Yeah, I am planning to propose the same thing in Wikipedia, but since I'm mostly active in the Indonesian WP, the articles in other Wikipedias like say, w:en:Hawaii County, Hawaii and w:en:Hawaii (island) would most likely still be separated without community's input (a discussion to merge the two didn't get any reply since 2013). But there are already island/s and its administrative unit of same area in a single article like w:en:Isle of Wight or w:en:Texel (Q9966). There would be issues when the article may talk about both subjects while the Wikidata item linked would only be one and only contain the information about the island or the administrative unit since the other info would be on a separated item. In these cases users also would have to choose whether to link it to the item for the island or the administrative unit and that is not that simple, just like the problem I encountered. RXerself (talk) 01:47, 11 July 2020 (UTC)
            • That's often referred to as the "Bonnie and Clyde problem", and Help:Handling sitelinks overlapping multiple items talks about how to handle it. The basic principle is that you need three items: one for the island, one for the administrative unit, and one to represent Wikipedia articles that talk about both. You then link each Wikipedia depending on which approach it took: the ones that have separate items link to each Wikidata item separately; the ones that combine both into a single page, link to the combined item. I have to frequently take this approach for items about elections, for example. It's fairly common for French and Spanish Wikipedia to have separate pages for a presidential election and a legislative election in a country that happened on the same day, whilst most others usually combine both into a single "General election" page. That means that in Wikidata we need to have separate items for all three, and then link them up accordingly. --Oravrattas (talk) 05:01, 11 July 2020 (UTC)
              • What information would the third combined item contain? If it's the same with the two, we would create duplicates. If it's empty or only contains one of them, the article wouldn't show up in some queries. RXerself (talk) 08:10, 11 July 2020 (UTC)
                • I believe that the "third, combined item" should generally be minimal, containing little more than links to the corresponding Wikipedia articles, and links to the constituent entities.
Personally, I'd like additional options for linking to the constituent entities, though. Bonnie and Clyde (Q219937) uses has part(s) (P527) to link to Bonnie Parker (Q2319886) and Clyde Barrow (Q3320282), and that's fine. It's less clear how to link a hypothetical physical-or-administrative-Nauru to either Nauru (Q30151675) or Nauru (Q697). Another example I've worked on (without fully satisfactory results) is calorie (Q87260855), which is really either small calorie (Q130964) or kilocalorie (Q26708069). —Scs (talk) 17:20, 11 July 2020 (UTC)
Say, we created a combined item for the archipelago of Selayar Islands (currently Q14619 and Q3289762) like calorie (Q87260855). The problem rises should I make a query for example, "List of regencies in Indonesia with its article in Javanese (jv) Wikipedia," using value regency of Indonesia (Q3191695) for instance of (P31) like the other regencies, the link for jv.wp (or any wp) would come up for other regencies but wouldn't come up for Selayar Isls since it's only linked to the combined item which doesn't use Q3191695. If we add Q3191695 to the combined item, it would create a duplicate. RXerself (talk) 17:35, 12 July 2020 (UTC)
I don't completely understand your example, but it sounds like you need to adjust your query to cope with cases where the Wikipedia page isn't specifically about the item, but might also be about a combination that includes the item. In the "Bonnie and Clyde" example, searching for Wikipedias with an article specifically about "Bonnie" is different from searching for articles of which "Bonnie" is either the main focus, or a part. --Oravrattas (talk) 18:56, 12 July 2020 (UTC)
  • Generally, yes: so please avoid merging them. Area is frequently different. The impact for countries or similar is obviously different than for the thousands of islands in Indonesia. --- Jura 13:32, 5 July 2020 (UTC)
    • For example Cuba and Madagascar consists of multiple islands, where the main island is only a part. An village may also contains a main island and other smaller islands nearby.--GZWDer (talk) 13:39, 5 July 2020 (UTC)
      • I haven't merged any iirc because I want to know the precedence. No, the islands and administrative units I'm talking about are only the ones that occupy exactly the same area. I am aware of countries or regions which have the same name but comprising different area. Such regions exist too in Indonesian administrative units and I am fine with their items being separated. RXerself (talk) 05:27, 6 July 2020 (UTC)
        • I still say two items, linked with coextensive with (P3403). - Jmabel (talk) 16:04, 6 July 2020 (UTC)
          • This would be a workaround but another issue rises when a Wikipedia article which talks about both the island and the administrative unit needs to be linked to only an item. (I described further above). RXerself (talk) 01:42, 11 July 2020 (UTC)
        • I agree it sometimes seems silly to have two different entities for "the same thing", but in the case of islands which are also municipalities, it really does make sense. Islands are physical features; municipalities are political. Islands are (usually) permanent; municipalities can come and go.
I believe it's also traditional to link the two with different from (P1889), although this bothers me, since in most senses they're "the same". —Scs (talk) 18:00, 6 July 2020 (UTC)
I have run into the same issue with the bot as well which creates separate articles for human settlements and political entities, such as Q22407536 and Amriswil (Q69768). While I understand that these can be separate entities, most people living there would not consider them separate and it creates a lot of confusion. So expect that at some point a bot will add a third entity that describes the human settlement (item 1), on the island (item 2) located in the administrative municipality (item 3) which all have identical names and area. --Hannes Röst (talk) 18:27, 6 July 2020 (UTC)
  Support separate items for the geographical islands and administrative territories "islands", btw, Hong Kong should also have separate items, one for the geographical Hong Kong areas (and describe its country (P17) values histories), another for the SAR, Macau ditto. --Liuxinyu970226 (talk) 23:58, 7 July 2020 (UTC)
Another nice (?) example is the island of Nantucket in Massachusetts, for which we have a town (Q18372627), a county (Q2991355), an island (Q49149), and a census-designated place (Q12185344). (I'm not sure what the distinction between the town and the CDP is. The county and the island are not actually coterminous, as the county includes two small neighboring islands, Tuckernuck Island (Q4000168) and Muskeget Island (Q2912934).) —Scs (talk) 12:31, 12 July 2020 (UTC)

Duplicate entries (geonames/Wikipedia)

I have come across an issue where there are duplicate Wikidata entries for a lake in Ontario: Chartrand Lake (Q22395320) and Chartrand Lake (Q22483765). I cannot merge them because each of two duplicate items has a Wikipedia entry in arz and sv language (bot entries). How should we proceed in a case like this? Just use said to be the same as (P460) would be one solution but actually there should only be a single Wikidata item. Any ideas? --Hannes Röst (talk) 18:19, 6 July 2020 (UTC)

a) Are you sure that the Wikipedia entries are the same? I can't speak the languages. Just wondering, the Wikipedia entries may are different thing.

b) If are the same, leave a message in Project chat in each Wikipedia, asking if the two articles are the same thing. If they are, they are going to merge them (one of them will be a redirect). After that, only one Wikidata item will have a Wikipedia entry so you will be able to merge them.

Data Gamer play 13:08, 7 July 2020 (UTC)

They look the same to me. As bot-created entries, I doubt that anybody on those wikis would notice/object if one was redirected to the other. Ghouston (talk) 01:01, 8 July 2020 (UTC)
I believe that a good way it is to ask to these Wikipedias, as I have written. They will correct the problem. Data Gamer play 01:13, 8 July 2020 (UTC)
What leads to conclude that this is the "best way"? --- Jura 05:47, 8 July 2020 (UTC)
Sorry, I changed it. Data Gamer play 11:31, 8 July 2020 (UTC)

In my experience, nobody from such 'script/bot-powered' Wikipedias as the arzWP will have objections if you merge duplicated articles into one without special discussions, with leaving a summary describing your actions of course. Contributors will surely be even thankful since bots or similar programs are doing mistakes from time to time. After that you can merge items here at Wikidata.
Starting the local discussions is long way since we don't know how active they are, how they are 'tolerant' towards discussions in English, etc. --Wolverène (talk) 12:34, 8 July 2020 (UTC)

w:sv:Wikipedia:Bybrunnen#Same lake? and w:arz:ويكيبيديا:صالون المناقشه#Same lake?. I try it, lets see. Data Gamer play 17:02, 8 July 2020 (UTC)

After my question, svwiki merge the two articles. Now, I am waiting arzwiki. Data Gamer play 17:44, 8 July 2020 (UTC)
Problem solved. Both wikis merged the articles. So, we can merged Wikidata items now. Data Gamer play 21:30, 8 July 2020 (UTC)
Thank you, but note that there are may be hundreds of such duplicated articles. --Wolverène (talk) 07:01, 9 July 2020 (UTC)
If you can, make a list please. Data Gamer play 12:27, 9 July 2020 (UTC)
Too long and boring, pls just look the content of categories like sv:Kategori:Robotskapade artiklar 2016-01 or sv:Kategori:Robotskapade artiklar 2016-02‎, then look for couples of pages like sv:Andrews Lake (sjö i Kanada, Ontario, Cochrane District, lat 48,59, long -80,14) & sv:Andrews Lake (sjö i Kanada, Ontario, Cochrane District, lat 49,30, long -81,30). There are definitely many of them. Unfortunately, the arzWP is also choosing (relatively speaking) the 'Cebuano Wikipedia path', so such mistakenly-made articles will continue to be created on and on. --Wolverène (talk) 12:49, 9 July 2020 (UTC)
Any suggestions to solve the problem? Data Gamer play 14:21, 9 July 2020 (UTC)
There are a lot of items created this way. I frequently found them when editing villages and islands in Indonesia which already has an article in a wiki and a Wikidata item but then data from Geonames are imported for bot-created articles in one wiki and created duplicates. And then someone else uses another bot to add articles to another wiki and created more duplicates. For example, this search would return several items of villages, some are the same but in different items linked only to different wikis. I don't know how to handle this other than listing and merging them, but problem like this is related to the use of bot just adding new items while creating the articles in the wikis. But I don't know how to make it to detect if there is already an item related to the subject either. RXerself (talk) 02:38, 11 July 2020 (UTC)
We need to encourage and empower bot maintainers to teach their bots to detect existing articles so that they don't create these duplicates. —Scs (talk) 17:24, 11 July 2020 (UTC)
My experience of svwiki is that it is not very welcomed that users who do not know the language tries to merge articles there. But requests like those Data Gamer made in the Swedish village pump are always welcome and are normally done within a few hours. The clean up project after Lsjbot does progress, but will take time. Se Special:Contributions/Kitayama. 62 etc (talk) 12:52, 11 July 2020 (UTC)
Fyi: Among Swedish Wikipedians, Lsjbot is typically considered to have done a great job on creating Swedish articles about all animals on the earth. The Lsjbot project on creating articles about all places in the world was interupted, and a lot of manual work has been done on solving issues in the articles. On Swedish Wikipedia, a project has started to slowly delete some of the geographical articles that were created by Lsjbot, if they have not been edited by any human, and only have interwiki link to Cebuano, and not are expected to be improved.Tomastvivlaren (talk) 15:10, 11 July 2020 (UTC)
Personally, I mostly good experiences with User:Matěj_Suchánek/markasduplicate.js. Articles eventually get merged (sometimes days, months, if not years later .. varies from one language to another). I mostly use it with items for people and I'm generally certain that it's the same person. --- Jura 11:25, 12 July 2020 (UTC)

Vandalism again

Today, around 01:30 UTC, an Australian Ip vandalized an item on Artemisia Gentilleschi [10]. The data from the item are used in commons:Creator:Artemisia Gentileschi (referenced from the item), which is transcluded in 50 files on Commons. At 5 UTC, another Australian IP nominates one of these files for deletion, citing a wrong info in the template commons:Commons:Deletion requests/File:Corisca and the Satyr by Artemisia Gentileschi.jpg. I happened to have one of the relevant pages on Commons on my watchlist, but otherwise the file (I hope) would be kept anyway, but may be in a week or may be in three months, and vandalism would stay intact all the time (or possibly noticed by someone who has no idea about Wikidata transclusions). Whereas this is an anecdotal case, it shows that vandals are already at the stage where they know that vandalism on Wikidata leads to large-scale vandalism on the projects (and is actually way more efficient), and, unless we want that all projects block Wikidata transclusions as the English Wikipedia has already (almost) done, we need to think about some good strategy.--Ymblanter (talk) 06:38, 8 July 2020 (UTC)

Would probably be easier if most of the recent changes weren't in spanish, arabic or russian. --Trade (talk) 07:23, 8 July 2020 (UTC)
@Trade: There are tools to help with that. SpeedPatrolling optionally only shows you unpatrolled changes written in certain scripts (e. g. Latin); Wikidata vandalism dashboard shows label/description/alias/sitelink changes in certain languages; reCh likewise lets you filter by language code(s) once you select terms edits. --Lucas Werkmeister (talk) 08:57, 8 July 2020 (UTC)
@Trade: What exactly is difficult about edits being in Spanish, Arabic, or Russian? They are not secret codes! - Jmabel (talk) 13:33, 8 July 2020 (UTC)
phab:T189412 (if ever implemented) would enable admins to do massive semi-protection Ghuron (talk) 12:24, 11 July 2020 (UTC)
I do not think it would ever be implemented. Nobody is working on it, and we have no means to force anybody to work on it.--Ymblanter (talk) 19:11, 11 July 2020 (UTC)
I'm trying to remind Lydia about it every office hour. May be at some point of time she'll decide to implement that just to shut me up :) Ghuron (talk) 07:17, 12 July 2020 (UTC)

Simple query?

I need some help creating a query. what I really want is a list of the alma mater for winners of the Honda sports award (Q5892712) when the article does not have a photo. That last aspect might be tricky (although maybe not because one could check to see if there is an image), but I already have a list of winners for which there is no photo, so if I have a list of all winners and their alma mater, I can merge that with my list and identify those without photos.

I see that as a field for educated at but it can include both high school and college. I'm only interested in college but if I get both I think I can figure out how to throw away the high school.

There are a number of Honda sports awards such as on the sports award for volleyball, on the sports award for basketball but all of them should be part of Q5892712.

My goal is to generate a list that I can dump into Excel for other purposes.

It may be helpful to point out that there are 487 award winners, 314 of which have Wikipedia articles. 144 these articles do have a photo, and 170 do not. --Sphilbrick (talk) 00:28, 12 July 2020 (UTC)

@Sphilbrick: I replied at Wikidata:Request a query#Honda_Sports_Award_alma_mater Vahurzpu (talk) 04:56, 12 July 2020 (UTC)
Thanks, very helpful. I have to run out for a few hours, will look closer this afternoon, but I think this does what I want.--Sphilbrick (talk) 12:59, 12 July 2020 (UTC)

Difference between Q16764963 and Q170285

The title of Q16764963 in Luxembourgish is Verkéiershiwäisschëld and in German is Verkehrshinweisschild. Everywhere I tried to find a translation of it seemed to come up as just traffic sign which is what Q170285 is. I don't speak Luxembourgish or German, so I don't know if there's some difference that doesn't translate or if they are the exact same. --TB5ivVaO1y55FkAogw1X (talk) 00:45, 12 July 2020 (UTC)

There is a list of different types of traffic signs as defined in Luxembourgish law here. There is a similar list with English terms at https://en.wikipedia.org/wiki/Traffic_sign#Categories. As far as I can tell, Q16764963 would be the union of directional road sign (Q2554368) and information sign (Q6031215) but excluding non-traffic signs. That makes these concepts somewhat different, leading to individual items. There are many such cases, such as school teacher in the Swedish school system (Q10497467). This query shows all professions that are somehow country-specific, many of which have near-equivalents in other countries. It's somewhat philosophical to wonder if a high school teacher (Q10511368) is the same as high school teacher (Q5758653), which is the Spanish version of it: the requirements, ages of students, stratum of society, etc. will differ between countries. But for purposes of querying and analysing the data, it's often annoying. And when neither a property nor your language's description mention the country-specificity, it's confusing. Matthias Winkelmann (talk) 03:15, 12 July 2020 (UTC)
We've got more or less the same problem in several different areas, and there may be common solutions.
  • In this previous thread [somebody please fix the link when it's archived] we're discussing the issue of distinguishing between the entity that describes an island, versus the entity that describes the country or municipality that's coterminous with the island, versus the "disambiguating item" that is either the island or the municipality, and exists solely to be tied to a Wikipedia article that doesn't distinguish between the island and the municipality.
  • Similarly, calorie (Q87260855) is an entity that describes one of two (actually several) different units, the small calorie (Q130964) and the kilocalorie (Q26708069). (There are potentially lots more of these, as there are multiple definitions of units like league (Q660818), ounce (Q48013), and bushel (Q216658).)
My point is that in the case of occupations, we could stand to have an entity for, say, "generic schoolteacher", that then linked to the various country- and/or language-specific variants. And similarly for Q16764963 and Q170285.
(I wonder if we need a new name for this situation? It's like the Bonnie and Clyde problem, but more generic.) —Scs (talk) 12:06, 12 July 2020 (UTC)

nv.wikipedia Template:@

Which item this is equal to? Template:@ (Q6133158)? Template:No spam (Q6021142)? Or neither? (I asked @Seb az86556: which is the only known Navajo native speaker, and the administrator of nv.wikipedia, but still no replies). --Liuxinyu970226 (talk) 12:12, 5 July 2020 (UTC)

It is used exclusively for files, apparently, to show where the files came from.--Snaevar (talk) 18:17, 7 July 2020 (UTC)
@Snaevar: So it's Template:Information (Q6466010)? --Liuxinyu970226 (talk) 23:46, 7 July 2020 (UTC)
No, it is a lot simpler. It only serves the purpose I mentioned, nothing more.--Snaevar (talk) 22:14, 8 July 2020 (UTC)

The example of usage is here. Hard to say with which item it's better to link but surely not with Template:File source (Q5617747) or any supposed above. --Wolverène (talk) 11:45, 13 July 2020 (UTC)

(upd) Most probably a primitive variant of Template:Imbox (Q5825560), so it may be merged into Q5825560 and seems it wouldn't be a great mistake anyway. --Wolverène (talk) 11:59, 13 July 2020 (UTC)

Removed a bogus image, did I folllow the intended process?

I just removed a bogus image from a WD entity. I asked about it in the discussion page of the entity, got no reply for a couple of days and proceeded to remove it. I am not sure if this is how it's done here, please advise if not. Moreover, I suspect the image came from some automated (and slightly buggy) extraction process and is likely to reappear. Is that so? For reference, the resource discussion: https://www.wikidata.org/wiki/Talk:Q638608#%7B%7BP%7C18%7D%7D Fkleedorfer (talk) 08:42, 13 July 2020 (UTC)

On WikiData many items were filled automagically. Sometimes the scripts were buggy, but also often the underlaying data (especially categories) were bogus. 100 edits would mean 1-2 errors, leaving 98 edits that were acceptable, a decent way to get WikiData filled quickly. Talk pages are often not watched actively, people that create items, usually create hundreds of them a day, with tools like PetScan or QuickStatements. If you see an error, just fix it, and hunt the next one! We have 80 million items, most of them are filled pretty much correct, but most of them also need some love and improvement. Edoderoo (talk) 12:14, 13 July 2020 (UTC)

Mapping Keyphrases to Wikidata Items

Has anyone done any work with mapping a keyphrase, like "Google" or "International Business Machines" or "IBM", to the most relevant wikidata items? I've started looking into it and don't want to reinvent the wheel. Entity resolution looks like a key service required to better integrate wikidata with other information on the internet.  – The preceding unsigned comment was added by Tbonza (talk • contribs) at 13:22, July 13, 2020‎ (UTC).

@Tbonza: There are several solutions for this; a lot of people have found OpenRefine (Q5583871) to work very well in reconciling a list of names to Wikidata items. ArthurPSmith (talk) 14:30, 13 July 2020 (UTC)

Adding a reference seems to be broken

Adding a reference seems to be broken "Reference URL" sits in a text field with no way to put in a value. The marker shows 1 reference while I am trying to add and then goes back to 0. Example Q64185991 trying to add a reference URL for https://books.google.com/books?id=6PtfAAAAMAAJ&pg=PA195 to date of birth Mary Mark Ockerbloom (talk) 13:48, 13 July 2020 (UTC)

Rosetta Stone for Class and Property Descriptions from External Ontologies?

I am working on a Canadian knowledge graph of performing arts (http://artsdata.ca). Our data model uses classes and properties from existing ontologies such as schema, foaf, FRBRoo, rda, ... and has mappings to Wikidata's ontology. However, I am running into a problem documenting our data model as I need both english and french descriptions of the classes and properties we are using from the existing ontologies. I would like to avoid creating my own translations.

PROBLEM

The general problem is that many ontologies don’t have their class and property definitions documented in the desired language of the reader. In my particular case I am looking for a french translation of several Schema.org classes and properties. This creates a barrier for reusing existing ontologies. Ideally the ontology itself would offer translations, and many do. However some ontologies, like schema.org, are only documented in English.

In my case, schema:Event is only documented in English on http://schema.org. Wikidata has an equivalent class (wdt:P1709) for schema:Event which is wd:Q1656682. However, the human description of wd:Q1656682 is not the same as the definition on schema.org. And in many other cases there may not be an equivalent class or property.

SOLUTIONS

The idea I am proposing is that Wikidata offer a crowd sourced “rosetta stone" for ontology class and property descriptions. For example, a new Wikidata entity could be created for the purpose of documenting schema:Event and the rdfs:comment made available in @fr and other languages by the community. However, the original description from the ontology would never be changed. The original texts of the ontology (in all languages the ontology supports) would need to be identified as read only.

Since this is a specific use of Wikidata for the purpose of translating documentation of external ontologies, I would like to know what the Wikidata community thinks, and learn about alternative approaches.

Thanks in advance. Saumier (talk) 13:13, 9 July 2020 (UTC)

When it comes to the actual content it seems to me that you confuse descriptions with definitions. The two are different things. If we take the example of Schema.org, they define the license of their content as CC-by-SA-3.0. While we consider the identifiers to be more free then that, the definitions are free text and might de-jure be copyright protected in a way that prevents us from simply copying their definitions. Given that copyright is stronger for free text then data we limit the amount of properties that store freetext on Wikidata. It seems to me that the project you are proposing would work better in a Wikibase instance outside of Wikidata that's itself CC-by-SA. ChristianKl ❪✉❫ 17:30, 10 July 2020 (UTC)

How to add a license number?

Hello,

I am pretty new to Wikidata so please excuse me if the answer is obvious to you guys.

I can see there are properties for some very specific licences (such as for gymnasts) but the generic "licence" begins with Q rather than P. So how can I add the operating licence number (which also has other characters than digits) to Kangal power station and hopefully other Turkish coal-fired power stations please?  – The preceding unsigned comment was added by Chidgk1 (talk • contribs) at 12:56, 12 July 2020‎ (UTC).

  • I'm not sure if we have an appropriate property, but that sounds like it's basically an identifier, so if you go to Wikidata property for an identifier (Q19847637) and ask to see the derived statements that will give a pretty good list. I'm a little surprised, though, that it's all instance of (P31), no subclass of (P279). - Jmabel (talk) 16:54, 12 July 2020 (UTC)
    • Thanks for your prompt reply. Yes there is an external database here which can be queried manually by the license number. However occasionally 2 power plants are authorised by the same licence, so is it in fact an identifier? I do not understand what you are saying about instances, subclasses and how to ask to see derived statements. Could you explain in terms suitable for someone new to Wikidata like me? Should I request that a new property is created specifically for operating licence numbers of Turkish power plants? I may well have made mistakes in editing Kangal power station. If you spot any or know an excellent example of a coal-fired power station please let me know. Chidgk1 (talk) 18:43, 12 July 2020 (UTC)

Page says nothing to translate

I know little about Wikidata but I've been translating into Dutch for a while now. Looking to do some more I came across https://www.wikidata.org/wiki/Wikidata:Wikispecies/nl. In small letters right below the title of the page it says translation is 100% complete. But if you look at the page, 90% is still in English. I clicked on the translate button, and then "untranslated", which tells me "nothing to translate"! What am I missing here? --Dutchy45 (talk) 15:18, 13 July 2020 (UTC)

  Notified translation administrators --SilentSpike (talk) 16:30, 13 July 2020 (UTC)

The page Wikidata:Wikispecies consists mainly of templates. To translate the text, go to the template pages, e.g. Template:Sister projects overview/part 5, and click there on the translate button. --Pasleim (talk) 21:26, 13 July 2020 (UTC)

Wikidata weekly summary #424

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Events
    • Upcoming: Next Linked Data for Libraries LD4 Wikidata Affinity Group call: New co-facilitator introductions and tools for editathons, 14 July. Agenda
    • Upcoming video: Wikipedia Weekly Network - LIVE Wikidata editing #12, 18 July Facebook, YouTube
    • Past: Celtic Knot Wikimedia language conference, online, July 9-10. Videos, documentation and slides of sessions are accessible from the program page.
  • Tool of the week
    • DragNDrop.js makes it possible to show a Wikipedia article and drag items to Wikidata.
  • Development
    • Continued building out documentation for Federated Properties (phabricator:T255651) and making interface improvements to the first stage of the feature (incl. phabricator:T246886, changes to special pages that interact with both Items/Properties, and phabricator:T255581, changes to Special:ListDataTypes when federation is enabled)
    • More work on the consistent design system
    • More work on decoupling the different Wikibase extensions from each other to make development easier
    • Finalizing research and interviews to better understand what could be improved in the way developers access Wikidata's data (APIs, SPARQL)
    • Testing the first prototype of the Simple Query Builder with some editors to get final input before coding starts
    • Sorting of language links on Wikipedia and the other Wikimedia projects was broken (presumably by a change in MediaWiki core). A fix is being worked on. (phabricator:T257625)

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.

successiveness

Hello. Mayor of Lysi (Q56459600). The position founded in 1962. After 1-2 years the position dissolved. The only mayor was Yiangos Kalli Souroullas (Q97321428). Then the (same) position was founded again in 1986. The first mayor after the re-creation was Andreas Tofia (Q94691858). Can we add that:

and

?

The sources are not helping me to decide. One source says that Mayor of Lysi (Q56459600) was the first mayor, maybe that helps a little. (Sorry for my English if something is not clear). Data Gamer play 15:50, 13 July 2020 (UTC)

@Data Gamer: There are a couple of different approaches usually used here. If we actually want to record that the position was dissolved in the early 1960s and then brought back in the 1980s, the usually way is have two separate items for the position. They can have the same name, but often in cases like this there are other differences we'd want to model too (e.g. the law the created the position). The other is to say that the position wasn't really abolished during that period, but simply vacant. In that case it's more usual to say that the later officeholder replaces (P1365) the older one, but with a twenty year gap, that seems a little odd, and even with a single office it may be more reasonable to use novalue inbetween. Personally I'd lean more to creating a new item for each period of the post with a gap like that. --Oravrattas (talk) 19:10, 13 July 2020 (UTC)
For the sources it is only one position... Thanks for the answer. Data Gamer play 20:23, 13 July 2020 (UTC)
Mayor of Wellington (Q52988199) has the same issue with a long break between William Guyton (Q8010296) and Joseph Dransfield (Q6282698). At one point I tried linking them with followed by (P156) but a bot kept replacing it, so I gave up: the link isn't really needed anyway, since a list can be easily constructed by sorting by date. Ghouston (talk) 01:31, 14 July 2020 (UTC)
Thanks. Data Gamer play 15:20, 14 July 2020 (UTC)

Deletion log

I remember that I have created a "Gino Sorbillo" item, now deleted. I tried to search on deletion log but no result back. How can I find deleted item and reason? --2001:B07:6442:8903:CC47:786A:4FF5:973D 14:57, 14 July 2020 (UTC)

Q93450079 (Gino Sorbillo, "Italian pizza chef") —MisterSynergy (talk) 15:07, 14 July 2020 (UTC)

Litoria gularis double entry

I see that there are two entries for Litoria gularis, a species of frog: Q3019219 and Q22112053. I believe they should probably be combined.

Wikispecies also seems to have this twice: Wikispecies as Nyctimystes and Wikispecies as Litoria, but that is a separate issue. Darkfrog24 (talk) 18:24, 11 July 2020 (UTC)

This is Wikidata, where nothing is guaranteed to make sense or be convenient for humans. Items for taxa and scientific names tend to be difficult to keep internally consistent and reconcile across other projects, but generally every name combination gets its own item, whether it is valid or not. Thus Litoria gularis (Parker, 1936) is separate from Nyctimystes gularis Parker, 1936 (and its parent taxon is Litoria, not Nyctimystes). Synonyms often have separate external identifiers, e.g. GBIF for L. gularis and N. gularis. The labels should be altered to reflect the right combination, and external identifiers may need to be shuffled. The accepted name(s) and basionym can be linked with taxon synonym (P1420). -Animalparty (talk) 20:19, 11 July 2020 (UTC)
Added the missing relationship. --Succu (talk) 20:46, 11 July 2020 (UTC)
I think this could gain in clarity:
@Succu: Could this be done by bot?
  • No, the alternitive autor citation is incomplete. Read Litoria gularis (Parker, 1936) Frost, Grant, Faivovich, Bain, Haas, Haddad, de Sá, Channing, Wilkinson, Donnellan, Raxworthy, Campbell, Blotto, Moler, Drewes, Nussbaum, Lynch, Green & Wheeler, 2006 [= Litoria gularis (Parker, 1936)].
  • I don't think there is consensus to change all the labels that way. Sometimes I do this manually. Especially if the addition of a label fails or I moved sitelinks to another item.
  • If you want to add original combination (P1403) you need a good source. I added basionym (P566) to a lot of items because I had some (witch had errors too), but still a lot of them are missing. --Succu (talk) 14:52, 12 July 2020 (UTC)
NB: People often mistake "instances of a name" for taxonomic synonyms. This is false. Some databases (for some reason especially herpetological ones) really like to record every instance a name was used, and then append the author of the work it appears in followed by a dash. This is not equivalent to a new name, and in this case Frost et al. are not authors, merely listers. If I published an article or field guide next week using the name "Litoria gularis, then some fastidious database may eventually include "Litoria gularis (Parker, 1936) - Animalparty 2020". This has no bearing on the validity of any name, it's just record keeping. -Animalparty (talk) 04:45, 13 July 2020 (UTC)
  • (1) Is it worth sorting out the sitelinks and fixing the labels? (2) Would there be an alternate way to link the two items? --- Jura 09:17, 13 July 2020 (UTC)

  WikiProject Taxonomy has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. --- Jura 07:29, 15 July 2020 (UTC)

Authority control proposals not transcluding properly?

At the bottom of the list of Wikidata:Property_proposal/Authority_control, I notice that there about five items that are being linked (and not included in the page index) instead of being transcluded into the page:

  • Wikidata:Property proposal/National Registry of Exonerations Case ID
  • Wikidata:Property proposal/QLD Biota ID
  • Wikidata:Property proposal/Australian Weed ID
  • Wikidata:Property proposal/PM20 geo code
  • Wikidata:Property proposal/PM20 subject code

I had figured that there was a typo that broke the transclusion, but it all looks correct on the code level to my admittedly-inexperienced eye. Anyone know what's going on with that? Thanks Kenirwin (talk) 18:24, 13 July 2020 (UTC)

The problem is there are too many proposals and the technical limits for the page have been reached. Our property creation system is not only technically inadequate, but also quite ineffective at filtering out unsuitable or redundant properties (some of which end up being deleted a few months later because they were created based on votes, sometimes just one, and not on the systematic fulfillment of clear criteria), and inefficient for the users involved. I'd say it's a process that delays the inclusion of certain data and consumes a lot of user time without all this serving to guarantee the suitability of the properties created. If anyone has ideas on how to improve the system, both its requirements and its implementation, they're likely to have my support. I'd make some suggestions to improve the decision-making processes, but those suggestions wouldn't be very useful if they can't be implemented due to technical (or, in general, resource) constraints. --abián 19:29, 13 July 2020 (UTC)
More participants in discussions would be a good thing. In the short term, as fix would be to split "Wikidata:Property proposal/Authority control" into, say, "Wikidata:Property proposal/Authority control for people" and "Wikidata:Property proposal/Authority control (other)". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:29, 13 July 2020 (UTC)
You should start by making a clear and coherent definition of what does and what does not count as authority control. @Pigsonthewing:--Trade (talk) 07:22, 14 July 2020 (UTC)
That sounds sensible to me Kenirwin (talk) 01:05, 14 July 2020 (UTC)

  Notified property creators --SilentSpike (talk) 05:28, 14 July 2020 (UTC)

  • I'd be okay with splitting the subpage up. I also think we could codify when to mark proposals as not done to get rid of some of the ones clogging things up. --SilentSpike (talk) 05:30, 14 July 2020 (UTC)
It's pretty mich a gamble as to when a proposal can get marked. As of right now it's somewhere between two weeks and two months.--Trade (talk) 07:24, 14 July 2020 (UTC)
  • Creations and (automated) archiving seem to have taken care of this. I closed a few stale ones as   Not done. --- Jura 10:10, 15 July 2020 (UTC)

Property misplaced as item

Hi! I think that item Q65972149 (founder of), should be listed as a property rather than an item. Property P112 (founded by) is listed as its inverse, which does not make sense if 'founder of' is an item. I don't know how to fix this or where the appropriate place to discuss this matter is, so I figured I'd place it here. Thanks! Jlevi (talk) 23:36, 14 July 2020 (UTC)

  • @Jlevi: So this is an inverse label item. These exist because not every property should have an inverse property. A given item would typically have one or few "Founded by" statements, whereas a given person/entity could theoretically have unlimited "founder of" statements. There is no real value in creating both as properties for a two way link, since the data can be queried in either direction with a one way link. You can see these inverse labels in action if you enable the relatedItems gadget in your preferences then scroll to the bottom of an item page and use the "Show derived statements" button.

    The one thing I would say is, I think Wikidata property (P1687) is frequently informally being used incorrectly on these inverse label items. --SilentSpike (talk) 09:50, 15 July 2020 (UTC)

Question about conflation in WorldCat identities

I've been side-tracked into trying to clean up some authority control links and have been discovering what many of you probably already know, that it's practically the Herculean labor of the Augean stables in the case of a common name, what with bots merrily linking to a slight resemblance or to a mistake that another bot already made. But it's usually possible either to move an authority to its correct item or to create a conflation item for it. WorldCat Identities puzzle me, though. They are usually linked to a Library of Congress authority record, but they then go on to list books supposedly by that author, which in most cases aren't, but are rather by multiple different authors. The LOC record doesn't name those books, so I don't know where WorldCat is getting its authorship information. Should I treat such WorldCat records as conflations? Levana Taylor (talk) 03:46, 3 July 2020 (UTC)

  • I've asked some folks from OCLC to help us with this question, but it's a long holiday weekend in the US so I wouldn't expect an answer for a few days. Gamaliel (talk) 13:24, 3 July 2020 (UTC)
    • From what I've found over time, I think that the system works as follows: from the identifier it gets a heading (if it's an LCCN then there's a unique associated heading, like "Baker, James, lieutenant-colonel" in the example above) and it then looks for that exact string in the "author" field of whatever library records it has in the database. Therefore, I don't think it would be possible to disambiguate manually. Take a look at this book in one local library's catalog. If you go into the MARC record (link at bottom of left column), it only has the author's name as "Berger, Warren"; nowhere does it have the LoC authority number "n96041686". If some other Warren Berger popped up, the record would become ambiguous unless updated. Given that WorldCat has thousands of member libraries, I don't know if "get every library to recatalog ambiguous works" would be very effective. Vahurzpu (talk) 14:28, 3 July 2020 (UTC)
      • I'm not sure that can be right--there's not enough randomness. In my second example, although the works span many authors and many topics, from airport security screening to tantalum ore to the sanitation of New Guinea, they are all either prepared by or for legislatures and government agencies, or else deal with technicalities of government (International Law and the Arctic for example has no government affiliation in its preparation or publication). It's not all the same government, so that can't be what's causing them to be grouped together. I wonder if the order of events from authorship to grouping wasn't the other way around: first have a machine create a group based on a cloud of similarities like government affiliation or government policy subject tags, then attach that group to the identity of the author of the most-commonly-shelved work in the group. I've checked a few more identities, and in all of them, the first work on the list has its author correctly identified, even if misfits get into the list farther down. Levana Taylor (talk) 16:16, 3 July 2020 (UTC)
  • The way I see it, It's not the role of Wikidata to be a scorekeeper of every inaccuracy on the internet. All databases have some amount of error (Wikidata, Worldcat, VIAF, Internet Movie Database (Q37312), Internet Broadway Database (Q31964) etc.), but I see little sense in modifying external identifiers with qualifiers. Yes, some bibliographic databases include works from the wrong author, and some performance databases conflate actors. We could tag every slightly inaccurate value with all the possible or potential conflated IDs, but it'd be moot once the source is updated, rendering Wikidata needlessly complicated (and also then incorrect in itself) and who's going to come back and fix all the Wikidata items that are now no longer conflations? I say do the best with what we've got, e-mail the database stewards if you feel up to it, but let's not bend over backwards to note that a particular error occurred at one time. -Animalparty (talk) 03:36, 4 July 2020 (UTC)
That way of looking at things certainly simplifies matters enormously. Just figure out what author the WorldCat identity was supposed to represent (easy in the case of their "lccn-" and "viaf-" numbers, and in the case of their "np-" numbers, I think it'll be the first one on the list), and that is that. It's useful to draw a distinction between conflation and other sorts of error: if lccn-n2013070019, for James Baker (Arctic scholar), contains one work by him and nineteen by others, it's not because he's been confused with those others, but because the automated process did a bad job of matching him with works in the catalog. Levana Taylor (talk) 14:35, 4 July 2020 (UTC)
I agree, however it is still a good idea to record that information to indicate that you are aware of conflation and either report this to the database in question or at least indicate in Wikidata. This would prevent other editors from adding the wrong information again or trying to "solve" the issue. There is currently a discussion ongoing on how to best indicate this: Wikidata:Property proposal/external error. This proposal would allow you to be specific about the type of error detected. --Hannes Röst (talk) 03:48, 5 July 2020 (UTC)
  • Given that it merely relys on two other identifiers, it seems that P7859 is a problematic and may not be suitable for Wikidata. --- Jura 07:14, 15 July 2020 (UTC)

Location of languages

I am informed that 6,405 items about languages (or subclass of) have coordinates; for example German is said to be located at 52°0'0.000"N, 10°0'0.000"E ("±0.000001°"). I've never been to that point, I admit (it's a field south of Hildesheim, in Lower Saxony - it's quite possible nobody has spoken any language there for weeks), but I have heard German used elsewhere - from Scotland to Australia, and from Qatar to the United States. The situation is ridiculous, and no doubt for some, insulting (are we going to tell the majority of Kosovans that their language belongs not there, but in Albania?). Can we reach consensus that "coordinate location" should not be used for languages? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:23, 2 July 2020 (UTC)

Please notify the WikiProject Languages. Visite fortuitement prolongée (talk) 18:34, 3 July 2020 (UTC)

@Visite fortuitement prolongée: I pinged the Linguistics project but it's not possible to ping the project Languages, I'll leave a message there. Cheers, VIGNERON (talk) 16:47, 4 July 2020 (UTC)
Thank you. Visite fortuitement prolongée (talk) 17:39, 4 July 2020 (UTC)

Identifying important items

I'd really like to be able to identify the most "important" items that have "issues". There's tons of items with problems but the vast majority of them are not that important. There's lots of ways to assess item importance but it would be helpful if that information were available within wikidata so I can author SPARQL queries around it (e.g. "show me the most important humans that don't have a birth date property"). This could also help with drop-down suggestions. So what would people think about doing something like:

  • Computing (say) the PageRank (Q184316) of every item (maybe across the major wikis and wikidata)
  • Creating a new property to hold this value (along with a timestamp)
  • Importing the importance values for the (say) top 5M most important items.

It seems this has been discussed before but went nowhere. And it's possible that Wikimedia already has some relevant importance metrics we could just import e.g. traffic stats.

I could do this entirely outside of wikidata but it would mean that only a few people would be able to leverage the information. I understand this is potentially an abuse of properties since this is information about the item not what the item represents (though that's blurry).

BrokenSegue (talk) 06:59, 9 July 2020 (UTC)

If you are looking to complete items, I don't think you can get around picking a field and then maybe filter items that have sitelinks to languages that use Wikidata (e.g. ca/cs/da/es/no/ru/sv/commons, etc.) --- Jura 08:52, 9 July 2020 (UTC)
  • Identifying and ranking "important" items is a really interesting, really important, and afaik really hard problem. It seems highly subjective, but of course we wouldn't want to make it a value people could just enter, because that'd be Original Research, and people would edit-war over the importance of their favorite items forever. So in practice we need some kind of fair, objective algorithm -- but it's a real research project to find and tune those.
For entities like islands (Q23442) and human settlements (Q486972), I've wondered if it would be possible to define a formula, based on statistics like area and population, with perhaps a bump for capitals (Q5119). —Scs (talk) 16:54, 11 July 2020 (UTC)

Birth name, full name, personal name, legal name, and so on...

Hi, everyone!  
I'm trying to understand how to use "personal name", "maiden name", "legal name". What property should I connect them to?
And the full names... I mean, in the property "birth name" we write the full name of the person when he/she was born. If someone was born named "John William Charles Smith", and then changed his name to "John William Evans". Is there a property where I add the person's new full name? Like official name (P1448) or any onter? If yes, how? I'm trying to understand, but I'm a little confused.   Minerva97 (talk) 19:18, 10 July 2020 (UTC)

official name (P1448) excludes human (Q5). name in native language (P1559) is better here. If the change is due to marriage, birth name (P1477) and married name (P2562) are more specific. Matthias Winkelmann (talk) 03:33, 12 July 2020 (UTC)
@ChristianKl:, @Matthias Winkelmann: I was thinking about cases like Miley Cyrus (Q4235). She was born Destiny Hope Cyrus, in 2008, she legally changed her name to Miley Ray Cyrus. "name in native language (P1559)" applies in this case? Minerva97 (talk) 16:14, 15 July 2020 (UTC)

Specifying countries represented among students

For Pomona College (Q7227384), I recently wrote a Simple Wikipedia page, and I'm trying to transfer some updating statistics to Wikidata so that I'll hopefully never have to update the Simple Wikipedia page again. Some of those are stretching a few limits, though. For the sentence About 1,700 students representing all 50 U.S. states and 63 countries attend as of 2020., representing the student count is pretty straightforward using students count (P2196), and I got the "as of" taken care of as well, but would there be any way to represent the 50 and the 63 in Wikidata? {{u|Sdkb}} talk 20:32, 14 July 2020 (UTC)

I have a similar issue with representing the number of majors offered and the number of courses offered. Should properties be created for those? {{u|Sdkb}} talk 22:19, 14 July 2020 (UTC)
@Sdkb: Creating property proposals for these sounds reasonable to me. ArthurPSmith (talk) 17:13, 15 July 2020 (UTC)

BREAKING CHANGE: removing special pageterms behavior on repo wikis, use entityterms instead

This is an announcement for a breaking change to the pageterms submodule of the query API module, which only affects Wikibase repository wikis. If you do not use that API module, or only use it on client wikis (e. g. Wikipedias) and not on repository wikis (Wikidata, Wikimedia Commons), you can ignore this message.

For years, the pageterms API module has served a double role: on client wikis, it returned the “terms” (Label, Description, Aliases) of the Wikidata Item linked to the given page(s), whereas on repo wikis, it would return the terms of the Item (or other Entity) on that page itself. For example, querying for the Label of Wikipedia:Village pump on English Wikipedia would return “Project:Village pump” (the Label of Q16503), but querying for the Label of Wikidata:Project chat on Wikidata would not return anything, even though that page is linked to the same Item – you would have to query for the Label of Q16503 instead. This behavior is inconsistent and also mixes repo and client concerns in a way that makes the Wikibase code harder to maintain.

To resolve this, we introduced a new entityterms API module (a submodule of the query module, just like the pageterms module) which has the same behavior as the pageterms module currently has for Item (or other Entity) pages, and which is only available on repo wikis. If you want to get the terms of Q16503, you can now use action=query&prop=entityterms&titles=Q16503 instead of action=query&prop=pageterms&titles=Q16503. (You can also use wbgetentities, which gives you much more control over the returned data; pageterms/entityterms may be faster and can also be combined with other submodules of the query module.) On or shortly after , we will remove the special repo behavior of the pageterms module, and it will then behave just like it always has on client wikis, and return the terms of the Item linked to a page, not the terms of the Item (or other Entity) on a page. (Because the new API module is already available on Wikidata, and you can start using it immediately, we are not making this pageterms behavior change available on Test Wikidata significantly before that date.)

If you have any issue or question, feel free to leave a comment at T257658. For more information, see also T115117, T255882 and T256255.

Cheers, --Lucas Werkmeister (WMDE) (talk) 16:06, 15 July 2020 (UTC)

Wikidata:Requests for comment/General semi protection for all property pages

Following from the previous discussion here, an RfC has now been started on general semi protection for all property pages - please comment there. Thanks. Mike Peel (talk) 20:07, 15 July 2020 (UTC)

Thanks to you and MS for putting this together - really glad to see it as a formal RFC. Andrew Gray (talk) 20:30, 15 July 2020 (UTC)

QuickStatements and Wikidata watchlist

When you create new WD items with QuickStatements these new items are not starred/put on your watchlist. Is it possible to somehow put all the Q-ids you created on your own watchlist? Ecritures (talk) 20:14, 15 July 2020 (UTC)

Users with fewer than 400k edits can download their created items with this tool and add them to the watchlist via Special:EditWatchlist/raw. —MisterSynergy (talk) 20:26, 15 July 2020 (UTC)
Thank you: that worked perfectly :) Ecritures (talk) 20:41, 15 July 2020 (UTC)

Help for Arabic typing

Hello to all! Could someone help me for Arabic typing? In Moroccan Arabic Wikipedia (Q97393767) there is svg version of Moroccan Darija (Q56426) Wikipedia logo and I wish to typing the words for adding the logo in en:Wikipedia:Slogans. The words I have typed by me are:

Wikipedia = ويكيپيديا

The Free Encyclopedia = لموسوعة الحورا

I have typed correctly?

Thanks in advance!!!

--2001:B07:6442:8903:649F:94AE:92A:F6BF 09:24, 20 July 2020 (UTC)

What does "semi-protected" mean here?

I've done a few dozen edits on Wikipedia, but I'm pretty new to Wikidata. I noticed that the Pittsburgh article on wikipedia has a bad link for the official pittsburgh webpage in the infobox, and realized I need to fix that in Wikidata. So I need to edit Pittsburgh to change https://www.pittsburghpa.gov/ to https://pittsburghpa.gov/ (remove the www). But I can't, because it's "semi-protected"? Am I correct that that is the reason why? Is there documentation about what "semi-protected" means in Wikidata ? It doesn't appear to be in the FAQ at Help:FAQ and I can't figure out search terms that get me the answer quickly.

Bsammon (talk) 01:56, 14 July 2020 (UTC)

Yes, you cannot edit that item due to the "semi protection". It is a page protection level that excludes IP editors and "newcomers" (fewer than 50 Wikidata edits or younger than 4 days I think) from editing. Once you reach 50 edits, you are automatically being promoted to the "autoconfirmed" user group which allows you to edit semi-protected pages as well.
The current protection of the item Pittsburgh (Q1342) was applied by User:Abián a while ago who also set up the Wikidata RfC to exclude IP editors and newcomers from editing "highly visible items" via semi-protection. That RfC was unfortunately accepted by the community and found its way to the Wikidata:Page protection policy, although our admins still do not have any means to verify whether an item applies for this sort of protection at all. —MisterSynergy (talk) 07:10, 14 July 2020 (UTC)
So, followup questions:
  • Should I add this to the FAQ?
  • Is there a "Why can't I edit this protected page and what do I do now?" help document I can refer to (either for my own bookmarks or for the FAQ)
  • (How) can I find out why this (or some other) specific page was protected?

Bsammon (talk) 17:38, 14 July 2020 (UTC)

Thanks to the Wikimedia Deutschland team, we (everyone) do have access to updated lists of Wikidata entities along with their usage numbers (see [12] and [13]). However, it's not convenient to paste these links very often as they provide the information necessary for vandalism to be particularly damaging on many pages, projects and external services.
@Bsammon: Please feel free to request on the talk page of the corresponding entity any changes you want to apply next time (if any) and add {{edit request}}. You will have to perform this step a very limited number of times until you have made enough edits or, instead, until your edits show an outstanding knowledge of the project, in which case we'll be able to assign you a confirmed flag in advance. Please ask us if you have any questions.
This process for requesting changes (for whatever reason) isn't ideal, but I think it will improve soon. The development plan for 2020 considers the task phab:T234976, on feedback loops, which includes phab:T229100, although the plan includes many other tasks and I don't know to what extent it's been possible to make advancements this year. @Lea Lacroix (WMDE): What do you have in mind? --abián 16:38, 15 July 2020 (UTC)
Ah shoot. I asked our designers to make some potential mock-ups but then dropped the ball in the next step to take it to the developers -.- Putting it back on the list now. Next step is to figure out how to make this possible technically because that's apparently a bit of a pain in that section of the page. Thanks for the reminder about it. --Lydia Pintscher (WMDE) (talk) 19:09, 16 July 2020 (UTC)

change property Boijmans identifier for artist to external identifier

Can someone change back the property Boijmans artist ID (P3888) to an external link? The property was decoupled as external identifier as the link didn't work for a while, but it is working again. See vor example https://www.boijmans.nl/collectie/kunstenaars/3283 --Hannolans (talk) 15:31, 14 July 2020 (UTC)

@Hannolans: It's still set as an external-ID, so I think all you need to do is change the formatter URL to be "preferred" or "normal" rather than "deprecated". (I would do it but I think only one should be "preferred", and there are two of them - I guess Dutch?)
The problem is that this will not automatically update the pages. Changes to a formatter URL need the pages to be purged (eg by an edit) to update the link on the item - see phab:T112081. I can run a script to purge all the pages though, once the formatter is fixed. Andrew Gray (talk) 15:58, 14 July 2020 (UTC)
Ah thanks! I have set the english language link as preferred. But can't change it to a linkable identifier also not by editing. If you can purge the pages that would be great. --Hannolans (talk) 16:14, 14 July 2020 (UTC)
Okay! This seemed to run a bit weirdly at first but I have finally got a page to purge and display the link - Q321245#P3888. Script is running overnight so let's see how it looks in the morning with a sample of items. Andrew Gray (talk) 22:32, 14 July 2020 (UTC)
Script is still churning slowly (the script slows down when things get lagged) but it's starting to show effects - clickable links are appearing! Andrew Gray (talk) 09:40, 15 July 2020 (UTC)

Wrong merge of scientists - how to best undo the substitution on articles

Hello, I detected a wrong merge of two scientists (Alexander D. Popov (Q67220493) and Aleksey Popov (Q67378272)) performed in February 2020. Is there an automatic way to undo the substitution on articles? - Valentina.Anitnelav (talk) 09:06, 15 July 2020 (UTC)

Looking at some of the articles linking to Q67220493, I suppose I have to agree with MisterSynergy --- Jura 10:29, 15 July 2020 (UTC)
Edits like "Author Disambiguator matching authors for" are hard to decode. --- Jura 10:37, 15 July 2020 (UTC)
  1. Bisect the offset parameter in the URL. For example https://editgroups.toolforge.org/?limit=50&offset=5000&tags=prop-P50&user=KrBot
  2. Name of every edit group have a pattern so that you can find a specific edit group such as [14]. In this case no items is affected.

--GZWDer (talk) 10:41, 15 July 2020 (UTC)

California DOJ Data Dump

Howdy! I'd like to add some data from the California DOJ to [15]. Would this be accepted? I was told my a fellow wikimedian that I should ask first.

  Notified participants of WikiProject Open Government Data

Carolynknights (talk) 20:20, 16 July 2020 (UTC)

This is data on police officer deaths in the line of duty from 2000 to 2019, broken down by month and county. However, I've been told it might be better suited to Wikimedia Commons. Carolynknights (talk) 20:20, 16 July 2020 (UTC)

Changing the domain of a property

  Notified participants of WikiProject Mathematics

Alexander polynomial (P5350) is currently defined as being a property of members of the set of knot (Q1188853). The Alexander polynomial actually seems to be defined over the set of link (Q1760728), of which knots are a subset. (See, for example, this paper.) Should I just boldly edit the property to change its domain, which as far as I can see would not break anything because of the existing set inclusion relationship between knots and links, or is there a procedure I should go through to do this? -- The Anome (talk) 07:51, 16 July 2020 (UTC)

Entities with multiple versions

There are many entities in WikiData that have multiple versions which are notable in and of themselves.

(version lists are not exhaustive and not intended to be exhaustive)

The current state of these items is a bit of a mess. Some problems are (some is true transitively):

So ideally these things should be fixed, but I think it is useful to have some way to not just find instances, but also find the collections of versioned entities. A specific case where this would be helpful is, if BFO was fixed to no longer be an instance of an ontology, but you still wanted some way to find BFO even though there are no items for BFO 1.0 and BFO 2.0. Having something to identify BFO as a collection of versioned entities that are themselves ontologies would make it easier to look for specific versions of ontologies or collections of versions.

As far as I understand one option to make entities HTTP (Q8777) and Basic Formal Ontology (Q4866972) distinguishable from other subclasses is with metaclasses (similar to cell type (Q189118)) however I'm not sure what to name the metaclasses and how they should relate to other entities to be used in these cases.

I think the best option of naming the metaclasses is "series", so there would be, for example:

Another way to solve this may be with series (Q20937557), e.g.:

I'm not sure what the best option is here, any input will be appreciated. I have come across this situation with entities with multiple versions often but I'm always reluctant to fix because I'm not sure what the optimal solution is and I also don't want a different solution for every versionable entity. Iwan.Aucamp (talk) 23:48, 9 July 2020 (UTC)

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

You can also consider editions, perhaps BFO 2.0 is an edition of an abstract work BFO, in the same way that a book can have separate items for its editions. They are all works of text, I guess. Ghouston (talk) 05:25, 10 July 2020 (UTC)
@Ghouston: instance of (P31)version, edition or translation (Q3331189) makes sense for HTTP/2 (Q739120), IPv6 (Q2551624), GNU General Public License, version 3.0 (Q10513445), BFO 2.0, Windows XP (Q11248) and others. However at the same time, to me at least, if we take the case of licenses, this is false GNU General Public License (Q7603)instance of (P31)open-source license (Q97044024) and this is true GNU General Public License, version 3.0 (Q10513445)instance of (P31)open-source license (Q97044024). So then the question still remains what to do with GNU General Public License (Q7603). Maybe GNU General Public License (Q7603)instance of (P31)intellectual work (Q15621286)of (P642)GNU General Public License, version 2.0 (Q10513450) makes sense? But does GNU General Public License (Q7603)subclass of (P279)open-source license (Q97044024) and GNU General Public License, version 3.0 (Q10513445)instance of (P31)GNU General Public License (Q7603) also make sense in this case? I'm guessing not, and maybe that is okay. Iwan.Aucamp (talk) 09:29, 10 July 2020 (UTC)
@Iwan.Aucamp: To me this is more along the fascinating question of whether an identifiable entity which changes over time is to be considered really as one thing (an instance of something) or many things (a subclass of something) - and I don't think there's any rationally justifiable reason to force one viewpoint over the other in general, and on Wikidata we have been very inconsistent about this even in restricted domains (governments, for example, is another case). version, edition or translation (Q3331189) is the solution that's been reached for books and (to some extent) other written works, but as you point out it doesn't quite extend as nicely as one might like. So I think we do need something that is an "instance of" a "thing which may change over time", i.e. a "series of xxx" or maybe a more general "series of states" item? ArthurPSmith (talk) 15:26, 10 July 2020 (UTC)
Lots of things change over time, e.g., humans change over time but somehow we are able to refer to a single individual without making them a subclass of human. Probably because there is only one item in Wikidata, so there's no need to make it a subclass of anything. In these other cases the changes are occurring in discrete steps and there's quite likely going to be a reason to refer to a particular version, and sometimes even Wikipedia articles about a particular version. We have multiple Wikidata items referring to the same entity as it exists at different stages of development. But both the human and the HTML specification are things that change over time. Ghouston (talk) 07:41, 11 July 2020 (UTC)
I suppose another important difference is that all of the HTML specifications written so far do exist simultaneously at the current point of time, whereas the changing instances of the human each only exist at one point in time. Ghouston (talk) 07:43, 11 July 2020 (UTC)
Yet another model is used for software like Linux kernel (Q14579), which has been released in some large number of official versions, and who knows how many unofficial versions, and many of these versions have been indicated with software version identifier (P348) statements instead of separate edition items. Ghouston (talk) 07:52, 11 July 2020 (UTC)
┌────────────────────────────────────────────────────────────────────────────────────────────────────┘
@Ghouston, ArthurPSmith: I think the difference is that most versions of Linux kernel (Q14579) for which software version identifier (P348) is used are not notable enough to have their own items, and I would say in general software version identifier (P348) should not be used if there are items for the versions, or if there are items for the versions and software version identifier (P348) is used then some other relation would still be needed.
However HTTP/1.0 (Q31207264), HTTP/1.1 (Q30500573), BFO 1.0 and 2.0 are notable (IMO) and should have their own items, especially since they have some different characteristics. I somewhat lean to having GNU General Public License (Q7603)subclass of (P279)open-source license (Q97044024) and GNU General Public License, version 3.0 (Q10513445)instance of (P31)GNU General Public License (Q7603) and GNU General Public License (Q7603)instance of (P31)series of software licenses (metaclass), mostly because it is simpler to express shared characteristics especially if those characteristics are expressed with instance of (P31). In part the problem is that if I make GNU General Public License (Q7603)instance of (P31)intellectual work (Q15621286) and GNU General Public License, version 3.0 (Q10513445)edition or translation of (P629)GNU General Public License (Q7603) then GNU General Public License (Q7603) is rather vauge, so I would need basically open source license worksubclass of (P279)intellectual work (Q15621286) and GNU General Public License (Q7603)instance of (P31)open source license work. On the other hand by using a metaclass and subclass of (P279)/instance of (P31) GNU General Public License (Q7603)instance of (P31)open-source license (Q97044024) makes sense. There is however some downsides there also, in some sense assigning common characteristics to GNU General Public License (Q7603) may have limited validity as those characteristics may not hold for new versions but of course the values can be deprecated.
Another complication with edition or translation of (P629) is that it does not really work well recursively, so you cannot have Windows 10 (Q18168774)edition or translation of (P629)Windows NT (Q486487) and Windows NT (Q486487)edition or translation of (P629)Microsoft Windows (Q1406). This is maybe not the clearest case for using subclass of (P279)/instance of (P31) but in cases where edition or translation of (P629) is used for versioned documents you have some weird (wrong IMO) situations where all of the following are instance of (P31)written work (Q47461344): ITU-T Recommendation (Q55935585), ITU-T recommendation version (Q55936923), IEEE standard (Q55755785), IEEE standard version (Q55936932), ISO standard (Q15087423) and ISO standard edition (Q55771109). This situation can of course be fixed, but it is actually not clear how. For example the problem is that the following are wrong as things stand: ITU-T Recommendation (Q55935585)subclass of (P279)international standard (Q1334738) (should be ITU-T Recommendation (Q55935585)subclass of (P279)written work (Q47461344)) and technical standard (Q317623)subclass of (P279)written work (Q47461344) (should be technical standard (Q317623)subclass of (P279)version, edition or translation (Q3331189)).
Not to belabour the point too much but this is somewhat of a mental block to me. Every time I want to deal with versioned items I just get overwhelmed by the very confused situation we have at the moment. I think I'm going to try rework the licenses to use instance of (P31)/subclass of (P279) and see how it goes. Iwan.Aucamp (talk) 21:52, 17 July 2020 (UTC)

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

Merge gadget with not functioning for me in Firefox 78

Hi. Just wishing to check whether other users of Firefox 78 are able to get the merge gadget to function. For me it fails to provide the pop-up box to undertake merge. Checked and no issue with Chrome.

Addendum #wikidata user says they have no issue with similar browser version. Console says 

Uncaught DOMException: Permission denied to access property "ownerDocument" on cross-origin object load.php:179
    jQuery 12
        getDocument
        getDimensions
        getSize
        getSizeProperties
        fitLabel
        emit
        setLabel
        getSetupProcess
        proceed
        execute
        setup
        openWindow

I hope that is helpful to someone.  — billinghurst sDrewth 00:48, 16 July 2020 (UTC)

@billinghurst: Same here, I reported it there:
Tracked in Phabricator
Task T258254
The RedBurn (ϕ) 10:24, 17 July 2020 (UTC)

Are they indentical

Are Marie Ahnighito Peary (Q21557921) and Marie Ahnighito Peary (Q21549681) identical? 15:12, 17 July 2020 (UTC)

Yes, they are identical. They are both about the writer born in 1893, daughter of Robert Peary. We have a lot of duplicate entries that need to be merged. I'll merge these later if someone doesn't beat me to it. From Hill To Shore (talk) 15:24, 17 July 2020 (UTC)

Vernacular Architecture in Spanish

Vernacular Architecture (Q930314) in spanish is Arquitectura vernácula (but has its own code: Q5556463, must be merged), is not Arquitectura popular (it this must have its own code). Please, fix links (of languages). Wikiedro (talk) 18:25, 17 July 2020 (UTC)

Klammerzusatz hinter Vornamen und Familiennamen

Welchen Sinn erfüllen die Klammerzusätze hinter Familiennamen. Hinter der englischen Beschreibung von Kennedy steht im Feld Alias z. B. „Kennedy family name“ oder „Kennedy surname“, meistens jedoch in der Form „Kennedy (family name)“ oder „Kennedy (surname)“. Häufig findet sich das auch bei den anderen Sprachen. Gerade fügt User:Edoderoo mit einem Bot hinter die niederländische Beschreibung „achternaam“ den Alias „Xxxxx (achternaam)“ hinzu. Soll das jetzt auch in allen anderen Sprachen erfolgen? User:Harmonia Amanda könnte das dann gleich in ihrem Script einbauen. Im Fall von User:Edoderoo habe ich jedoch eher den Eindruck, es geht ihm hier um möglichst viele Bearbeitungen, da mir zu dem Klammerzusatz bisher kein Konsens bekannt ist. 158.181.76.68 16:05, 16 July 2020 (UTC)

(What is the meaning of the brackets behind family names. Behind the English description of Kennedy in the field Alias e,g. "Kennedy family name" or "Kennedy surname", but usually in the form of "Kennedy (family name)" or "Kennedy (surname)". This is often also found in the other languages. User: Edoderoo is adding a bot behind the Dutch description "achternaam" to the alias "Xxxxx (achternaam)". Should this now also be done in all other languages? User: Harmonia Amanda could then incorporate that into her script. In the case of User: Edoderoo, however, I have the impression that he is concerned with as many edits as possible, since I am not yet aware of any consensus on the addition of brackets. – Translation by Google)

Well, your impression is wrong. For some lastnames it is easy to to find them back, for for many others it's not, because the name has also 26+ other meanings. Then you can find them back easily with searching for XXX (achternaam) when this is an actual alias. I plan to do this for firstnames and Dutch streetnames (then with the municipality name) as well. For doing many edits I do not have to be concerned, there are hardly no other users with more edits. But that does not mean that every edit of mine is not meaningful, isn't it? Edoderoo (talk) 16:24, 16 July 2020 (UTC)

I'm open to corrections on this, but my impression was that there is pretty strong consensus that labels must not contain any disambiguation hints, and that this extends to aliases (which should be of the same data type as labels). Instead, such information belongs in the description, and it's the data consumer (i. e. search dropdowns) responsibility to surface these hints whenever necessary and appropriate. Here's the relevant section from Help#Labels. Quote:

When a page title includes disambiguation, either through commas or parentheses, the disambiguation should not be included in the Wikidata label. Disambiguation information should instead be part of the description.

Matthias Winkelmann (talk) 19:31, 16 July 2020 (UTC) For labels, yes. For aliases, why? I already explained why to add them. Please tell me why we should forbid it in an alias? Edoderoo (talk) 04:21, 17 July 2020 (UTC)

"Alias" is literally defined as "an alternative name", implying that the type and format requirements are identical to those labels. The label is just one primus inter pares that we/someone believes to be a common choice. It's to allow Barack Obama as well as Barack Hussein Obama, and Barack Hussein Obama II, and even Big Barry Bo, theoretically. All these aliases are actual names (/nicknames) that have been used to refer to the person. Often, there are several that could, conceivably, get the top spot as the label. The use of middle names and/or initials is not always consistent: George HW Bush got the extra "H" only after his son became more well known; Obama's "Hussein" was used exclusively by conspiracy theorists to harm him, and by him to ridicule their silliness. More often, the full name is used for formal occasions, or a stern mother's scolding. Marilyn Monroe is the obvious label, but Norma Jeane Mortenson is not wrong either.
So we try to make a sensible choice for label, but cannot possibly find the one label that is appropriate in every context, because there is no one label that is. Aliases then allow the client to use the advantage they have in knowing the context to make a better choice. I often substitute short(ish) aliases, for example, because my tables have the tendency to outgrow my screen's width.
But that sort of flexibility is only safe as long as every alias has at least some claim to be an actual identifier of the person. When the list of aliases starts to approach cebwiki-like nomenclatural pandemonium (Aach (munisipyo sa Alemanya, Baden-Württemberg Region, Freiburg Region, lat 47,85, long 8,85) ), you're more likely to make the wrong choice, and the aliases are no longer useful for this purpose.
For finding items, there are similar problems: by simple birthday paradox logic, the number of items that have a word in common with your query increases exponentially relative to just a linear increase in aliases per item. Ban Ki-Moon will get lost among Europa (Moon) and Kale (Moon).
My main objection, however, is simply that there is a field, namely "description", that is intended for exactly the purpose you have in mind! Someone thought about exactly these issues at the very start, and they separated these two types of information, thereby making it possible to do disambiguation exactly as you want it, wherever it's needed. It should be almost trivial to display the description along with labels in auto-suggestion lists, for example, while it is either hard or impossible "undo" a mixing of descriptions and names.
It's like being served black coffee with milk on the side: yes, you will need to add it yourself. That might be inconvenient, but it is possible. Removing the milk from the coffee, however, is not. Matthias Winkelmann (talk) 10:09, 17 July 2020 (UTC)
  • Aliases are mainly here to search and select items. It was found that this sometimes not possible without including the dab. Help:Label doesn't apply to aliases. --- Jura 10:38, 17 July 2020 (UTC)
    • My addition of (achternaam) is exactly meant for searching, description is only helpful when there are not more then 5 items, with 50 items (and several of them missing descriptions) it is not going to work. And lastnames/firstnames have to be entered on too many items, to take a bit of extra work by opening/searching in another tab. Edoderoo (talk) 13:43, 17 July 2020 (UTC)
    • That's not correct. Aliases are mainly here to store existing aliases and make them available to users of our data. They do, of course, help immensely finding the correct item when the object in question is listed under other names/titles in external sources/databases. But that's just a very welcome side effect of having this kind of data, not why we have/store this data in the first place. And muddling externally available, pre-existing data with made-up aliases without any way of separating the made-up ones isn't good practice. --Kam Solusar (talk) 15:58, 18 July 2020 (UTC)
      • I think that Alias names are also for improving the search to find the correct item easily. I think that for external data users it is important to understand the structure of the used data. If the used data includes also the Alias names then it is helpful to understand how they are created to identify the ones who are here mostly to improve the search. After I think that currently it is not possible to separate this two parts of Alias names, there should be helps for the data users to extract the Alias who are used mostly for search. For example in the case of the surnames and family names to add a description page for the structure where it is mentioned that there are alias names who are made of the Item Label and the instance of the item in brackets. --Hogü-456 (talk) 20:46, 18 July 2020 (UTC)
    • I understand your frustration with search, because I frequently have the same experience. But it's important to realise that it is the search that is broken, not the data structure. If you search for [cupid moon] in the search field at the top of this page, the quick-pick dropdown that appears is empty, because that function searches in names and aliases only. But hit enter, and you will see the result of a different search that not only includes the description but also uses it to score results, and lists the moon above the (otherwise more famous) god.
You chose to add occupations for disambiguation. But it's far more common to use years of birth and death. And typing names into the google search box, the suggestions add cities or countries more often than not, indicating that many people include these in their strategy for disambiguation. So following your logic, all of these should really be included as aliases as well. At that point, people will start for specific formatting to allow splitting up all that information again, and you have essentially recreated wikidata in the alias.
So please file a ticket with an improvement request for whatever search function is broken. They always need to include descriptions in both the fields searched as well as any result shown. Here on the site, it appears only the quick suggestions in the main search field has that problem? I just tried the UI for adding statements, and it does this perfectly fine, I believe? Which search function do you use where you have these problems? Matthias Winkelmann (talk) 22:31, 18 July 2020 (UTC)
  • I think it would be an error to have "Kennedy (family)", but not "XYZ (surname)" merely because LOC has an entry on one family name, but not the other. Also, please bear in mind that most uses of Wikidata don't have descriptions available for selecting items beyond display. If there is some tool that relies on the presence of "Kennedy (family)" and the absence of "XYZ (surname)", it's probably a sign that it should be using structured statements instead of aliases. --- Jura 10:03, 19 July 2020 (UTC)

Secretary of State Q736559

This item should be splitted. In some languages/country it is a political function e.g. the Netherlands and the US. In other countries it is a civil servant. I would not know how to executed this, perhaps somebody wants to do some study on this point. Ellywa (talk) 16:02, 18 July 2020 (UTC)

    • I hadn't looked at what inherited from it. Yes, there does seem to be a confusion or conflation here of multiple concepts, and the English-language article, at least, seems to be about a term, rather than a concept, hence inevitably a conflation of concepts. - Jmabel (talk) 06:53, 19 July 2020 (UTC)

Archive url

Hello. Do you know if there is a way to archive [16]. I can't find the informations anywhere else. Data Gamer play 15:10, 15 July 2020 (UTC)

@ChristianKl: hello. Theodosis Groutas (Q94618106) date of birth, place of birth, educated at, academic degree, number of children. Data Gamer play 19:07, 17 July 2020 (UTC)

@ChristianKl: I have tried archive.is and web.archive.org but in that case, can not archive the content. Data Gamer play 06:42, 20 July 2020 (UTC)

Permanent links when using Wikipedia articles as a source

English is not my first language, I hope this may be understandable anyway: Sometimes Wikidata uses Wikipedia as a source. I really think there should be a requirement using permanent links when using a Wikipedia-article as a source. Wikipedia is unique in its way presenting permanent links together with extensie version handling, and it would really simplify when reviewing sources used in Wikidata. Has this been discussed or has it just been overseen by the community?--LittleGun (talk) 09:41, 17 July 2020 (UTC)

OK, that was not my point. My point was: When Wikdata is using Wikipedia as a source, I really think permanent link should be used. Has this been discussed or can you do anyway you want? Of course it is possible to do the dig afterwards. But the actual source, is the version that was current when added. Then that one should be stated.--LittleGun (talk) 15:11, 17 July 2020 (UTC)
It's not generally done. The idea is to add actual references rather than to state which Wikipedia it transited through. --- Jura 16:36, 17 July 2020 (UTC)
I know it is not generally done and the idea to use "actual" sources is good. But, Wikipedia is currently used as source. And when it is, I think permanent link should be used. I am surprised the lack of interest for that.--LittleGun (talk) 04:11, 18 July 2020 (UTC)
I'm sure you have some great application in mind. What is it? In general, I think we have a problem with an over-reliancy of tertiary sources. --- Jura 09:08, 18 July 2020 (UTC)
@Jura1: I believe you are missing the point. This is not about promotion about a tool, but a principal discussion on our policy on sourcing. Should Wikidata as a project allow for tools that merely uses imported from Wikimedia project (P143)? Right now we do, but should we really? What if we banned such low quality sources and required any imports to actually do a proper reference using permanent links and made a policy that the use of tools that didn't provide this better quality sourcing also are banned? Existing tools would need to be rewritten, no solution for this is in place yet for all the different tools available. Ainali (talk) 09:56, 18 July 2020 (UTC)
The question is still why? Beyond "really think" LittleGun hasn't provided much of an explanation. You mention "low quality sources", but the quality of Wikipedia doesn't change if you modify the link format. --- Jura 10:02, 18 July 2020 (UTC)
Not of Wikipedia, but of the sourcing statement. Just compare a link to a specfic version of a page with one to a project when used as a source, it is clearly better/more useful to the reader. It's plainly just a lot easier to know where to look if you want to verify what the source says is true. Ainali (talk) 10:26, 18 July 2020 (UTC)
Somehow your sample is (deliberatly or not) incomplete: I think you should compare actual statements. I don't think https://en.wikipedia.org/ is (or should) ever be given. --- Jura 10:30, 18 July 2020 (UTC)
If a statement has the reference imported from Wikimedia project (P143)English Wikipedia (Q328) then that link is all you have (and still one click away!). But if the permanent link to the article that was used to deduce whatever value is being sourced (for example where he was educated) it is much easier to read and see if that claim is made there. Ainali (talk) 10:39, 18 July 2020 (UTC)
In that case, you are doing it wrong. Rather than click on Q328, you should click on the sitelink of the item. In a query, obviously, one can output that directly. --- Jura 10:42, 18 July 2020 (UTC)
That method might work in some cases but imports are based on all sorts of stuff like lists, so the deduction might not even be in the article about the item. And even if it is, it might not be in the current version. In those cases, the permanent url would be very helpful. Ainali (talk) 10:51, 18 July 2020 (UTC)
Jura: Sorry, I did not understand you read something else in to my question. I did not think the advantage using a permanent link required any explanation. It is used just to be transparent to what version was used as a source and to save the reader from the effort to do wiki-archeology comparung datestampes in search for what version was used. I mean as a reader of an fact, checking its source, a totally manual part of being a Wikipedian. With a requirment to Wikidata users to always use an permanent link when using an Wikipedia article as source this is made easier and more transparent. That is the answer to your question "Why", which really was never asked before you intrepreted my answer to it as being "because I really think so".--LittleGun (talk) 11:01, 18 July 2020 (UTC)
I see. np. Maybe we are too much focused users who don't want anything imported from Wikipedia. --- Jura 11:10, 18 July 2020 (UTC)

One "odd" usage Larske was doing a SPARQL query in WIkidata finding all Wikipedia links that were depreciated because of wrong value i.e. when we added a better sources Dictionary of Swedish National Biography ID (P3217) that claimed something else Larske created a SPARQL query that checked when we had conflicting values from a wiki and the "trusted source i.e. we could find errors in the Wikipedia rather easy. If we as LittleGun suggest has good links then we could easy find those errors when a more trusted source gives another value - Salgo60 (talk) 10:12, 18 July 2020 (UTC)

@Jura: dont remember it was > 2 years ago.... I guess user Larske has better memory... the cool thing with this SPARQL approach is that then we can start tracking errors in more wiki languages when we get better sources in Wikidata... maybe we get this for free with the new "Denny Vrandečić project" Abstract Wikipedia (Q96807071) --> we will have less language versions that needs to be changed - Salgo60 (talk) 10:44, 18 July 2020 (UTC)
What did you do when you found different dates? Complete the statement with the actual reference(s) found in Wikipedia and in your preferred tertiary source? Or merely delete the statement not found there? --- Jura 10:50, 18 July 2020 (UTC)
Dictionary of Swedish National Biography ID (P3217) is 99% of the times correct (video about P3217) so I checked the Wiki article and if I understood the language I changed in the Wiki article... but lesson learned was that it is useful to have Wikipedia links in Wikidata even if its not a good source - Salgo60 (talk) 11:10, 18 July 2020 (UTC)
So you systematically discarded referenced dates (contrary to Help:Ranking)? I suppose the Wikipedia link was most of the time the sitelink present on the item. --- Jura 11:16, 18 July 2020 (UTC)
I dont follow you
  • Wikidata had a lot of errors when we added facts about > 8000 people
  • when we had a mismatch we checked and mostly ranked the Dictionary of Swedish National Biography ID (P3217) data as preferred
  • when I had the language skills I changed in the wiki article to the SBL value when I trusted it
  • this is mostly "simple facts" like death born --> we tried to add templates in Swedish Wikipedia to have WD value displayed....
- Salgo60 (talk) 11:36, 18 July 2020 (UTC)
Nice: sounds like a great improvement. Supposedly not all diverging dates had no reference at all, but if you re-ranked them, that's fine.
About the question of sitelinks: what I meant to say is that you could just follow the Wikipedia article link on the item of the person, so a permalink wouldn't really have added anything (worse: you would have had to skip forward to check the current version as well). --- Jura 11:01, 19 July 2020 (UTC)
You are still assuming that imported facts are only based on information in the article about the subject. But it could be based on other things like list articles etc. It can quickly turn into detective work when a permalink at least lets you know why it was added at the time of the edit, and then you can easily jump to the recent version to see if it has been corrected since. If it has been corrected since and a permalink wasn't added you'll have to dig through the history to understand if was added as an error or if there's an error in the Wikipedia article. Ainali (talk) 06:37, 20 July 2020 (UTC)
No, see my question above (10:22, 18 July 2020 (UTC)) --- Jura 06:46, 20 July 2020 (UTC)

Wikidata weekly summary #425

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Events
    • Upcoming: The First Wikidata Workshop: Second Call for Papers. Papers due: August 10, 2020 | Workshop date: October 29, 2020
    • Upcoming: next Wikidata office hour, July 21st at 16:00 UTC (18:00 CEST) in the Wikidata Telegram group. Query Service special with guests from WMF Search Team.
    • Upcoming: Wikidata Lab XXIV: Posicionamento digital relativo with Ederporto - July 23 17:00 UTC (14:00 BRT). In this technical training, we'll study the possibilities and functionalities of relative digital positioning in images and do practical activities on this topic using historical photographs of the city of São Paulo. The event will be held in Portuguese. Join us!
    • Upcoming video: July 21 - Wikipedia Weekly Network - Entity Schemas and Shape Expressions (ShEx) Facebook YouTube
    • Upcoming video: July 25 - Wikipedia Weekly Network - LIVE Wikidata editing #13 Facebook YouTube
    • Upcoming: Kidok-Workshop, online workshop about church building data. In German, non-native users welcome. Currently looking for a date in the upcoming week and people to help!
    • Upcoming video: Beyond Wikipedia - Knowledge that even a computer can understand. July 22, 2020 at 17:00 UTC. In this talk, Zbyszko Papierski will present Wikidata Query Service as one way that developers can interact with Wikidata.
  • Development
    • More work on the consistent design system
    • More work on decoupling the different Wikibase extensions from each other to make development easier:
    • Prepare the first development phase for the Simple Query Builder
    • Reference game: run a scraper, work on a dashboard that will gives an overview of the potential references that have already been judged in the game
    • Fix a bug related to maximum limit of search results (phab:T256885)
    • Resolve some configuration issues regarding Entity Sources (phab:T254315)
    • Federated Properties:
      • Changes to special pages that interact with both items/properties when federation is enabled (phab:T246886)
      • Changes to Special:ListDatatypes when federation is enabled (phab:T255581)
    • Add new monolingual language codes: mic (Miꞌkmaq), gil (Kiribati)

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.

Wikidata reconciliation service unusable.

Am I the only one that finds the Wikidata reconciliation service unusable lately. Even the most trivial of tasks (such as matching countries) takes forever and the results returned do not correspond to the parameters of the query. I am matching against items of type country and the service does not find half of the countries. I am using OpenRefine but i strongly suspect that the issue is on the Wikidata side. Does someone else have similar issues? Has something changed with the service lately? Thanks (and sorry for the rant) :) Nikola Tulechki (talk) 13:44, 20 July 2020 (UTC)

@Nikola Tulechki: There is a new URL to use for OpenRefine reconciliation with Wikidata - https://wikidata.reconci.link/en/api ('en' can be replaced by another language to reconcile in that language). ArthurPSmith (talk) 19:16, 20 July 2020 (UTC)
@ArthurPSmith: Thanks!!! That was exactly the answer I was hoping for! Nikola Tulechki (talk) 06:28, 21 July 2020 (UTC)

Every single wikinews article gets an item?

e.g. Q17869060. For real?--RZuo (talk) 22:46, 20 July 2020 (UTC)

See Wikidata:Wikinews/Development.--GZWDer (talk) 05:15, 21 July 2020 (UTC)

Meaning of the language code "mul"

Hello. What, exactly, is the intended use/meaning of the language code "mul" (multilingual)? Does it mean:

  1. applies to various/some other languages (no clear definition)
  2. applies to all languages
  3. applies to every language unless there is a different value given for a specific language
  4. applies to all languages, even if there are additional values for a given language
  5. no particular meaning, mostly as (informal) fallback
  6. ...

I started wondering about it when some editors started removing the "en" (English) value of unit symbol (P5061), claiming that it is implied by the presence of the same value under "mul" (multilingual). This would be justified if "mul" were intended to be used in the sense 2) or 3), but I could not find any documentation which confirms that meaning. I'd prefer reducing duplication by providing only a single value, for many language, if possible. However, I don't see an easy way to define "mul" in a way that accomplishes that. For instance, the SI Brochure (9th edition) (Q68977219) states that "the symbols for SI units ... are the same in all languages". Would "mul" be appropriate for that? Looking at newton (Q12438), according to the SI, the symbol should be "N" in all languages, but in Russian the letter "Н" seems to be used (in addition to N or exclusively? - how would we model that?). For degree (Q28390), there is currently only one value under "mul": Does that imply that there exists no language in which the symbol might be different? I doubt that such a strong statement could be made. But if the answer is no, then what can be conclude about an arbitrary language "x" (does it use ° or some other value which is missing)?

To conclude, I tend to think that, at the moment, the only viable solution, given the under-definedness of "mul", is to list values for individual languages, even if "mul" is present with the same value.

Let's discuss and potentially document our conclusions. Toni 001 (talk) 08:54, 9 July 2020 (UTC)

  • I thought the point of having P5061 was that people could define one for every language and Russians can read it in Cyrillic --- Jura 09:04, 9 July 2020 (UTC)
  • For instance of (P31)anthroponym (Q10856962) such as Johan (Q10989273)native label (P1705)Johan I think it means sense (2) to me. If someone's name is Johan for example, it may be transliterated (transliteration or transcription (P2440)) to Russian, but that is distinct from the actual name. Where would be a good place to document it? Iwan.Aucamp (talk) 13:28, 9 July 2020 (UTC)
  • "applies to every language unless there is a different value given for a specific" seems to me the most useful definition for labels/aliases/descriptions. When it comes to properties I would expect both the mul name and the language specific one to be valid for every language. It provides us with the ability to reduce the amount of data we store and reduce the amount of edits that are made to the same label in another language while at the same time allowing us to make exceptions when necessary. As far as places to document such a decision https://www.wikidata.org/wiki/Help:Multilingual feels like a good location. ChristianKl ❪✉❫ 20:27, 9 July 2020 (UTC)
    About option 3: Let's say a value "abc" is available for language "mul" and there are no other values. This now implies that whoever entered the value has checked all the hundreds of languages and confirmed that they don't use another value. Whenever support for a new language code is added one would have to check all items where a property value uses "mul" - an impossible task. More generally, as far as I understand, Wikidata works under the open world assumption: The absence of a value is not interpreted as the value not existing (instead, one would explicitly enter a value or the special "no value"). Therefore I'd be cautious about option 3 and rather prefer the informal option 5. Toni 001 (talk) 09:57, 11 July 2020 (UTC)
  • Personally, I'm fine with interpretation (3), although your points about its potential ambiguity are well taken. But for me, the inability to distinguish between "There is no exception for language 'xx'" versus "The exception for language 'xx' has not been listed here yet" is not a crushing one -- especially since, as you say, trying to determine in advance that there are no exceptions is a practically impossible task anyway.
Another example is ohm (Q47083), which currently lists a symbol of Ω for English and Ом for Russian. But Ω clearly isn't "English". (Although arguably it's "multiple" in a different sense than N is for newton (Q12438)...)
Perhaps in the fullness of time it would make sense to have, er, multiple forms of "mul", one for sense (2), one for sense (3), and perhaps one for symbols like Ω. But I'm not sure the Mediawiki developers could be convinced of this, and I'm not sure future editors could be convinced to distinguish and use them correctly. (So maybe I'm really advocating for interpretation (1) or (5).) —Scs (talk) 16:45, 11 July 2020 (UTC)
  •   Comment Why not try to convert case 1 to use "und" instead? --Liuxinyu970226 (talk) 07:23, 17 July 2020 (UTC)
    • The core idea of having mul is that for an item like the one for a person who's name is language independent we don't want to 100 edits to enter the name of the person in 100 different languages and just store on edit. Using und in this way doesn't help with that. ChristianKl ❪✉❫ 13:05, 21 July 2020 (UTC)

Long URL for source

For Q97395427 I have a source for the altitude and the creation date: However this seems to long for Wikidata: pdf.sciencedirectassets.com The Analysis of Railway Line M 202 Zagreb – Rijeka. Can some shoter URL be found?Smiley.toerist (talk) 19:40, 15 July 2020 (UTC)

https://www.sciencedirect.com/science/article/pii/S1877705817326401 ? Visite fortuitement prolongée (talk) 20:17, 15 July 2020 (UTC)
@Smiley.toerist: For a scientific article, please create an item for it, such as Proposed Solutions for Increasing the Capacity of the Mediterranean Corridor on Section Zagreb - Rijeka (Q97412149).--GZWDer (talk) 02:38, 16 July 2020 (UTC)
Its a bit a heavy-duty source, for specifying the altitude of the station. I changed the source. The 1873 date is more problematic as the source does not specificaly state that this station was opened in 1873, only the railway line was opened in 1873. Theoreticaly the station could be opened later. This is very unlikely as at this point the railway (westwards) has a much steeper gradient. In steamtraction times extra steamengines would be added to the train, or the train would be split. For these actions a station is necessary.Smiley.toerist (talk) 11:08, 19 July 2020 (UTC)
The source change with the P1476 property does not really work. You get the article title text displayed, but you cant click through to the item Proposed Solutions for Increasing the Capacity of the Mediterranean Corridor on Section Zagreb - Rijeka (Q97412149) itself.Smiley.toerist (talk) 11:22, 19 July 2020 (UTC)
title (P1476) isn't the right property to link the article. It would be stated in (P248) Proposed Solutions for Increasing the Capacity of the Mediterranean Corridor on Section Zagreb - Rijeka (Q97412149). ChristianKl ❪✉❫ 08:51, 21 July 2020 (UTC)

display resolution of mobile phones: impossible to set

Hi, currently looks like there is no way to set a display resolution (like 1920x1080) for smartphones. Vitaly Zdanevich (talk) 14:37, 20 July 2020 (UTC)

If so, then a property proposal would be welcome. --Matěj Suchánek (talk) 14:39, 21 July 2020 (UTC)

Multiple Publishers and Publication Dates

I have encountered several popular video game (Q7889) from the 1980s that were published on 10-15 different gaming platform (P400) by 5-10 different publisher (P123) around the world at differing times. My practice so far has been to add all the publisher (P123) with qualifiers of platform (P400) and place of place of publication (P291), as well as all the publication date (P577) with qualifiers of platform (P400) and place of publication (P291) and maybe some publisher (P123). Then yesterday I read a discussion from 2016 that I don't fully understand, but seems to indicate "duplicate data" may be an issue.

This process took a great deal of time and I've already resolved to only add the first publisher and publication date in the future, and set it to be "preferred" rank if there are others. However, there are already several video games that I have added all the publishers and publication dates, and I am wondering what changes I should make to them. Rampagingcarrot (talk) 21:09, 20 July 2020 (UTC)

Defender of the Crown (Q1182594) is a particularly grotesque example I just finished (for the sake of completeness). Oh, and I've been doing this for developer (P178) too (though not for Defender of the Crown (Q1182594)) --Rampagingcarrot (talk) 01:01, 21 July 2020 (UTC)

This kind of use cases, to me, begs to create items for each platform-realization of the game − but that’s its own can of worms. See eg User:Diggr/Data Models of Video Games for some thoughts on the topic.
Meanwhile, what you have been doing with the qualifiers looks fine to me ; I’m not sure I see the problem with it?
Jean-Fred (talk) 20:11, 21 July 2020 (UTC)
Spplitting up 35k+ WD items doesn't sound like an easy task.--Trade (talk) 01:19, 22 July 2020 (UTC)

Best property to use

Items with Amazon Standard Identification Number (P5749) will sometimes display "Item model number" and "Batteries" ("4 AA batteries required") under "Product information"

Are there any way to add this information to WD?--Trade (talk) 13:05, 16 July 2020 (UTC)

one of two

How can I define with property constraint (P2302) that a person either should have political prisoner (Q217105) or prisoner of war (Q179637). Pmt (talk) 17:01, 20 July 2020 (UTC)

see Help:Property constraints portal/One of. --- Jura 18:12, 20 July 2020 (UTC)
I believe Help:Property constraints portal/Item is closer but I don't remember now if it works with multiple values (and treats them as OR). --Matěj Suchánek (talk) 14:31, 21 July 2020 (UTC)
Good point. At least if the property isn't <prisoner type>. --- Jura 14:55, 21 July 2020 (UTC)
It's not impossible for someone to be both a prisoner of war and a political prisoner at different times. There are, for example, thousands of (ex)-Taliban who were, arguably, prisoners of war when captured by US soldiers but were subsequently (re-)imprisoned by Afghan government. Matthias Winkelmann (talk) 18:24, 21 July 2020 (UTC)
Seems to work fine with constraint/Item. The thing is that there exist a Norwegian prisoner register person ID (P8269) for person arrested by the occupation forces. The persons arein two categories, either political prisoners or military prisoners, and both in the same register. Tks a lot. Pmt (talk) 20:49, 22 July 2020 (UTC)

Duplicate entries, state of Mexico and federative entity of Mexico

I believe the proper entity is state of Mexico (Q15149663) while the duplicate, federative entity of Mexico (Q20528428), contains nothing of value and should be deleted. The creator of this item, User:Eldizzino has been blocked indefinitely. Jc3s5h (talk) 02:00, 21 July 2020 (UTC)

Eight million people

Now we have more than eight million items with instance of (P31)=human (Q5).--GZWDer (talk) 10:09, 21 July 2020 (UTC)

Amazing, here's to a million more.*Treker (talk) 13:10, 21 July 2020 (UTC)
Can you also count how many of them have any other property except for the OrcID? Edoderoo (talk) 16:46, 21 July 2020 (UTC)
8 million - 1233226 [17]
8 million - 1557508 [18] --- Jura 19:47, 21 July 2020 (UTC)
Great, so we're at about 0.05% coverage of all humans that have lived? Thanks. Mike Peel (talk) 20:02, 21 July 2020 (UTC)

Only 8 mil? There are lots more phone books, year books, and raw birth records of people who have lived at some point in history. Let's get crackin'! -Animalparty (talk) 02:59, 23 July 2020 (UTC)

How to determine if a given work of art has metadata on Wikidata

Hello all- I posted a question on a Commons template talk page that might best be answered by someone from here. I'd be grateful if someone would take a look: Commons:Template_talk:Art_Photo#Guidance_re_syntax_1_vs_syntax_2. The template page itself: Commons:Template:Art_Photo. Eric talk 19:10, 22 July 2020 (UTC)

instance of: Mediterranean country

Currently we have 23 items with instance of (P31):Mediterranean country (Q51576574) (12 of which can be found with a wdt:P31 query; the others being trumped by something having higher rank.)

This doesn't seem to be the correct conceptual level for P31. Should we be using something like located in or next to body of water (P206):Mediterranean Sea (Q4918) for this instead? (Most, but not all, currently also have that). The only other example I can see that is currently similar is 9x instance of (P31):country bordering the Baltic Sea (Q63791824), but there are obviously lots of other potential values that could appear here if this is reasonable modelling. --Oravrattas (talk) 09:46, 22 July 2020 (UTC)

  • There is no requirement for P31 to have just a single view and discard concepts from some thesauri. --- Jura 09:53, 22 July 2020 (UTC)
    • Should we also add P31s for being an ASEAN country, CEMA country, E-9 country, OECD country, Andean country, Islamic country, French speaking country, and all the other similar categorisations from the UNESCO thesaurus? --Oravrattas (talk) 11:00, 22 July 2020 (UTC)
  • I feel somewhat uneasy about this on occastion. Holocaust victim (Q5883980) is often used in P31, and it may be the most significant aspect of this person's life from our perspective. But does that define the person in such totality that it becomes their identity? Alexandra Paul (Q268294) might hate the attention her sister always gets, yet we feel empowered to say she is a twin in a way that must be fundamentally different than being just any sister, because sisters only get sibling (P3373). At the same time, concepts can also be too generic, and if I ever come across a human settlement (Q486972) of people categorising everything from a campsite to New York City as such, I'll burn it down with the same disregard for scale they keep showing (metaphorically). It's a continuum, and Mediterranean country (Q51576574) is probably straddling the border. It's not entirely unreasonable because it's a significant category and people from those countries do feel more commonality with each other than, for example, the OECD countries you mention.
As long as these classes are properly organised as subclasses of human (Q5) etc, it's probably not too terrible in practice. It's somewhat annoying that a query for city (Q515) won't get you a complete list of, well, cities. But it just doesn't seem realistic to find any sort of structure where everything you want can be found under a single precise term, without traversing the tree. And located in or next to body of water (P206):Mediterranean Sea (Q4918) is even more important in that regard, because people traversing along such properties are extremely unlikely to notice when information is hidden away in P31.
I see @Oravrattas: has been removing Italy (Q38)instance of (P31)country (Q6256) and others, and while I get the reasoning, I somehow feel all countries should be labeled as such, and all people as people, irregardless of some resulting redundancy? --Matthias Winkelmann (talk) 11:16, 22 July 2020 (UTC)
I agree. I consider these as a relic from old times, when the relevant properties were not available. I think that it is now just unnecessary duplication. Same case as member states of the United Nations (Q160016) (see the discussion] or (still used) million city (Q1637706).--Jklamo (talk) 12:32, 22 July 2020 (UTC)

Aside: Unless I'm very mistaken, Holocaust victim (Q5883980) should not be used in P31. Humans should be

⟨ FOO ⟩ instance of (P31)   ⟨ human (Q5)      ⟩

and not

⟨ FOO ⟩ instance of (P31)   ⟨ anything else ⟩

. - Jmabel (talk) 15:41, 22 July 2020 (UTC)

Agree with this. There's an open "context of death" property proposal which will let us store this somewhere more appropriate than P31, hopefully. Andrew Gray (talk) 15:58, 22 July 2020 (UTC)
This was my understanding too, but I couldn't quickly find where that's actually spelled out. I suspect it probably is in a Help page or equivalent, but if not, it would be good to actually make that explicit somewhere we can easily point to when things like this come up. --Oravrattas (talk) 16:15, 22 July 2020 (UTC)

Two items about Mario Kart Arcade GP

Please merge the following: Q2575214 and Q76990601.

Thank you. Naddruf (talk) 01:24, 23 July 2020 (UTC)

@Naddruf: they are not the same. one is about the series and one is about the game. BrokenSegue (talk) 02:22, 23 July 2020 (UTC)
@BrokenSegue: If they are different, then maybe they should have different names. I was trying to understand the difference between en:Mario Kart Arcade GP and simple:Mario Kart Arcade GP, and it was confusing. Naddruf (talk) 02:46, 23 July 2020 (UTC)
@Naddruf: Lots of things have the same name but different meanings here. You should look at the descriptions/types to figure out what the item represents. BrokenSegue (talk) 03:10, 23 July 2020 (UTC)
I guess then it's the Wikipedia articles that need qualifiers, so we know what the differences are. Naddruf (talk) 05:10, 23 July 2020 (UTC)

Manufacturer Part Number

Hi there,

what do I map the Manufacturer Part Number property (https://schema.org/mpn) to? There is Part Number but is that actually equal to a manufacturer assigned part number? That seems more specific.

Also Part Number is a node, not a property.  – The preceding unsigned comment was added by F-ludwig (talk • contribs) at 07:05, 19 July 2020‎ (UTC).

Modification of values for Guía Digital del Patrimonio Cultural de Andalucía ID (P3318)

Hi:

I'm in the process of a significant batch import. The source has modified their URL schema simplifying it from i[1-9]\d* to [1-9]\d*. I would like to update the property and the values to the new schema. One plus is the property will be easier to understand for new contributors because never has been explicit. The question here is: how this process should be done? Should I create a new property and migrate the data? Should I just update Guía Digital del Patrimonio Cultural de Andalucía ID (P3318) and reload the correct codes? Thanks. Olea (talk) 08:43, 21 July 2020 (UTC)

I assume the old code does not work anymore, and the new code does: please replace the non-working code with the new working code in the existing property. Edoderoo (talk) 16:47, 21 July 2020 (UTC)
@Edoderoo: The old code still works because the source added an HTTP redirection. The issue is the i letter now is absolutely meaningless for anyone in the world except familiarized wikieditors.
If there is not any opinion against, I'll modify the existing property and will update the related elements. Olea (talk) 14:10, 23 July 2020 (UTC)

Property for related laws

Is there a property for related laws? For example, copyright (Q1297822) was described in copyright law of the People's Republic of China (Q10872186) and An act respecting copyright (Q5169221). copyright term (Q5169289) was described in copyright law of the People's Republic of China (Q10872186) section "第三节_权利的保护期". In Wikidata:List_of_properties/judiciary, I can see laws applied (P3014), but it's not the same.--The Master (talk) 07:43, 23 July 2020 (UTC)

Q10872186 already has main subject (P921) set to Q1297822; qualified as of (P642)=People's Republic of China (Q148). That seems both correct and adequate. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:05, 23 July 2020 (UTC)
For reference works related to a subject, there are both properties to link subject to a work (main subject (P921)) and to link a work in a subject (described by source (P1343)) like in Q2#P1343. However, for laws, there is only one way as Andy said. There is no direct way to see how a subject was described in different countries' laws. I think there should be new property for laws like described by source (P1343) for reference works.--The Master (talk) 02:50, 24 July 2020 (UTC)

Number of deaths per year in a country

I want to record how many people die in a country per year. As a first example I tried it for United States of America (Q30) I used number of deaths (P1120) as a property, but it seems not to be the correct way to do it. How would you do that? Franzsimon (talk) 21:41, 22 July 2020 (UTC)

might make sense to add a new different property to record this or else modify the existing one to accept countries. BrokenSegue (talk) 03:12, 23 July 2020 (UTC)
The property Birthrate [[19]] for a country have been proposed. Also the property Deathrate for a country should be proposed. Pmt (talk) 13:08, 24 July 2020 (UTC)

Two spaces

The header of the page now has two spaces instead of one space between the description and the Q-number Label <> (Q1234567). Is that the intention or is it a software error? The change has only been made a few days ago. --HarryNº2 (talk) 14:47, 24 July 2020 (UTC)

author citation (zoology) (P835) monolingual or mutilingual

Should author citation (zoology) (P835) be multilingual? Please see the problem at Property talk:P835#Monolingual or multilingual?. Thanks in advance. Paucabot (talk) 14:58, 24 July 2020 (UTC)

Constraints about given and family names

Hello. Is there a property constraint we can use about given name (P735) and family name (P734) with language? There a lot of items that have values with given name (P735) and family name (P734) that are wrong. I don't know how this can be done. Some ideas (that may be wrong):

a) if a given name has language of work or name (P407) -> Modern Greek (Q36510) then the person's item must also have name in native language (P1559) -> something in el language and vice versa

b) if a given name has native label (P1705) -> something in el language then the person's item must also have name in native language (P1559) -> something in el language and vice versa

c) if a given name has writing system (P282) -> Greek alphabet (Q8216) then the person's item must also have name in native language (P1559) -> something in el language and vice versa

I can't think something else. I think we need something to warn users that they are adding wrong values in given name (P735) and family name (P734). Any ideas?

Data Gamer play 11:34, 23 July 2020 (UTC)

You say that those names are different, but I'm sure that pleny of other users would suggest to merge those items, because they are the same, only the script is different. Ofcourse names look different in Greek/Russian/Latin script, but Andreas written in Greek script is the same first name as Andreas in English/Dutch/German/French-latin script. Edoderoo (talk) 14:01, 23 July 2020 (UTC)
You are saying that Andreas (Q4926263) (for example) must be used for all cases, in all languages? That is not the answer I took in Wikidata talk:WikiProject Names#Help for Greek names. @Jura1:, @Infovarius:. Data Gamer play 17:03, 23 July 2020 (UTC)
My first name is Alexis / Αλέξης but I am no greek (even if I was thinking about being landlord in Castellorizo) :) Bouzinac (talk) 19:28, 23 July 2020 (UTC)
Wikidata:WikiProject Names#Basic principles: "each string should have a distinct item: the most important property is therefore native label (P1705). Wassyl is not Василь is not Васіль is not Vasyl;" This is the rule, isn't it? Data Gamer play 22:01, 23 July 2020 (UTC)


One constraint could be gender. For example, if sex or gender (P21) -> male (Q6581097) and the given name item is instance of (P31) -> female given name (Q11879590). Data Gamer play 20:16, 24 July 2020 (UTC)

German question

I know that in german ß=ss. Wonder if Neiss (Q37313871) has alias Neiß or should Neiß and Neiss be different family name items ? Bouzinac (talk) 12:03, 24 July 2020 (UTC)

These are two important spellings. A separate data record must be created for each spelling, compare e.g. Rössler (Q56540502), Rößler (Q16866247), Roessler (Q21507330), Roeßler (Q97649163), Roßler (Q97055925) or Rossler (Q37460999). There should be no other spellings in the Aliases field for name entries. --HarryNº2 (talk) 14:46, 24 July 2020 (UTC)
Ok, done here Fritz Neiss (Q94871580) Bouzinac (talk) 15:22, 24 July 2020 (UTC)

ß is not ss. It was not before the Rechtschreibreform, it is not after it. --SCIdude (talk) 15:28, 24 July 2020 (UTC)

Proper names are not affected by the german spelling reform. --HarryNº2 (talk) 15:39, 24 July 2020 (UTC)
According to the entry in the DNB and VIAF, he has several name variants. The lemma itself should be Fritz Neiss, the other variants belong in the field of aliases. --HarryNº2 (talk) 15:39, 24 July 2020 (UTC)

4448 Humans who died before they were born

This is my first contribution/post, so please bear with me. Today, I found that 4448 people were born after they died according to WD. What to do with this insight? Is there any benefit in manually correcting some of these or is there a systematic error somewhere?

Query: https://w.wiki/Xhw

The query should be limited to date precision, but I'm not enough good at sparql to make this query work. I'm pinging Dipsacus fullonum‎, he might be able to help out.
SELECT ?item  ?born ?death
WHERE 
{
  ?item wdt:P31 wd:Q5;
        wdt:P569 ?born;
        wdt:P570 ?death;
        
         ?born wikibase:timePrecision "11"^^xsd:integer . # date precision = day
      ;?death wikibase:timePrecision "11"^^xsd:integer . # date precision = day
  FILTER (?born > ?death)
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
Try it!
Bouzinac (talk) 20:40, 23 July 2020 (UTC)
@Bouzinac: Here is what you tried to do: 
SELECT ?item ?born ?death
WHERE 
{
  ?item wdt:P31 wd:Q5 ;
        p:P569 [ psv:P569 [ wikibase:timePrecision 11 ; wikibase:timeValue ?born  ] ; a wikibase:BestRank ] ;
        p:P570 [ psv:P570 [ wikibase:timePrecision 11 ; wikibase:timeValue ?death ] ; a wikibase:BestRank ] .
  FILTER (?born > ?death)
}
LIMIT 10
Try it!
But I cannot get it to run without timeout if I remove the LIMIT. Maybe it can be split in several queries for subsets of human e.g. after for different birth or death years. --Dipsacus fullonum (talk) 21:19, 23 July 2020 (UTC)
There are 2 related reports: Wikidata:Database reports/items with P569 greater than P570 and Wikidata:Database reports/items with P569=P570 - check queries there. It took me some time to bypass timeouts with subqueries and exclude false positives due to decade/century/etc. precision. --Lockal (talk) 09:41, 24 July 2020 (UTC)
@Lockal: I see that the report with items with P569 greater than P570 doesn't include items born after they died but in the same year and with enough precision to exclude the possibility of it happening on the same date. So there is room it improve the report, but it may have to be split into several reports to cover all cases without timeouts. --Dipsacus fullonum (talk) 07:20, 26 July 2020 (UTC)

question about FAST

Can someone help me with a question about FAST IDs? Worldcat has information about how to create them, but I couldn't find how to search for ones that are already in use. All I know about FAST 104019 is that it's a subject heading for a person named James D. Baker, which is the name of an awful lot of people. Can I find who the creator intended to refer to by finding how it's actually being used? — Levana Taylor (talk) 19:08, 26 July 2020 (UTC)

@Levana Taylor: I'm not an expert on FAST, but if you copy the "SKOS Preferred Label" and run it through the searchFAST tool, you will get details of any ID links. In this case we get the VIAF ID https://viaf.org/viaf/25961991/. VIAF points us to Library of Congress at http://id.loc.gov/authorities/names/n82209161.html and LoC points back to the Wikidata record you created at James D. Baker (Q96925507). From Hill To Shore (talk) 20:45, 26 July 2020 (UTC)
OK, thanks, that search tool is what I was looking for. — Levana Taylor (talk) 21:40, 26 July 2020 (UTC)

Anti-abuse measure

" As an anti-abuse measure, you are limited from performing this action too many times in a short space of time, and you have exceeded this limit. Please try again in a few minutes."

I have constantly been facing this error while performing arbitrary edits. Has there been any changes to some edit filters? 1234qwer1234qwer4 (talk) 23:38, 25 July 2020 (UTC)

I got his recently too, many times in one session. It thought it might be because I was simultaneously running a QuickStatements batch in the name of my main account. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:56, 26 July 2020 (UTC)

Dates on multi-volume works

Some works are published over serveral years, e.g. the 63 volumes of the Dictionary of National Biography (DNB) between 1885 and 1900. The actual publication of the first volume was preceded by several years of preparation (intro: "in 1882 the publisher .. planned"). The question is how to use the following properties:

Currently publication date (P577) has 1885. I think 1882 can be added with inception (P571). "1885 to 1900" with start time (P580) and end time (P582). --- Jura 23:56, 15 July 2020 (UTC)

The editions of the volumes have publication date, the series does not, though you if you are talking the first series and the first run, then the publication dates would be 1885, 1886, 1887, ... 1900 as it was published in each year, rather than a first and last. The guidance at Wikidata:WikiProject Books seems reasonable.  — billinghurst sDrewth 00:25, 16 July 2020 (UTC)
What part of it do you have in mind? It does suggest to use first publication for the work.
I suppose adding 1900 (or all years) with P577 could be another option. --- Jura 07:12, 16 July 2020 (UTC)
I guess we could complete that there accordingly. --- Jura 09:50, 21 July 2020 (UTC)
@Jura1: I read "first publication" as the first date or set of dates of publication, so in this case it is 1885 ... 1900, rather than the reprints in the 1920s. I don't read that as the first publication date of the first volume.  — billinghurst sDrewth 06:19, 27 July 2020 (UTC)

Setting up for military intelligence files for UK NA series, need guidance and fixing

Hi to all. I have been uploading the occasional file from the National Archives series: Dublin Castle files

Example: Piece 207/141: Joseph McBride (1922) (Q97608857)

It started with one or two files, though I have added several more and need better guidance how to set up a parent item for a split file series, and then subsidiary files utilising the parent. There are a number of constraint issues, and I would like some to help me get this better set up.

You can see the couple of handful of files at this Piece 207/ search.

Thanks to our experts.  — billinghurst sDrewth 22:38, 22 July 2020 (UTC)

Any takers? If not, where might one find the requisite support among the projects, I am not certain where it would belong.  — billinghurst sDrewth 06:20, 27 July 2020 (UTC)

Lists on English Wikipedia which also exist in another language

Hello,

At en:Talk:List of active coal-fired power stations in Turkey#Keeping this list in step with other articles and lists I propose that if a list exists in a non-English Wikipedia the corresponding English list should be allowed to use Wikidata for its tables.

Your comments there would be welcome.

Regards

Chidgk1 (talk) 06:13, 26 July 2020 (UTC)

You are right (as I guess co-benefits for other languages would be minor) but it was suggested I advertise it widely. Chidgk1 (talk) 06:36, 27 July 2020 (UTC)

Show two different entries Commons category

Hi all,

I am running the WLE in BR and facing a lot of new challenges, but one thing that I notice is: categories that are depicted as "Wikimedia category" can combine two qid and give in their box info about two subjects as: c:Category:Protected areas of Bahia

But some places, that are important two have 2 information also, we can not do that right now, as Q96741007, that is just a small portion of Q96610485, there is an away to link info as in the previous example?

Thank you. Rodrigo Tetsuo Argenton (talk) 16:39, 26 July 2020 (UTC)

The display of two QIDs happens only when Template:Wikidata Infobox is used on a Commons category page that is designated as a Wikimedia category (Q4167836) on Wikidata. Category:Protected areas of Bahia is in such a situation but Category:Parque Natural Municipal das Andorinhas is not, which is correct in both cases. And then it seems reasonable not to display the full content of Environmental Protection Area of Cachoeira das Andorinhas (Q96610485) in the page you want it to show up in. Commons categories using the template are already taking a huge amount of time to load. Thierry Caro (talk) 07:23, 27 July 2020 (UTC)

Property for noting the sports team of a college/university

Do we have any property for this? I found athletic nickname (Q4813767), but that's not a property. We tie sports teams to their institution by using represents (P1268) at the team page, but I don't know of any property to use at the institution page. {{u|Sdkb}} talk 22:47, 26 July 2020 (UTC)

  • I don't think we should particularly have an overt property in that direction. A college or university may have dozens of sports teams; a team is associated with exactly one university. While our approach here isn't an RDBMS, still a one-to-many relationship is usually better handled by putting one statement each on the side where there are many, not to form a long list on the side where there is one. - 01:14, 27 July 2020 (UTC)
    @Jmabel: Oops, I should've been more precise with my language. I was referring to "sports team" in the sense of the athletics program for an entire college/university, not for an individual sport. For instance, Georgetown Hoyas (Q3141652) is the athletics program for Georgetown University (Q333886), including all the different teams in individual sports. {{u|Sdkb}} talk 04:17, 27 July 2020 (UTC)

Labels/Aliases with brackets

Theres is a little dispute on Sharknado (Q21528034) wheather brackets in labels or aliases are useful for linking or searching an item. In Help:Label I didn‘t found an explicit rule against that, but in my opinion Label/Alias should only be used for real titles. Despite of that one can find a lot of those „helping“ aliases, while labels with brackets are often falsely imported from wikipedia sitelinks. Queryzo (talk) 20:22, 7 July 2020 (UTC)

I don't see why we should exclude brackets from titles at all, either in aliases or the main name.*Treker (talk) 20:38, 7 July 2020 (UTC)
If the actual name has brackets/parentheses, then they are fine. But if they are just disambiguation information, then they belong in the description, not the item name. - Jmabel (talk) 00:33, 8 July 2020 (UTC)
We frequently include them. Otherwise it might not be possible to select the item without using search. --- Jura 05:46, 8 July 2020 (UTC)
That shouldn't happen because aliases aren't primarily for internal use as a search help, but are intended for use outside of Wikidata to display alternative titles/names, etc. I've removed quite a few of these aliases in the past, because I thought they were remnants of bot edits or merges that hadn't been properly cleaned up. --Kam Solusar (talk) 12:25, 8 July 2020 (UTC)
But WHY remove them? Like I and Jura have pointed out it becomes a problem when trying to link to the correct page.*Treker (talk) 21:55, 8 July 2020 (UTC)
Because aliases aren't indended, designed or presented to the outside world as data for purely internal usage. If that were the case, it wouldn't be a problem to add all kinds of aliases in the form of "name"+"vague descriptor". But it indeed is presented to outside users as usable data that can be used to display, well, actual aliases. Currently, external users of our data have no good way to tell actual, pre-existing aliases (like working titles, alternate titles, etc.) from aliases created by users solely to help with their Wikidata-internal searches. Even if they became aware of the problem, they would need to create filters specifically just to remove all the user-created stuff like these "XYZ movie", "YYZ (movie)", "XYZ film", "XYZ 2020", "XYZ movie 2020", etc. aliases.
Yes, finding the correct item can sometimes be a bit cumbersome if you can't immediately find it among the values in the input field, but IMO that means that maybe the search function should be improved, not that it's a good idea to pollute our data with search terms. --Kam Solusar (talk) 00:28, 9 July 2020 (UTC)
Alternatives should be defined as statements, not aliases --- Jura 07:52, 9 July 2020 (UTC)
Yes and no. Official alternative titles of works should of course be used as values for title (P1476) with the appropriate qualifiers. But they should also be added as aliases, because that's what they are and what the aliases data is meant for. But there might also be other commonly used names that aren't official and shouldn't be stored as title (P1476) values. Like different spellings of numbers for sequels - e.g. "2" vs "II" vs "two" ("Saw 2" for Saw II (Q270410)). With people, it can also be used to store combinations of name, full name and pseudonyms/nicknames, like '"John 'Bob' Smith". But not mixed with made-up aliases like "John Smith politician", because that makes aliases unusable for their intended purpose - to provide a list of actual aliases that can be used outside of Wikidata items themselves. --Kam Solusar (talk) 15:11, 10 July 2020 (UTC)
Well either way we really need to create some way to make navigation an searching easier for the editors, I constantly find myself unable to locate the correct item.*Treker (talk) 15:35, 10 July 2020 (UTC)
I think the consensus is/was that one should actually be able to type something in an entity selector to get to the correct item without using full text search. So "XYZ (film)" is perfectly valid. --- Jura 07:46, 11 July 2020 (UTC)
Yes, through a combination of label and a good description, to make it distinguishable from other items with the same label. That way, you may have to scroll through a bunch of items with the same label, but you can eventually find it in the list of items. But I don't think there's a consensus to do it via adding vague search keywords to data that's exposed to outside users who don't know about such uses. --Kam Solusar (talk) 00:17, 13 July 2020 (UTC)
@Kam Solusar: even if not intended aliases are used so and very useful for that. I support Jura here, and sometimes even "XYZ (film, 2000)" should be added because of additional value for search when there are hundreds films with title "XYZ". --Infovarius (talk) 22:26, 13 July 2020 (UTC)
But how do you mark those search keywords as data that shouldn't be exposed to outside users of our data? Those that use the listed aliases like they would use aliases on items about persons (e.g. "XYZ", also known as "ABC", "BCD" or "FGH", is a...) --Kam Solusar (talk) 16:36, 15 July 2020 (UTC)
It seems to me that this might be a problem that needs attention from someone with expertice on how Wikidata is built.*Treker (talk) 13:13, 21 July 2020 (UTC)
Maybe it would help if users could provide some examples where such search keywords are indeed needed. I've worked on a lot of items about movies/TV seres and fictional entities, and I haven't personally encountered many cases where I'd actually need to regularly enter movies/series as values of properties where things like page names of sitelinks aren't searched. Even then, the problem with finding the correct items in my experience were primarily due to missing/incorrect/incomplete labels and missing/incomplete descriptions. If a movie item has the titles used in movie databases as its label or alias and a proper description á la "2019 film directed by Bob Smith", I usually didn't have a problem quickly finding what I'm looking for, though you might have to click "more" a few times, which takes a few seconds.
In the case of Sharknado (Q21528034) linked by Queryzo in the opening post, I'm not even sure how "Sharknado (film series)" would be helpful with finding it, especially with the parentheses. If I enter "Sharknado" in the item selector, this item is currenlty the second result and displays the label "Sharknado" and the description "film series". The other result being the movie of the same name (Sharknado (Q13794921)) and its sequels. So it's already incredibly easy to find. To take advantage of this shortcut, other Wikidata user would need to know that they'd have to add "(film series)" to the title and even then it only saves seconds in most cases. There are indeed some more rare cases where you might have dozens of movies and other works of the same title, but I quickly learned that you can't rely just on the year in the description/alias as years can and do often vary by a year or more between different sources/databases. You have to go by director AND year to be sure, so these generic search keywords wouldn't even help. --Kam Solusar (talk) 16:36, 15 July 2020 (UTC)
https://www.wikidata.org/wiki/Help:Aliases doesn't clearly state whether or not aliases are mainly for internal or external use, however if an external user is interested in multiple titles that a work has the can query title (P1476). If an external user doesn't want to use the main label the likely have an idea about what kind of name (P2561) they want and can query the relevant statements. Aliases are for use-cases where the nature of a name is not relevant like Wikidata search hits. ChristianKl ❪✉❫ 12:57, 21 July 2020 (UTC)

This is one of the (rare, to be enjoyed) cases where both sides of the debate tend to understand the other's point, and occasionally feel their pain, and to disagree only on priorities. So instead of arguing whether we need to define "alias", or if we can just agree on the 2000-year tradition of universally understood meaning that word carries, I want to emphasize how we can solve this without compromising the data model, by simply focussing on getting the tools fixed on the client side.

There are two related but distinct issues here:

  • a wish to "tag" an item with certain descriptive terms to either get them included in matching searches, or to boost their rank within search results.
  • the bad experience of being presented with a list of search results that show only labels, and therefore often show several indistinguishable items.

Both issues stem from choices made by tools, namely to privilege labels (and inter alia, aliases) over descriptions and properties. To a certain degree, that is how it is intended, and why labels are somewhat privileged space. Unfortunately, in some circumstances, only labels and aliases are considered/shown right now, leading to these problems.

Adding more words to the privileged space is then the immediate solution to the specific pain you are experiencing. It is not, however, actually helpful! Over time, it will just devalue labels by inflation, increase the number of false positives in search, and, like any good arms race, lead to calls for extra-extra-special labels, starting the cycle anew. (Edit: already happening) Where tools chose to include labels only in their search indices or user interfaces to improve performance and preserve space, respectively, they will see diminishing results over time. At the limit, all item information will be stuffed into the label by people trying to work around some issue, and the only difference to the start point is that it's just prose again, instead of structured properties.

It seems pretty obvious that the far better approach here is to improve the tools. Indeed I am somewhat certain descriptions were included from the start for exactly this use case. Unfortunately, the documentation, API, and other affordances did not gently guide toolmakers to a set of sensible defaults. There should be some prominent guidance specifying that, in lists, descriptions should almost always be shown alongside labels. And searching should default to including descriptions and property values, with appropriate discounting, wherever feasible.

Because I kept running into this issue using OpenRefine, I asked for descriptions to be included by default in the spec. This being an effort to establish a universal spec for reconciliation services, the (incredibly fast) fix should in due course improve the situation across many services and client tools. If you run into this problem in other circumstances, try bringing it up with the service or software you are using. It's liable to be a major pain point for many of their users, and probably has the sort of effort/reward ratio that makes it an instant star in their issue tracker. Matthias Winkelmann (talk) 11:09, 27 July 2020 (UTC)

Do Genes have authors?

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

I came across a bunch of items added by @Leishmania-CBMSO: such as

LBRM2904_01.0090 (Leishmania braziliensis protein-coding gene) with

instance of (P31)protein-coding gene (Q20747295)author (P50)Mauricio Alejandro Reinoso Dueñas (Q95993101).

Here are 290 of these. Now I'm wondering:

  • Should genes have authors? Apart from a possible counterclaim by any item in deity (Q178885), some sort of credit might not be entirely absurd, even if it would need to be closer to the sense of "discovery" than authorship. But, currently, there are no other instances with listed authors, not even BRCA1 (Q227339) which won a Nobel IIRC.
  • Should these genes have authors? Many of them are described as hypothetical, meaning that this they aren't individually studied, but their function is inferred from similarity of their sequence to better-known genes in other organisms.
  • Should they have items, at all?
  • Does it make a difference that the user seems to be linked and/or identical to the claimed author Mauricio Alejandro Reinoso Dueñas (Q95993101) who, to complete the circle, was in turn created by a user of that name (@Mauricioreinoso96:)? Matthias Winkelmann (talk) 20:15, 25 July 2020 (UTC)
@JMRR-UAM: from the same institution who might also help resolve this matter. Mahir256 (talk) 20:20, 25 July 2020 (UTC)
Probably discoverer or inventor (P61) is meant to be used.--GZWDer (talk) 23:18, 25 July 2020 (UTC)
These are entities produced by natural processes, any authorship is nonsense. The user is messing around for some time (e.g. Q65500432). --SCIdude (talk) 08:12, 26 July 2020 (UTC)
As to "should these have items?" the WikiProject Molecular Biology has not given the scope of its endeavor, so anything goes, as long as there are no duplicates. However, the items are quite obscure, from yet another variant of an organism that is already present. Needed would be imports like Arabidopsis or Dicty, but please contact the project before you start. --SCIdude (talk) 08:23, 26 July 2020 (UTC)
This is a really interesting question. Mostly, genes have been naturally produced (through an evolutionary process) by the organisms which contain them into their genomic material. Nevertheless, the concept has been created by scientists, and its structural features defined after experimental analysis. Therefore, the scientist, in some way, can be considered as the author for a given gene. Nevertheless, after reading some comments above, I agree that term like ‘descubridor’ (P61) may be more appropriate. JMRR-UAM (talk) 07:27, 27 July 2020 (UTC)
The Wikidata item is about It's not necessary to be a deity to create entirely new genes and then synthesize them but the genes in question are the result of natural evolution and thus shouldn't be listed as having an author. discoverer or inventor (P61) works better for the use-case. ChristianKl ❪✉❫ 11:50, 27 July 2020 (UTC)

revise regex for WorldCat Identities ID

The regex constraint for WorldCat Identities ID (superseded) (P7859) needs to be revised to take into account values like this, which is in fact valid. — Levana Taylor (talk) 15:24, 26 July 2020 (UTC)

Do you have any clue why the property proposal does not mention them? I mean, it might have been intention or oversight. --Matěj Suchánek (talk) 11:02, 27 July 2020 (UTC)
Definitely oversight - we'd seen 'np-' examples, but not this 'nc-' format. I'll fix the constraint. ArthurPSmith (talk) 14:25, 27 July 2020 (UTC)

Ability to soft inherit property OF THE PARENT rather than replicate to all subparts

Is there, or can there be, the ability to have an item "soft inherit" a parent's property/qualifier rather than having to hard replicate to a daughter item? So here I am talking about a work that has an item, and then the chapters of the work have their items, some the same as the parent, some different. Numbers of the items of the daughter items should inherit the parent's field, or I would like to designate it to be so.

Living examples

At this time, where there is no difference between parent<->daughter version of the property/qualifier, I leave the daughter items bare.  — billinghurst sDrewth 06:16, 27 July 2020 (UTC)

@billinghurst: Thank you for bringing this up! As I understand it, a number of property values assigned to an item are assumed to apply also to its daughter items, unless another explicit statement replaces that value. For example, if a physical object such as a major building has a geographic location (country, coordinates, nearest public transit stop and so on), there is little reason to believe those property values to be different for every single section, office, or other part of the sane building (the library, the restaurant, the parking garage etc).
However, you cannot automatically retrieve all those "inherited" (or "inheritable") property values with a single query, and I don't think it would be easy to implement, because the inheritability (isn't that a word?) of any single property can be disputed, and we are approaching 10,000 different Wikidata properties...
Two weeks ago I began working on a Lua module to experiment with dynamic property inheritance algorithms on my own, in the hope that some of it could be useful to Wikidata usage in general. I'm new to Lua and using this as a learning exercise, so please have patience with any poor coding style you may encounter, but any comments or suggestions are welcome. Also, don't expect the code to be usable out of the box, as it's far from ready, and I think various caching mechanisms and optimizations will be required to avoid overburdening the Wikidata servers with dozens of non-localized queries for each property value.
But it's not just about the code. It's also about defining the inheritance characteristics of nearly 10,000 Wikidata properties, and how they relate to the transitive (and possibly other) property paths they are inherited via; subclass of (P279), part of (P361) and so on. Should the inheritance be exclusionary (closer values replacing more distant ones), additive (values accumulating on top of each other), or conditional (dependent on various qualifiers and other characteristics)?
And it all depends very much on the quality of those inheritance paths. Every single Wikidata item is of course either an instance of (P31) or a subclass of (P279) some class item leading back to the entity (Q35120) class root... not! So maybe this won't be ready for launch tomorrow. Or next week. Or... never mind. But I'd like to look further into this. --SM5POR (talk) 19:19, 27 July 2020 (UTC)
Wikidata:Item classification seems slightly out of date, but gives a reasonable overview of how the concept of inheritance applies (or doesn't) within Wikidata. --Oravrattas (talk) 20:23, 27 July 2020 (UTC)

Sports

Would anyone here mind if add sport (P641): racing (Q878123) to video gsmes in the racing video game (Q860750) genre? I noticed that it gave a constraint violation --Trade (talk) 11:22, 27 July 2020 (UTC)

I fixed the question @ChristianKl:--Trade (talk) 14:00, 27 July 2020 (UTC)

Wikidata weekly summary #426

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Tool of the week
    • We would love suggestions for tools to include in this section of the weekly summary. Please add your suggestions directly under Status updates/Next#Backlog after checking that the tool isn't already listed.
  • Development
    • Changed the size of image previews to 1024 in the gallery view of the query service to avoid some images not loading sometimes (phabricator:T258241)
    • Added an actual space between the entity title and the name of the fallback language (if any), so that the fallback language isn't selected anymore when double-clicking the entity title for copying (phabricator:T256857)
    • Fixed the directionality of text pieces in placeholders that mix LTR and RTL (phabricator:T253812)
    • Continued work on first pieces of design system to make coding new features easier in the future
    • Continued untangling the code of Wikibase Client and Wikibase Repo to make it easier to develop on them
    • Finished first piece of research on how to make it easier to access Wikidata's data for programmers - more work to be done
    • Preparing to start coding on the Query Builder to make it easier to create queries without having to know SPARQL
    • Finished running the scraper that gets potential new references for unreferenced statements and preparing it for publishing

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.

impossible to set email because of error `This URL misses a scheme like "https://":`

Vitaly Zdanevich (talk) 15:51, 27 July 2020 (UTC)

A Wikidata Lookup Bookmarklet from any URL?

Here's my idea. You are browsing the web, and find a page on any other site about an item that should have a Wikidata item. You click a bookmarklet, which instantly tells you in a side panel which identifier property should be linked, and if it is linked, which Wikidata item the ID is linked from. Wouldn't that be amazing? How many of you would use this?

The idea was spurred by User:rdmpage in this video (thanks!).

As far as I can see it should be technically possible, but I'm sure you friendly hivemind can improve dramatically on this in-principle sketch:

  1. Bookmarklet runs javascript to open a side panel or overlay div, in which the content is an iframe containing an auto-running WDQS query which takes a single input, the current URL.
  2. The SPARQL query finds out whether that URL matches a formatter URL (P1630) or a third-party formatter URL (P3303), and if so, extracts the ID from the URL.
  3. Then it queries whether that ID is listed for that property, and returns the Wikidata item(s) with that ID for that property. QED.

I've worked on this a bit, and while I'm good enough at SPARQL, and have written a working demonstration query, (this shows the result from 4 different URLs - I'm aware that it doesn't work for *all* URL schemes, but I think it already covers over 90% of them). I don't have sufficient .js skills to get the bookmarklet working at the moment, and I'm sure some of you can do a much better job anyway. Would anyone be willing to help work on this? I think it would be so useful to many of us.

What do you think? --99of9 (talk) 13:26, 18 July 2020 (UTC)

That sounds quite similar to phabricator:T253201. --Lydia Pintscher (WMDE) (talk) 18:43, 18 July 2020 (UTC)
Thanks. Yes, I figured it could go further in that direction, but didn't want to be too ambitious at first! So are bookmarklets or browser extensions the more suitable approach? Who's keen to help? --99of9 (talk) 11:49, 19 July 2020 (UTC)
Here's a video demo of my first version! @Lydia Pintscher (WMDE): --99of9 (talk) 00:49, 23 July 2020 (UTC)
@99of9: That's really awesome! I'm so happy you picked it up. --Lydia Pintscher (WMDE) (talk) 08:17, 23 July 2020 (UTC)
My initial name "Wikidata Connect" was rejected because it's "best to avoid using Wikidata as the first word in the app title". Has anyone got any other suggestions? Brainstorm time! --99of9 (talk) 14:46, 25 July 2020 (UTC)
That's really nice. I look forward to being able to use it. A few questions:
  • In the case where there is no match, could you offer search results before suggesting item creation? That would reduce duplicate items (and duplicated effort).
  • How does item creation work? Does it link to an auto-import tool that will pick up fields in a source-specific way?
  • Does the tool distinguish between "Page does not match any formatter template" and "Page matches template but identifier not found"?
Cheers, Bovlb (talk) 14:53, 27 July 2020 (UTC)
@Bovlb: Thanks for your thoughtful engagement. At the moment the "new item" link just opens a new tab to Special:NewItem. So both of your first two suggestions would be a huge improvement, but because they would need to parse the page for at least a title of an item, I think that would be too challenging for me to build in at this stage. I'm currently only using the URL itself. Your third point is what I'm working on right now. It wasn't distinguished in the demo, but it will be in the next version. The idea is that the icon will light up in amber if you go to a page that matches the formatter template and regex (without actually running any query, so your web browsing privacy is protected). If you then click the amber light, it sends the query and either turns green or red. Green is a full match, and shows the dropdown. Red is a template match but identifier not found which should then prompt toward search/creation. How does that sound? It's super helpful to think through this kind of question, thanks. --99of9 (talk) 11:09, 28 July 2020 (UTC)
@99of9: That sounds good. I believe there are other projects that have worked on the source-specific scraping, like meta:Mix'n'match. It's possible that you would be to be able to exploit those scrapers directly for new item creation, but I don't know the details. Bovlb (talk) 14:50, 28 July 2020 (UTC)

property for a deputy?

In 1818, James Baker (Q96798633) became chancellor of the Diocese of Durham; he was living in Durham at the time. In 1825, he became rector of Nuneham Courtenay and moved to Oxfordshire, but he didn't give up the chancellorship; instead, he appointed James Raine (Q6141791) as his "Principal Surrogate in the Consistory Court of Durham." Is there a property that would indicate the relationship between the two men: someone who is delegated to carry out someone else's administrative responsibilities when they can't be physically present? — Levana Taylor (talk) 04:05, 28 July 2020 (UTC)

@Levana Taylor: substitute/deputy/replacement of office/officeholder (P2098)? --Oravrattas (talk) 06:07, 28 July 2020 (UTC)
No, that's the property for titles, "Job-Title-B is the title for the substitute for Job-Title-A." I'm looking for something like "Person-B was deputy for Person-A; qualifier: in job X" or some other arrangement of those facts. I suppose I could create two items, "Principal Surrogate for the Chancellor of a Diocese" and "Chancellor of a Diocese," related by substitute/deputy/replacement of office/officeholder (P2098); then if Wikidata contains information showing that Baker was Chancellor of Durham from 1818 onwards, and Raine was Surrogate to the Chancellor of Durham from 1825 onward, you can deduce that Raine was surrogate to Baker. How would you find out, looking at the information from Baker's end, that he had a surrogate and who it was? The necessary information would be there, but indirectly. — Levana Taylor (talk) 07:01, 28 July 2020 (UTC)
OK, I tried creating those items Spiritual Chancellor (Q97768814) and Principal Surrogate in the Consistory Court (Q97768920) and it seems like it isn't appropriate to link them with substitute/deputy/replacement of office/officeholder (P2098) -- the constraints on the property show that it's thought of as purely applying to political offices that have a term length and a political jurisdiction within a country. Without that, there is no way of indicating a relationship either between the administrative roles of chancellor and chancellor's surrogate, or between the people who fill those roles. — Levana Taylor (talk) 13:56, 28 July 2020 (UTC)
I think part of the issue here is that those items are a bit too broad. For the example you gave above, you would be better with items for "Chancellor of Durham" and "Surrogate to the Chancellor of Durham" (which can use subclass of (P279) to connect the items you have created). Those could then have a jurisdiction of the Diocese of Durham, and if there are no explicit term lengths for these you can explicitly set that to novalue. --Oravrattas (talk) 14:29, 28 July 2020 (UTC)

Special:Contributions/Oribarak123456

This user's edits need some attention.--GZWDer (talk) 16:37, 28 July 2020 (UTC)

  • Looks probably well-intentioned but not understanding properties etc. very well (e.g. a person is not a instance of (P31) of a country). Someone needs to take this person under wing, preferably someone who speaks Hebrew. - Jmabel (talk) 00:47, 29 July 2020 (UTC)

WikiProject Amarna

If enough people are interested, we should create WikiProject Amarna or WikiProject AmarnaMD (Metadata), to create and maintain a repository of metadata about all the items (people, places, objects, etc...) from the Amarna Period (Q455151) of Ancient Egypt (Q11768).

This period includes Nefertiti (Q40930), Akhenaten (Q81794), & Tutankhamun (Q12154), but hopefully this project can help expand their relatives, consorts such as Ankhesenamun (Q230863), architecture, culture, technology, and the work they produced and patronised like the examples of Amarna art (Q2740302).

The project itself could be divided into three parts; Pre-Amarna (1391 to 1353 BCE), Amarna (1353 to 1332 BCE) & Post-Amarna (1332 to 1322 BCE) periods, which can help expand the context of the era.

Eventually, the project can grow or become a part of a dedicated project for the entire Eighteenth Dynasty of Egypt (Q146055), which would include Hatshepsut (Q129234) and her contemporaries and works, such as the Mortuary Temple of Hatshepsut (Q660692). Further, it can grow to encapsulate the entire New Kingdom of Egypt (Q180568) and then most of Ancient Egypt.


The idea for the project came when, I wanted to and will start to add and expand statements in the individual Amarna letters (Q235502). But I thought that more eyes should be present on these items, as I do not have enough expertise in this area, but have a keen interest in the topic. And seeing that no dedicated Wikidata Project exists for Ancient Egypt (unless I have missed something), I thought this could be a good starting point. There is also only a finite amount of information regarding this era, and as such the project could be semi-completed, until more information is uncovered. I am open to any and all suggestions, including a different title for the project to future proof it, or creating a chat group rather than a WikiProject.

I do not posses the knowledge to create a WikiProject, as such am posting this here. If this information belongs somewhere else, please guide me. And if anyone is interested in the above please reply to this or leave a message on my talk page. Wallacegromit1 (talk) 16:54, 28 July 2020 (UTC)

Layout issue : at certain font sizes, entry field disappears

Tracked in Phabricator
Task T259183
 
Display Error Wikidata Screen Shot 2020-07-21

I have recently had an issue with the display which seems to be a result of font size and column layout. Since I am short-sighted I usually increase the font several times. At certain font sizes, the field for entering a value for a reference URL is covered by the remove icon. (i.e. the remove icon and word are on top of the entry field, which is effectively invisible.) As far as I can tell the entry field's display is only 1 character wide. If I increase the font size further, the remove button actually ends up to the LEFT of the entry field, next to or on top of the "reference URL". So a good fix might involve rethinking the width and placement of the columns and fields displayed. I was able to work around by decreasing the font size, but this is awkward for someone shortsighted. Mary Mark Ockerbloom (talk) 15:41, 21 July 2020 (UTC)

I've had this same problem on my iPad, and only for the last few days. It was never a problem before that. - PKM (talk) 19:33, 21 July 2020 (UTC)
@Mary Mark Ockerbloom, PKM: Thanks for letting us know. We're going to look into it. -Mohammed Sadat (WMDE) (talk) 19:58, 22 July 2020 (UTC)
Is there a Phabricator ticket for this? It’s impossible to add qualifiers via iPad at the moment. - PKM (talk) 22:56, 27 July 2020 (UTC)
@Mohammed Sadat (WMDE): should I open a Phabricator ticket for this? The problem has been happening for a couple of weeks now. It’s really frustrating. - PKM (talk) 19:34, 28 July 2020 (UTC)
@PKM: Yes. Please open a Phabricator ticket, and screenshots in it will be helpful. -Mohammed Sadat (WMDE) (talk) 08:37, 29 July 2020 (UTC)
@Mohammed Sadat (WMDE), Mary Mark Ockerbloom: Done, Task T259183. - PKM (talk) 19:09, 29 July 2020 (UTC)

Problems with Listeria

Hi, I see an issue with Listeria that it's going on since hours and it has not stopped yet. Pages like this one are not updated. Any clue when it might resume?--Alexmar983 (talk) 17:25, 22 July 2020 (UTC)

I don't think ListeriaBot updates pages so rapidly by itself. The updates might have been triggered manually (by visiting "Aggiorna manualmente" link). --Matěj Suchánek (talk) 10:08, 23 July 2020 (UTC)
It was broken yesterday afternoon. I left a note for Magnus and it came back later that day. --- Jura 10:12, 23 July 2020 (UTC)
I was referring to the manual update. If you tried to click there, you could have noticed yourself. Thank you Jura for the note, so I should write directly to Magnus in the future? I use Listeria a lot in July-September and these types of issue appear once in a while, I assume 2 hours of trouble is enough to leave a note. --Alexmar983 (talk) 16:50, 29 July 2020 (UTC)

Clarifying abstract entity schemas

The entity schemas E3 ((Wikidata Item), E5 (Statement), E6 (Language mappings), E7 (Citation), E8 (External RDF), and E9 (Wikidata-Wikibase) are all blank.

  1. There should be something there, even if it is all comments and optional elements.
  2. Is there a quick way to find other blank entity schemas?
  3. Was there discussion about this when schemas were being created for the first time?

Sj (talk) 16:44, 24 July 2020 (UTC)

See phab:T229775.--GZWDer (talk) 17:18, 24 July 2020 (UTC)
Thank you @GZWDer:, helpfully clarifying but that discussion too is blank... AKlapper, as the only one (auto)subscribed to that ticket - is this on the roadmap? Is it controversial that none should be blank? Warmly, Sj (talk) 17:30, 24 July 2020 (UTC)

Other Qs about entity schema identifiers and properties

  • EntitySchema labels don't seem to be working for me - do they need to be updated to a new format?
  • There are a few schemas with odd numbers: E734 (family name), E735 (given name), E999 (Borked), E11424 (film). Are these actually in use? What was the mechanism of generating those EIDs, do we want to keep it, will the ID incrementing pass over them smoothly?

Sj (talk) 17:00, 24 July 2020 (UTC)

See phab:T223997.--GZWDer (talk) 17:18, 24 July 2020 (UTC)
Thanks again. @Andrawaag: can we create initial schemas for these? Sj (talk) 17:30, 24 July 2020 (UTC)

Cannot submit an item

Hello! I represent the Marketing team of Changelly, the international cryptocurrency exchange platform. We have tried to submit an item describing our company, but it was tagged as “spammy”. However, I believe it was a misconception: our service is respected by thousands of cryptocurrency users all over the world and we would really like to get a mention on Wikidata. We work hard every day to provide the best service in this innovative field to our clients and to make our platform as transparent as possible to exclude the cases of abuse of any kind. You have articles about Binance and Coinbase, so we do not understand why you cannot include our service as well. I hope you will at least consider my request. Thank you!  – The preceding unsigned comment was added by Paulina Szen (talk • contribs) at 17:39, 29 July 2020‎ (UTC).

      • @ChristianKl: @Vahurzpu: The item was submitted by a non-registered person (me). Could anyone here help me solve the problem? I don't see why the filter doesn't let me submit the item considering that similar items exist on Wikidata.

translate

Pages that link to "Qxxxxxxx". Displayed xxxx items.

Where can I translate "Displayed xxxx items"? Data Gamer play 18:03, 30 July 2020 (UTC)

mediawikiwiki:Translator hub ? Visite fortuitement prolongée (talk) 19:22, 30 July 2020 (UTC)
Data Gamer, generally speaking, MediaWiki interface messages are translated on TranslateWiki as mentioned on mw:Project:Translation (also linked from the page Visite fortuitement prolongée linked above). Assuming your were using Greek as your interface language, I can see that the translation not yet been done at translatewiki:MediaWiki:Whatlinkshere-count/el. If you're interested in helping with translations there, it could eventually make its way to this and many other wikis.  Hazard SJ  00:19, 31 July 2020 (UTC)

(sub)class or instance of

Regarding Lindabrunn Conglomerate (Q1825913), I was reverted about changing von subclass of (P279) to instance of (P31). With comment is a class, not an instance (particular piece of rock) If it were a particular piece of rock, I would use made from material (P186). While objects made of Lindabrunn Conglomerate (Q1825913) are a class, the material itself is not. IMO. Your opinions? @Yger, Вадзім Медзяноўскі: with similar cases. --Herzi Pinki (talk) 14:45, 28 July 2020 (UTC)

Given item is about an abstract object (a class) of which a concrete object (an instance) would be e.g. a particular chunk of material mined in a quarry, or chunk cut into a particular statue. See Help:Basic membership properties. I suppose it depends on topic which instances should use P31 and/or P186 for material. 2001:7D0:81F7:B580:2C09:1DBA:9CDE:4D10 15:09, 28 July 2020 (UTC)
@Herzi Pinki: There seems to be a common misconception that an abstract object without a physical identity and location in space can never be an instance, but is always a class, and I dispute this. Therefore I disagree with the explanation given for that revert (at least in the way I read it); an instance need not be a "particular piece of rock" but may well apply also to a particular kind or type of rock, like the Lindabrunn conglomerate.
However, the latter is true only when "rock" is interpreted as "type of rock", i.e. the class of all rock types (granite (Q41177), conglomerate (Q191704), quartzite (Q237883), tuff (Q484924) etc). Somewhat dependent on which language is used, a term like "X" may either describe the most generic type of X, or the "type of X" property itself, which can be utterly confusing.
The most generic type of rock (Q8063) would in my opinion be stone (Q22731), which I'd say is an instance of (P31) rock (Q8063). Wikidata currently treats "stone" and "rock" as synonymous labels for it plus defines it as a subclass of (P279) rock (Q8063) (which makes little sense in Swedish, as "bergart" clearly means "type of rock", not just "rock" or "stone"; it's like saying that "organism" is synonymous to or a subclass of "species"), and en:Stone redirects to en:Rock (geology).
All immediate or indirect subclasses of stone (Q22731), down to the Lindabrunn conglomerate and beyond, would also qualify as instances of rock (Q8063) (even though you don't need to spell this out for every subclass). However, due to the conflation of stone (Q22731) with rock (Q8063), they have been made subclasses instead. If most editors can't tell (or disagree about) the distinction between "instance" and "subclass", or A and B, why confuse matters by using different words or names for them..?
As a general rule of thumb, you can subdivide a class into as many subclasses you like without upsetting the interpretation of the tree. However, you need to keep track of the number of instance of (P31) in a statement path between two items, as each instance of (P31) represents a conceptual jump between two different levels of abstraction.
The latter kind of relationship can be illustrated by this chain of instance of (P31) statements in a related domain pertaining the Koh-i-Noor (Q212797), whicḧ is a diamond (Q5283), which is a mineral species (Q12089225), which is a rank of mineral taxonomy (Q55084935), which is a third-order class (Q24017465), and so on back to variable-order class (Q23958852), which is an instance of itself.
Thus I actually agree with this revert, although I dispute the explanation given for it. Further discussion on this topic should preferably take place at Help talk:Basic membership properties, to provide feedback and suggest improvements to the Wikidata documentation. --SM5POR (talk) 10:45, 29 July 2020 (UTC)
Yes, a class, in addition to being a subclass of something, can be an instance too, i.e. an instace of a metaclass (class of classes). In edit summary I meant that "instance of" relation doesn't apply as conglomerate (Q191704) isn't a metaclass.
As for whether rock (Q8063) should be regarded as a metaclass, I think it shouldn't. While Swedish term "bergart" sounds more like "rock type", this isn't the case for most other languages, is it? Rock is a kind of physical object (Q223557) and hence its subclass. If as a metaclass it was instead in a subclass tree of class (Q16889133), then particular rock types, in addition to being instances of a metaclass, should still be in a subclass tree of physical objects as well. Currently there doesn't seem to be an additonal classification scheme for the latter distinction, similar to subclasses of mineral (Q7946) that additionally are instances of some rank of mineral taxonomy (Q55084935).
Relation between rock (Q8063) and stone (Q22731) is qustionable, but I wouldn't go deep into it here, other than noting that the first one is more clear-cut concept in geology, while it is more ambiguous what the latter item is about. 2001:7D0:81F7:B580:E81C:A170:2BE7:7B3D 08:19, 30 July 2020 (UTC)

bringing in metaclasses and classes of classes seems to be a strawman argument. It does not help. (Arguing with numbers is rather tedious   and does not really work cross languages, if terms are slightly different.)
Lindabrunn Conglomerate (Q1825913) and conglomerate (Q191704) are rock (Q8063), not a stone in the sense of a peace of stone or a diamond (Q5283). While conglomerate (Q191704) is a general type of rock (Q8063), Lindabrunn Conglomerate (Q1825913) is a very concrete, very local subtype of conglomerate (Q191704). It is not an abstraction. Just as phosphorus (Q674) instance of (P31) chemical element (Q11344) and not a subclass of chemical element (Q11344). ev. Lindabrunn Conglomerate (Q1825913) instance of (P31) formation (Q736917). best --Herzi Pinki (talk) 15:55, 30 July 2020 (UTC)

I also wouldn't have brought up cases where item is both instance of and subclass of something, and hence metaclasses, as this initally wasn't the issue. Though, if you argue that one class (Lindabrunn Conglomerate (Q1825913)) is instance of another class, then this indicates that this another class (conglomerate (Q191704)) should be a metaclass. Currently at least the latter isn't classified as metaclass, the way chemical element (Q11344) is in your example.
Both, what you describe as "general type" and "very local subtype", are still classes (i.e. types of things, or abstract objects in sense of en:abstract and concrete). While distinct concrete objcets (described as "instances" in Wikidata) like chunks of material from distinct quarries or distinct sculptures can be described as being of rock (Q8063) class and of Lindabrunn Conglomerate (Q1825913) class. 2001:7D0:81F7:B580:E8A5:4F50:CF11:3311 18:05, 30 July 2020 (UTC)
@2001:7D0:81F7:B580:E8A5:4F50:CF11:3311, 2001:7D0:81F7:B580:E8A5:4F50:CF11:3311, Herzi Pinki, SM5POR: before starting to write what is the correct relations or not, please write in the same page the definition of all the items involved in the discussion. The main problem at the origin of such discussion is a lack of definitions or some misunderstanding about concept definition. Snipre (talk) 06:12, 31 July 2020 (UTC)

backtranslating in Q and P terms

Hi, it is quite tedious to write comments here in Q- and P-terms (items and properties). Sometimes it is more confusing than helpful, as when I'm writing in English here with uselang=de, I have no idea what it would look like in e.g. Cebuano. And whether terms do match exactly, nearly or not at all. That said, is there an automated way to backtranslate natural language to the typical wikidata slang using Q- and P-terms? Or at least get a proposal on mouseover? best --Herzi Pinki (talk) 16:01, 30 July 2020 (UTC)

It's not possible to go from natural language term to item/property ID because labels are not unique. I haven't seen any autocomplete in the standard editor. That said, you can jump to the search field and use it to search, then right-click on any of the suggestions to copy the linked URL. The search input also registers "f" as its access key. Check this table to find the eight buttons to press in addition to get that working.
If you want to show labels and/or description in a specific language to, the LD ("Label and Description") template works: error de software (problema en un programa informático o software que desencadena un resultado indeseado) should be in Spanish. The code for that is {{LD|Q179550|es}} And you may find some help among the gadgets, such as MediaWiki:Gadget-Descriptions.js which offers a preview-on-hover for linked items. Matthias Winkelmann (talk) 22:32, 30 July 2020 (UTC)
I agree that it can be very confusing to understand some discussions here at Wikidata. It often happens that labels does not fully match each other in different languages. When I look into some dicussions it often look like if you are arguing about if the Pope is catholic or if he is catholic. If he is shaved or has no beard. I often have to switch to English to fully understand exactly what you are arguing about. And I guess this often could be the reason behind some misunderstandings here. For example, in Swedish, we do not have different words for city and town. On the other hand, in English you have only one word for grandmother, while we have two. 62 etc (talk) 08:41, 31 July 2020 (UTC)

Wikidata list with wikipedia articles

Hello. User:Data Gamer/Names with WP I want to have one column with the en.wikipedia article (if exist) of each item. But I don't want a url. I want the form to be like Ioannis Mitroulas. Is that possible? Data Gamer play 21:07, 27 July 2020 (UTC)

Yes, but then you will probably have to make a template to use with the parameter row_template in {{Wikidata list}} as that is the only way that I know of to control how links are made in the template. --Dipsacus fullonum (talk) 07:06, 28 July 2020 (UTC)
Is there an example? Data Gamer play 14:51, 28 July 2020 (UTC)
There is a trick: create a wikilink directly in the query, bind a variable, have it returned from the query service and make a separate column for it.
You can lookup wikipages with ?article schema:about ?item; schema:isPartOf <https://en.wikipedia.org/>; schema:name ?title and do the trick with (CONCAT('[[', STR(?title), ']]') AS ?link). --Matěj Suchánek (talk) 13:52, 31 July 2020 (UTC)

Kven

Tracked in Phabricator
Task T259340

Moi!

From Wikipedia, I am trying to get the Wikidata label in Kven (Q165795), fkv. Test to see how {{#invoke:Sandlådan/Sextvåetc/Labels|what}} (from sv:Modul:Sandlådan/Sextvåetc/Labels) works in sv:Porsangers kommun for example. I can see in the api, that there is a Kven label in that page. But I cannot get it through a module on sv.Wikipedia. 62 etc (talk) 09:23, 29 July 2020 (UTC)

Hi Sextvåetc, to clarify, are you saying that the output you get from using the code you provided above on sv:Porsangers kommun doesn't include the sv language code? I tried it out and saw it in the output list ("sv sv Porsanger"), so perhaps you got it working or I misunderstood your question.  Hazard SJ  01:49, 30 July 2020 (UTC)
Hazard-SJ It is the Kven label (fkv), a minority language related to Finnish spoken in parts of Northern Norway I am looking for. I added that label yesterday to this item. My second guess was because there isn't support for languages without a project of its own. But that is not the case here. The South sami (sma) label in Røros Municipality (Q108999) could be found without problem. 62 etc (talk) 04:56, 30 July 2020 (UTC)
I now found a similair problem in Tysfjord Municipality (Q493851). The Lulesami (smj) label can not be found. 62 etc (talk) 05:04, 30 July 2020 (UTC)
Same problem can be identified in Douglas Adams (Q42). With LUA you can get 157 languages, and 161 by API. rwr, sje, smj and sms are missing in LUA. Where are the developers? 62 etc (talk) 18:35, 30 July 2020 (UTC)
Sextvåetc, thanks for clarifying, I saw what you mean from your first example (I haven't looked a the later ones). I'm not sure offhand why some languages aren't being included, but Wikidata:Contact the development team is a better page for reaching the developers.  Hazard SJ  23:48, 30 July 2020 (UTC)
When I look at Porsanger Municipality (Q483885) I see no label in "kv", and when I look at Tysfjord Municipality (Q493851) I see no label in "smj". I see statement values for official name (P1448) and native label (P1705) in these languages, but no labels. --Dipsacus fullonum (talk) 09:11, 31 July 2020 (UTC)
@Dipsacus fullonum: The lang-code for Kven is fkv, not kv. 62 etc (talk) 11:21, 31 July 2020 (UTC)
Tysfjord Municipality (Q493851) does have an smj label – you can see it in Special:EntityData/Q493851.json, and also on the entity page once you search for the value (Divtasvuodna) and discover that the displayed language name is julevsámegiella (the autonym, I assume). --Lucas Werkmeister (WMDE) (talk) 09:18, 31 July 2020 (UTC)
That said, in the Lua debug console, both =mw.wikibase.getLabelByLang( 'Q493851', 'smj' ) and =mw.wikibase.getEntity( 'Q493851' )['labels']['smj']['value'] print Divtasvuodna for me, so getting the label also seems to work… --Lucas Werkmeister (WMDE) (talk) 09:23, 31 July 2020 (UTC)
Yes, sorry I see it now. The labels are also available in the API: https://www.wikidata.org/w/api.php?action=query&prop=pageterms&titles=Q483885&uselang=fkv, https://www.wikidata.org/w/api.php?action=query&prop=pageterms&titles=Q493851&uselang=smj. --Dipsacus fullonum (talk) 09:26, 31 July 2020 (UTC)
@Sextvåetc: I think I figured out what’s wrong. Those language codes (and a few others) aren’t supported by MediaWiki by default, and so we have some special configuration to enable them – but only on Wikidata, not on Swedish Wikipedia. So on Wikidata, those language codes can be used, but on Swedish Wikipedia, the software thinks they’re invalid language codes and doesn’t return any data for them. I’m not sure what the best way to fix this is, but I created T259340 to track the issue. --Lucas Werkmeister (WMDE) (talk) 11:08, 31 July 2020 (UTC)

Help with translation to EN

Hi guys,

I'm running the WLE, and I'm facing translation issues, and as we do not have a Babylon, or at least I did not find, I'm here.

We have some names of conservative units that I do not know the sequence that sounds more native to you.

The list of issues:

  • Municipal Natural Park or Natural Municipal Park (pt: Parque Natural Municipal)
  • Private Reserve of Natural Heritage or Natural Heritage Private Reserve (pt: Reserva Particular do Patrimônio Natural)
  • Area of Relevant Ecological Interest or Relevant Ecological Interest Area (pt:Área de Relevante Interesse Ecológico)

I faced all this names, sometimes I even faced both names in the same article, so I am kind lost here. All the last options are for me more natural, but I am not native speaker... Municipal seems not even a common word.

Any light here? Thank you. Rodrigo Tetsuo Argenton (talk) 08:09, 31 July 2020 (UTC)

Well firstly, "Relevante" should be translated as "significant" not "relevant." The translation has to be "Area of Significant Ecological Interest." Secondly, Parque Natural Municipal definitely has to be "Municipal Natural Park" (or "Municipal Nature Park" would be even better I think), the other word order is wrong. And I find from the web that "Reserva Particular do Patrimônio Natural" should be "Private Natural Heritage Conservation Area," that's standard. — Levana Taylor (talk) 14:11, 31 July 2020 (UTC)
As a native English speaker (with roughly a typical Spanish-speaker's understanding of written Portuguese), I agree with all of Levana's wordings. - Jmabel (talk) 14:27, 31 July 2020 (UTC)
Grammar note: why does one of these use a word order with "of" and the others not? It's because of how the pieces of the phrase relate to each other in each case.
  1. [Private [Natural Heritage [Conservation [Area]]]]: There is one noun that is modified by a bunch of other adjectives and noun phrases. The adjective, which is more optional than the other modifiers, has to go on the outside. Think saying "what is that? it is a Natural Heritage Conservation Area, a private one, a public one, a potential one, etc."
  2. [Municipal [Nature/Natural [Park]]] -- exactly the same.
  3. In [Significant [Ecological [Interest]]] "significant" modifies "interest" not "area". That is, the phrase contains two nouns that are modified by other nouns or adjectives, and therefore I have to be careful to make it clear what is modifying what. If I wrote "Significant Ecological Interest Area" it wouldn't be entirely clear whether the structure was [[Significant [Ecological [Interest]]] Area] (right) or [Significant [Ecological Interest [Area]]] (wrong). But in "Area of Significant Ecological Interest" there is only one way it is possible to interpret it -- [Area of [Significant [Ecological [Interest]]]] -- so we prefer to say it that way. But wherever there isn't this chance of confusion, the standard basic English pattern is to string the modifiers together one after another in front of the noun; I chose that standard basic pattern for the first two examples. Mind you, "natural heritage" is also a phrase like "significant ecological interest" so you may be wondering why I didn't separate that out with an "of", to make sure no one could wrongly think it was "a heritage conservation area that is natural", [Natural [Heritage [Coservation Area]]]? The reason is that "natural heritage" is such a short and tightly-linked phrase that no one will ever make a mistake about it. You can, and should, treat short phrases like this as if they are a single adjective. Whew, complicated!!! [User:Levana Taylor|— Levana Taylor]] (talk) 14:11, 31 July 2020 (UTC)

software history

Please compare

  WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.   WikiProject Informatics has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. Visite fortuitement prolongée (talk) 14:33, 31 July 2020 (UTC)

One value of two

How is it shown that a property has one value or another but not both at the same time? Thanks. --Romulanus (talk) 15:13, 30 July 2020 (UTC)

This isn't making sense. The qualifier goes on a particular value. Right now, both values have

but that or any other qualifier can be on either value without being on the other. - Jmabel (talk) 00:59, 31 July 2020 (UTC)

Properties for railway rolling stock

Hi all, I've tried to added some locomotive and Diesel Multiple Unit classes to Wikidata (e.g. CP Class 5600 (Q856403) and Ghana Railway Diesel Multiple Unit (Q97777947)), but I've noticed, it's really hard, as many properties are missing or are unclear how they can be used on railway vehicles. I've created a subpage for that here: Wikidata:WikiProject Railways/Properties for rolling stock. Would be great to get some feedback from you! Please have a look! --Jcornelius (talk) 22:54, 31 July 2020 (UTC)

tree-lined

Hi, suppose that I have a pedestrian walkway, alley, avenue, promenade, road (sometimes also a parking lot or a square or a building) and I want to encode that it's "tree-lined" (in my language, "alberato"), how do I do that? It's quite a common description and sometimes they are considered scenic and part of the cultural heritage also because of this feature. So, how do we cover this information in The Avenue, Raglan Castle (Q24256091), Hamilton Terrace (Q47490961), Chaussee (Q41397963) or the next item I would like to create for Wiki Loves Monuments?--Alexmar983 (talk) 22:16, 25 July 2020 (UTC)

You could use avenue (Q7543083) for streets, not sure about parking lots etc. Ghouston (talk) 00:39, 26 July 2020 (UTC)
I don't want instances for a specific concept that is tree-lined, I want to add "tree lined" to different concepts. For example an avenue is in a urban area, you can have country road with trees and they are not avenues. Should I try a qualifier? I create an item for the concept? --Alexmar983 (talk) 03:49, 26 July 2020 (UTC)
avenue (Q7543083) is not necessarily urban, and most urban streets with "avenue" in their name are not examples of Q7543083. FWIW: "avenue" in the Q7543083 sense is probably not common U.S. English, I'm not sure we even have a word for it. Pretty common in the UK there, maybe the main meaning of avenue there. - Jmabel (talk) 16:02, 26 July 2020 (UTC)
There are many avenues in the UK that are tree-lined in the UK but there are also many other streets with different names that are tree-lined. There are also large numbers of avenues that aren't tree-lined. Whether the term historically had any connection to trees, I think it has lost that connection in British English. From Hill To Shore (talk) 18:20, 26 July 2020 (UTC)
@From Hill To Shore: So are you saying that UK English now does not use this sense of "avenue" any more than American English? If so, is there any word for this in contemporary UK English? Avenue (landscape) doesn't really have much to say about contemporary usage of the word, other than just its use in "the usual suite of words used in street names". - Jmabel (talk) 21:29, 26 July 2020 (UTC)
@Jmabel: People in the UK may use that sense of the word in some contexts, but if you are wanting a word that will be instantly recognised as "tree-lined" then Avenue isn't the one from a UK perspective (I know several avenues in my area that don't have trees). The text at the top of c:Category:Avenues in England perhaps puts it more clearly; "Avenues - the garden and landscape architecture feature... This is not the category for streets or roads which include the name 'Avenue', but for deliberately planted parallel rows of trees, hedges or other flora, unless the named 'Avenue' photo is focused on a regularly spaced twin line of trees." I can't think of a specific phrase that we would use in the UK other than, "tree-lined." It is possible that there is a word that fits, but I can't think of it. From Hill To Shore (talk) 23:00, 26 July 2020 (UTC)
@From Hill To Shore: would you agree, though, that avenue (Q7543083) is certainly this sense of the term? - Jmabel (talk) 01:10, 27 July 2020 (UTC)
It seems to be how it was intended, but items like Sixth Avenue (Q109873) seem to be instances based on name alone. Ghouston (talk) 04:48, 27 July 2020 (UTC)

All this discussion simply proves besides my original comment that you need to encode "tree-lined" as a separate concept. So how do we do it?--Alexmar983 (talk) 00:32, 29 July 2020 (UTC)

So no clue? I have two more items of tree-lined alley to create, I put them in a personal list and i am ready for whatever solution you want ot propose.--Alexmar983 (talk) 16:32, 6 August 2020 (UTC)

Thanks in any case.--Alexmar983 (talk) 00:13, 11 August 2020 (UTC)
P2670 seems to be the way to go. --- Jura 00:29, 11 August 2020 (UTC)
Should I create an item to be its argument?--Alexmar983 (talk) 22:43, 11 August 2020 (UTC)

IF you know how to use properly OpenStreetMap tag or key (P1282) this is the tag on OSM.--Alexmar983 (talk) 00:44, 12 August 2020 (UTC)

Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2020/07&oldid=1260202645"