Shortcuts: WD:PC, WD:CHAT, WD:?

Wikidata:Project chat

From Wikidata
Jump to navigation Jump to search

AGB, Allgemeine Geschäftsbedingungen (Q34582) wouldn't allow adding T&Cs for English language

[edit]

The item seems to be blocked by a rarely populated term (merely three languages in wikidata Q5165703) the T&Cs redirects to in English Wikipedia. This causes no language link for English language showing up in German Wikipedia lemma Allgemeine Geschäftsbedingungen at all, and also Allgemeine_Geschäftsbedingungen not showing up as Language link in English Wikipedia entry T&Cs. Seem I can't fix that, Wikidata wouldn't let me add the English Wikipedia lemma T&Cs to Q34582. Not sure whether the English Wikipedia redirection from T&Cs to the more general Contractual term is proper at all, causing the trouble in the first place here. Please could somebody more experienced fix this issue, so the T&Cs English Language link shows up in German Wikipedia Lemma Allgemeine_Geschäftsbedingungen? --92.117.131.30 00:31, 25 November 2024 (UTC)[reply]

Sitelinks to redirects can only be added via the sitelink for redirect badges. Those in turn can only be added by autoconfirmed users. If you want to contribute to Wikidata in such a way, it makes sense to register an account. ChristianKl01:19, 2 December 2024 (UTC)[reply]

"Modeling assumption" being a subclass of "lemma" is nonsensical and give rise to surreal inclusion chains

[edit]

In the last edition of The Open Wikipedia Ranking we noticed some very weird results. We rank Wikipedia pages using standard centralities measures, and then let the user select subsets corresponding to Wikidata. As we explain in the FAQ, since 2024 the best theorem on Wikipedia is "Conspiracy Theory", which makes no sense. The problem lies in the subclass relationship between "modeling assumption" and "lemma", which causes absurd chains such as foreign policy ⊆ public policy ⊆ policy ⊆ principle ⊆ axiom ⊆ lemma ⊆ theorem, or conspiracy theory ⊆ reductionism ⊆ simplism ⊆ modeling assumption ⊆ lemma ⊆ theorem.

We tried to deprecate the faulty subclass relationship, to no avail. It should be removed, as it is simply not true. Lemma and theorem in mathematics are interchangeable, but a modeling assumption is certainly not a lemma. And subclass causes a semantic collapse of a very large number of concepts into being transitive subclasses of “theorem”. Sebastiano Vigna 18:16, 27 November 2024 (UTC)[reply]

I was actually able to remove the subclass. Let's see if it comes back... Vigna (talk) 12:09, 28 November 2024 (UTC)[reply]
@Vigna: If you have any further issues of this sort I'd suggest raising it with Wikidata:WikiProject Ontology which deals with anomalies in the subclass tree. ArthurPSmith (talk) 12:57, 28 November 2024 (UTC)[reply]

Three is a crowd ... so how to fix?

[edit]

I would like to link the page Foo and the category Foo on some project to the the category Foo on Commons. That is ofcourse not possible. When I can't have both I need to chose. But who to pick? Is there a good practice for match making?

The relevant pages are Category:Printedition (Q33697781) and Wikinews:Print edition (Q33124259). I could chose to link the category to c:Category:Wikinews Print Edition - English because the files are in English and do the same for the Persian category.

I had this problem many times before and I always gave up and thought someone else would probably fix it but it would be nice to know how to solve such matching problems. MGA73 (talk) 11:50, 28 November 2024 (UTC)[reply]

Don't know which data type to take with proposal there are several possibilities that fit

[edit]

I am preparing a proposal for Audio tour

But have the problem that multiple data types can be used for this property proposal, so don't know how I should handle this.

Sometimes the audio tour can only be listened to on site, e.g. the MSKGent where there are several audio tours included in the museum ticket.
Another time one uses a site with multiple podcasts describing a city and where you can download them as mp3 or listen to them live via URL see e.g. vienna
In addition, you have museums that use an app for an audio tour

So then you need something that can indicate whether you can only listen to the audio tour physically on site, whether you can download it e.g. as an audio file,
further whether there is a URL to a podcast or the like and lastly whether you need to use an app to listen to the audio tour.

Can anyone tell me how best to approach this? Thanks in advance. Brechtd (talk) 17:31, 28 November 2024 (UTC)[reply]

I think you needed to have created one proposal, not three, properly fleshed it out, and invite discussion there, rather than come back here. Once your proposal is ready we'd be notified of it by the weekly message system. Vicarage (talk) 22:55, 28 November 2024 (UTC)[reply]

Merging or deletion needed

[edit]

I created Krazy Kat Goes A-Wooing (Q131343177) without realising it already existed under Krazy Kat Goes A-Wooing (Q124536127). I ported over all the data, so Krazy Kat Goes A-Wooing (Q131343177) should be kept and Krazy Kat Goes A-Wooing (Q124536127) should be removed. Can anyone help? Edo999 (talk) 10:05, 29 November 2024 (UTC)[reply]

→ ← Merged RVA2869 (talk) 10:25, 29 November 2024 (UTC)[reply]
That was quick. Thank you! Edo999 (talk) 10:26, 29 November 2024 (UTC)[reply]

how to describe new objects 'studio equipment', 'lighting equipment', 'camera accessory'

[edit]

For an archive system we need objects that describe what technical equipment was used to produce a film. Camera, lens, tripod, etc. already exist and can be used perfectly. But then there are a lot of smaller pieces of equipment that need to be grouped together. We are thinking of generic items such as 'studio equipment', 'lighting equipment', 'camera accessory', etc.

I assume I need to create these objects. What properties should I use? As a subclass of 'equipment' Q10273457 ? Any other ideas? Thanks! Graefestrasse (talk) 11:14, 29 November 2024 (UTC)[reply]

My first thought would be uses (P2283). Sjoerd de Bruin (talk) 21:57, 29 November 2024 (UTC)[reply]
There's also captured with (P4082) for describing which equipment was used to create a work, although that's best used for cameras, microphones and the like.
But I think @Graefestrasse's question is more about how to create/model new objects for generic groups of things that are used in the same context (but not necessarily for the same immediate purpose)? woodworking tool (Q11789812) kind of goes into a similar direction, maybe that could help a bit? El Grafo (talk) 15:15, 1 December 2024 (UTC)[reply]
Yes, Q11789812 sounds good. I guess I could create i.e. "studio equipment" and say it's an instance of type of tool as well as Q10273457 (equipment). Graefestrasse (talk) 10:19, 2 December 2024 (UTC)[reply]

Ethical Considerations: home addresses

[edit]

When describing a historic home that is still in use today what are the Wikidata guidelines for including a home address? H.Crummey (talk) 16:27, 29 November 2024 (UTC)[reply]

I'm not sure we have a policy, but for listed buildings in the UK at least their address is part of the public listing information, and so could be considered to be acceptable inside WD. For houses that are not architecturally notable, but historically so, as once the residences of famous people, again the public Blue Plaques (famous person lived in this London house) have addresses, suggests that addresses are not private information. I'd not include celebrity trivia, but if a building its considered notable enough to feature on WD, its address can be recorded. Vicarage (talk) 20:09, 29 November 2024 (UTC)[reply]
Okay that is a helpful clarification. Thank you. H.Crummey (talk) 23:56, 29 November 2024 (UTC)[reply]
When it comes to matching a person to the home there's residence (P551)living people protection class (P8274)property that may violate privacy (Q44601380), so according to LP: "Living individuals with records in Wikidata are for the most part not famous or celebrities; their privacy should be respected. Values for living individuals should generally not be supplied unless they can be considered widespread public knowledge or are openly supplied by the individual themselves (otherwise hidden supporting references are not sufficient). As an example, the fact that someone's address is accessible by looking at a domain name registration doesn't imply that it's considered widespread public knowledge for the sake of this policy" ChristianKl11:49, 2 December 2024 (UTC)[reply]

Sources for DOB

[edit]

I'm a Wikipedia editor and done very little directly in Wikidata, but I have noticed this a couple of times, when trying to check on things or find a source, and wonder if it has been discussed before. Over on WP, the rules about stating and citing dates of birth for living people (see en:WP:DOB) are very stringent, and specifically exclude IMDb as a reliable source. I have noticed over here that the references for DOB often don't live up to the same standards, for example see Antoinette Byron Q5697422. Refs are Spanish Wikipedia (which does not cite the DOB at all) and IMDb. (I have been trying to untangle Antoinette from Annie Byron, which had become partly confused over there.) This seems to be a problem if editors are picking up info from Wikidata that is not reliable. Thoughts, anyone? Laterthanyouthink (talk) 07:31, 30 November 2024 (UTC)[reply]

Wikidata is still basically the Wild West. There are scant policies and guidelines, lax enforcement/oversight of these, and most people seem to operate on the wild west model of "anything goes until someone shoots you and steals your horse, and then things get done differently for a while." Consensus, if found anywhere, tends to arise in the bowels of talk pages, merely to be quickly archived and rarely seen again. Also, much like like Wikipedia, Wikidata cares not one iota about truthfulness, but verifiable data for the sake of data, right or wrong. You might find some items have a half dozen different birth dates, each sourced to references of various levels of credibility and reliability. Nobody in their right mind should ever assume that the info on Wikidata is necessarily true, accurate, complete, or up to date. -Animalparty (talk) 03:28, 1 December 2024 (UTC)[reply]
I think that's partly the result of the scale of the task, and the paucity of volunteers. Personally I avoid working on people, but when I'm sorting out naval history, I want to get every ship with a smattering of information in 2-3 years, not spending a lifetime crafting one ship a day. I expect others, especially bots, will come along and add references to the facts I copy from wikipedias, based on the 3rd party links I also add.
12 years into the project, and still with gaping holes in coverage, I think comprehensiveness trumps accuracy, especially as machines can check and document the latter given the former. Vicarage (talk) 06:04, 1 December 2024 (UTC)[reply]
At the beginning, copying from Wikipedia was the easiest and so is copying from IMDb (which is already well matched to Wikipedia pages). Chances are these are just leftovers from the early days and the item has already been matched to a better source. Or the item is still waiting for a good match. Or it's a niche personality who is notable on at most one wiki and nobody knows where their DOB was taken from.
As for the reliability:
  • Collaborative. Data is entered and maintained by Wikidata editors, who decide on the rules of content creation and management. Automated bots also enter data into Wikidata.
  • A secondary knowledge base. Wikidata records not just statements, but also their sources, and connections to other databases. This reflects the diversity of knowledge available and supports the notion of verifiability. (Wikidata:Introduction)
If a statement references a source you don't trust, you don't have to trust that either. If a source supports a statement, add it with the source. --Matěj Suchánek (talk) 17:57, 2 December 2024 (UTC)[reply]

New property proposal - Bing Maps ID

[edit]

I'm proposing a new Wikidata property for Bing Maps ID. This property would allow linking Wikidata items to specific Points of Interest (POIs) on Bing Maps. Please review the proposal and share your feedback. Your input is valuable! David Osipov (talk) 09:57, 30 November 2024 (UTC)[reply]

 Info thank you, Lewis Hulbert , for stressing out that the ORSIDs are temporary IDs of Bing maps POIs. Please, ignore this proposal for now - I'll try to find a permanent ID of Bing maps. --David Osipov (talk) 10:53, 30 November 2024 (UTC)[reply]

Semantic Scholar Id

[edit]

How can I get Semantic Scholar Id Seemadarade (talk) 13:57, 30 November 2024 (UTC)[reply]

@Seemadarade: If you mean to obtain one for yourself, you'd need to go to their website and ask them. If you mean something else, please clarify your question. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:04, 1 December 2024 (UTC)[reply]

Allowed-entity-types constraint

[edit]

Hi everyone, I am new-ish to wikidata and describing commons uploads on wikidata. What is the allowed-entity-types constraint? For example on File:2022 Australian federal election Wright alluvial diagram.svg and commons:File:World War II Allies to Axis GDP Ratio.svg.

On the alluvial diagram pages (one for every seat) I have office contested (P541): member of the Australian House of Representatives (Q18912794), with qualifier electoral district (P768): Wright (Q2594455) (or whatever seat it is)

On the allies to axis chart, I have category combines topics (P971): World War II (Q362) & economics (Q8134)

Both of these say allowed-entity-types constraint, saying "The property should not be used on this type of entity, the only valid entity type is Wikibase item.". Can someone explain? I have tried reading the help pages, although I don't really understand still. Is it because they are images not items? How would I otherwise describe these in a structured data way, or should I just be not going that deep with it? MarkiPoli (talk) 15:02, 30 November 2024 (UTC)[reply]

@MarkiPoli: I don't understand why you're getting that error, but maybe somebody with more experience with the Commons structured data should comment. But to answer the question on entity types - it's basically within Wikidata determined by the prefix letter. Qxxx is a Wikibase item, Pxxx is a Wikibase property, Lxxx is a lexeme, etc. ArthurPSmith (talk) 16:00, 2 December 2024 (UTC)[reply]
Properties have statements indicating where and how they are intended to be used. Properties like creator (P170) or copyright status (P6216) have the statements property constraint (P2302)allowed-entity-types constraint (Q52004125) with the values Wikibase item (Q29934200) (use on Wikidata items) and Wikibase MediaInfo (Q59712033) (use in Structured Data on Commons). The properties you used on those files, on the other hand, weren't created or intended to be used on Commons to model the content of files. So they don't have the Wikibase MediaInfo (Q59712033) value and throw an error when used there.
category combines topics (P971) for example is intended for use on Wikidata items about categories, not on files. You'd probably have to ask over on Commons (commons:Commons talk:Structured data) about how they want to model this kind of file content. It's indeed possible that the Commons community didn't think about modelling the contents in such a detailed way. --2A02:810B:581:C300:FD46:81F6:126F:B70E 07:19, 4 December 2024 (UTC)[reply]

Merge of Q74524855 and Q71538638

[edit]

Recently, User:Daask merged most frequent value (Q74524855) (before: "most frequent value") and most frequent value (Q71538638) (before: "generally used form"). I wonder if that was right since Q7452455 was defined as being a facet of (P1269) point estimation (Q1192065), while Q71538638 was a facet of (P1269) name (Q82799). What do you think? Dorades (talk) 16:48, 30 November 2024 (UTC)[reply]

I don't know if point estimation (Q1192065) should have been there, but Q71538638 was originally only about forms of a name, not for other values, although it was being used in other ways, in some cases where Q74524855 would have been better. Peter James (talk) 20:48, 30 November 2024 (UTC)[reply]
I feel like both have the same meaning as Wikibase reason for preferred rank (Q71533077) and can't think of an example where it's worth distinguishing the two concepts. ChristianKl15:50, 4 December 2024 (UTC)[reply]

Do we have this Wikidata:Property proposal/role named as but for movies?

[edit]

I was recently notified about Wikidata:Property proposal/role named as which is about video game credits. Do we already do something like this for movie cast in how their roles appear in the movie credits? SuperUltraHardCoreGamer (talk) 10:21, 1 December 2024 (UTC)[reply]

Date of Women and Men vote suffrage

[edit]

We can probably use Wikidata to store the date when women+men suffrage was introduced by each country.

How to do that? Where to discuss this?

At the moment I'm proposing something like this.

To say Australia=1902 and Switzerland=1990 we could do:

Thanks for your opinion dear Wikidata hackers --Valerio Bozzolan (talk) 13:28, 1 December 2024 (UTC)[reply]

P.S. thanks to "Nikki" for the above tip from Wikidata Telegram chat https://t.me/c/1224298920/137467 Valerio Bozzolan (talk) 13:30, 1 December 2024 (UTC)[reply]
There are a lot of event that you could possibly store on the items of countries, but Wikidata is not doing well with having thousands of statements. I would rather have an item about the introduction in a given country and give that item the relevant statements. ChristianKl01:13, 2 December 2024 (UTC)[reply]
Introducing 200 new items seems less efficient than 200 statements. But it might be better adding those statements to introduction of universal suffrage for men and women (Q131367891) with country (P17) qualifiers than start a trend for adding events to countries, which could be huge and make the item unreadable Vicarage (talk) 01:30, 2 December 2024 (UTC)[reply]
@Vicarage: What property do you suggest? --Valerio Bozzolan (talk) 06:59, 2 December 2024 (UTC)[reply]
introduction of universal suffrage for men and women (Q131367891) point in time (P585) 1902 applies to jurisdiction (P1001) Australia (Q408) as its a legal change Vicarage (talk) 07:53, 2 December 2024 (UTC)[reply]
@Vicarage 200 statements to a single item is certainly better than adding it to individual countries.
It's worth noting that universal suffrage for men and women means "all man in women" and many countries. That does include prisoners and many countries do not allow them to vote, Ireland for example only allowed that in 2006. The US does not have it.
If you want to focus on the date where women where allowed to vote, women's suffrage (Q205204) is the better item. The personhood of prisoners is worth respecting. ChristianKl11:42, 2 December 2024 (UTC)[reply]
Its certainly a can of worms. https://commonslibrary.parliament.uk/research-briefings/cbp-7461/ show that even as late as 2018 England (but not Scotland)
were changing the rules. But better to record the broad brush details than not at all. Devolution to Scotland shows why applies to jurisdiction (P1001) is better than country (P17) Vicarage (talk) 11:59, 2 December 2024 (UTC)[reply]

Wikidata weekly summary #656

[edit]

Rotten Tomatoes and Wikidata

[edit]

When filling in the "Critic's response" in a film's en.wikipedia article, I update the Wikidata for the film, then use the following in the WP article: {{RT prose|{{RT data|score}}|{{RT data|average}}|{{RT data|count}}|prose prose prose prose|ref=yes|access-date={{RT data|date}}}}

Short and sweet, very very convenient, automatically updates the WP article if some user later updates the Wikidata fields.

A challenge is that, if the information for the Rotten Tomatoes average rating has "/10" included in the Wikidata field, it results in a double "/10/10" appearing in the WP article. For example, for the 2021 film Oxygen https://en.wikipedia.org/wiki/Oxygen_(2021_film), putting 6.8/10 in the Wikidata field results in the en.wikipedia article displaying:

... of 106 critics' reviews are positive, with an average rating of 6.8/10/10.

The only way to avoid this is to not include the "/10" when filling in the Wikidata field, resulting in the en.wikipedia article properly showing:

... of 106 critics' reviews are positive, with an average rating of 6.8/10.

While this "workaround" resulted in a nice output for en.wikipedia articles, it has been brought to my attention, by de.wikipedia editor Eiragorn, that leaving out some of the extra detail in the cross-language shared Wikidata is resulting in issues for the templates used in the WP articles in other languages. The "workaround" to make the en.wikipedia template work is hurting other ill.wikipedia templates.

I should point out that, until very recently, this "double display" issue would also happen if the Tomatometer score, in the Wikidata file, had the "%" symbol included, resulting in a double "%%" appearing in the en.wikipedia article. For the Oxygen example, this produced:

On the review aggregator website Rotten Tomatoes, 90%% of 106 critics' reviews...

It appears this has been fixed in the en.wikipedia template. In fact, it is "beyond fixed", as you can now do a Wikidata input with the digits by themselves (90), or the digits and the symbol (90%), and the result in always a single "%" appearing in the en.wikipedia article. For the Oxygen example, this now always produces:

On the review aggregator website Rotten Tomatoes, 90% of 106 critics' reviews...

I checked with an en.wikipedia administrator today, who recommended discussing this with you good folks at Wikidata first, and asking here what the proper format for recording the information is in Wikidata. They pointed out that Wikidata might have some different way of recording the /10 denominator and that it was best to check if a user should - or shouldn't - include it in the Wikidata field. It was mentioned that you folks would also have the ability to introduce constraint violations or make mass edits to enforce the proper format. Once that's done, each of the Wikipedias would then work with that as a baseline.

If I have placed this discussion item on the wrong Wikidata page, please let me know and I will redo at the proper spot. Thank you, in advance, for your help and direction. Jmg38 (talk) 01:31, 3 December 2024 (UTC)[reply]

review score (P444) usage seems to be X/Y for Rotten Tomatoes, and there is no constraint applied on the input to limit the characters used, so I think the onus is on WP to accept that as is. Clearly you could have a special WD review property "TripAdvisor stars", forced to 1-5, and WP formats that to add a '*', but that's not the case here, and I doubt we'd want lots of separate properties. Vicarage (talk) 10:31, 3 December 2024 (UTC)[reply]
@Jmg38: Doesn't this work better: {{RT data|prose|consensus=prose prose prose prose|ref=yes|access-date={{RT data|date}}}} - Difool (talk) 13:27, 3 December 2024 (UTC)[reply]
Thank you, Vicarage and Difool. And a thank you to Indagate, who also pointed out that the long clunky nested templates I came across years ago were unnecessary, with just one minor reduction from the already significant improvement that Difool included above: {{RT data|prose|consensus=text text text…|ref=yes}} Overall, another terrific outcome from the wiki community. Thank you all. Jmg38 (talk) 19:01, 3 December 2024 (UTC)[reply]
The examples of review score (P444) all are worded like X/Y. Wikipedia templates should be able to handle that. ChristianKl16:04, 4 December 2024 (UTC)[reply]

'Founder of' property or similar

[edit]

Hello everyone! I am scratching my head and can't come up with a reasonable alternative to 'founder of', which doesn't exist as a property presently. 'Founded by' already exists, so I am looking for its opposite.

Before I propose a new property, I wondered what alternatives there are? eg. Octavia Hill would be 'founder of' the National Trust; Alexander Keiller would be 'founder of' the Morven Institute of Historical research; Wallace Heaton would be 'founder of Wallace Heaton Ltd; Ivison Macadam would be 'founder of' the National Union of Students...

With thanks! Medievalfran (talk) 10:14, 3 December 2024 (UTC)[reply]

We generally don't record symmetrical properties, as the query system makes it easy to query ?founded wdt:P112 ?founder to find the relation both ways. It keeps the number of statements down and reduces clutter. Vicarage (talk) 10:22, 3 December 2024 (UTC)[reply]
This is previously proposed six times: 2013, 2016, 2016, 2017, 2023, 2024.--GZWDer (talk) 15:25, 3 December 2024 (UTC)[reply]

P31 = numeric identifier on Wikidata properties

[edit]

Hello all friends of Wikidata properties! See my proposal regarding moving numeric identifier (Q93868746) from instance of (P31) to has characteristic (P1552) on Wikidata talk:WikiProject Properties. Thanks. Samoasambia 22:03, 3 December 2024 (UTC)[reply]

Adding (Ancient) Egyptian lexemes from Thesaurus Linguae Aegyptiae

[edit]

The Thesaurus Linguae Aegyptiae (Q122748326) project seeks to donate about 15T of approx 40T lexical entries of the Egyptian (Q50868) language in Leiden Unified Transliteration (Q131362896) and Egyptian hieroglyphs (Q132659) to Wikidata. This is part of a cooperation project funded by the consortium Text+ (Q98271443) of the German National Research Data Infrastructure (Q61658497).

In a first batch, we are focusing on the main parts of speech, largely excluding proper noun (Q147276), title (Q216353), and epithet (Q207869). Examples: ḥꜣ.tï/𓄂𓏏𓏭𓄣 (L1391335), sw/𓇓𓅱 (L1385334), ẖr.ï/𓌨𓂋𓏭 (L1391360), ky/𓎡𓇋𓇋 (L1391327) wḏꜣ/𓅱𓍑𓄿𓂻 (L1385371); selected frequent proper nouns and titles, e.g. Wsꞽr/𓊨𓁹 (L1391321), Ppy (L1391324), nb-Tꜣ.wï/𓎟𓇾𓇾 (L1391331), ẖr.ï-ḥꜣb.t/𓎛𓌨𓃀 (L1391337); see also: https://situx.github.io/paleordia/language/?q=Q50868&qLabel=Egyptian, https://ordia.toolforge.org/language/Q50868.

We also created a language modeling page for Egyptian.

We are currently requesting a bot flag for our ThesaurusLinguaeAegyptiaeBot (talkcontribslogs), so that we can perform maintenance tasks more efficiently in the future: Wikidata:Requests_for_permissions/Bot.

Any comment and support is very welcome.

Dwer (talk) 13:04, 4 December 2024 (UTC)[reply]

[edit]

Are Wiktionary sitelinks in items for Unicodes allowed? Since they are inherently translingual/mul, would a sitelink to wikt:℥ in (Q87523885), along the other interwikts, be fine? LIrala (talk) 03:38, 5 December 2024 (UTC)[reply]