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

Wikidata:Project chat

From Wikidata
Jump to navigation Jump to search

Wikidata project chat
A place to discuss any and all aspects of Wikidata: the project itself, policy and proposals, individual data items, technical issues, etc.

Please use {{Q}} or {{P}} the first time you mention an item or property, respectively.
Other places to find help

For realtime chat rooms about Wikidata, see Wikidata:IRC.
On this page, old discussions are archived after 7 days. An overview of all archives can be found at this page's archive index. The current archive is located at 2023/09.

Is there a way to get all STEM pages on wikipedia?[edit]

I've been going through the categorylinks table a fair amount, and walking the tree/graph you end up pulling in almost all of wikipedia when you start from science as your root category and recursively pull in subcats/pages. Someone pointed me to wikidata as a way perhaps to solve this problem, but I'm unconvinced as building a proper categorization of something like STEM pages I think would require a very specific effort. Wikiqrdl (talk) 19:04, 10 September 2023 (UTC)Reply[reply]

One idea I proposed was 'category views'. Sort of like portals, where an individual or a group takes ownership of a category view and they control what gets included in it. It could use only existing tags / categories, but the connections between them and to pages would be controlled by the view. My thoughts are that it would lie generally outside the main flow of things, similar to wikidata. Importantly, the category views would be 'purpose built' by domain experts who understand how something might be useful to similar practitioners. Some sort of category view rating system might be useful, perhaps along different dimensions. Wikiqrdl (talk) 19:50, 10 September 2023 (UTC)Reply[reply]
As an example for the above, let's say a professor of education wanted to create a high quality list of all the pages for STEM on wikipedia. First they'd categorize it and then then they'd include the pages they felt met their bar in the categories. Something like this -> but more complete and with a very well thought out hierarchical arrangement that would serve someone going into STEM field of study might be find credible and useful. Importantly, it wouldn't get sidetracked by things like botany->plants->coats of arms with plants (as an example) Wikiqrdl (talk) 19:54, 10 September 2023 (UTC)Reply[reply]
Wikidata itself is not organized into categories or tags. ChristianKl❫ 23:09, 11 September 2023 (UTC)Reply[reply]
@Wikiqrdl: If you only care about the English Wikipedia, use See w:WP:PROJDIR/S for a full list of science WikiProjects. Dexxor (talk) 10:32, 12 September 2023 (UTC)Reply[reply]
Very cool, the wikiprojects/portals are exactly what I was looking for. Is there a database dump of the projects I can import? I'm trying to create a compilation of all the stem pages. I have all the other dumps imported. Can I traverse the wikiprojects and their subcats/subpages via the categorylinks table? Wikiqrdl (talk) 07:14, 18 September 2023 (UTC)Reply[reply]
Thanks, that's what I thought. Wikiqrdl (talk) 07:12, 18 September 2023 (UTC)Reply[reply]
fwiw, I tried to create a sparql to solve the problem, but it looks like the tagging just isn't complete. eg: instance of science, part of education. 8 results? Wikiproject / portal folks could help here. Wikiqrdl (talk) 02:29, 25 September 2023 (UTC)Reply[reply]
@ChristianKl @Dexxor One idea that came to me is that things like wikiprojects and portals could be organized around sparql queries. This would help get domain expertise to help inform the wikidata project. Eg, say a portal around botany was created, they would topline the query they used to capture the articles of interest. They would work with the wikidata project to make sure their wikidata was expressive enough and that their query was not overly broad / in appropriate. Wikiqrdl (talk) I proposed this here - 1:54, 25 September 2023 (UTC)

there is no subtopic for data flaws / wrong data in the large[edit]

In the heading {{Discussion navigation}} there is a lot of subtopics, but astonishingly no subtopic for systematic data flaws or wrong data in the large. Is there a place where kind of work packages can be dropped (for those interested in the correctness of the data) or is it intended to have the triage: love the data, correct the data, or leave wrong data as they are? best --Herzi Pinki (talk) 08:16, 17 September 2023 (UTC)Reply[reply]

Kannst du das bitte nochmal auf Deutsch schreiben? Ich (Englisch C1) verstehe nur Bahnhof. Welches heading? Was meinst du mit subtopics? Was ist "love the data"? Jonathan Groß (talk) 08:21, 17 September 2023 (UTC)Reply[reply]

Sure, I can, but this is the English project chat. Sorry for my English. Ich suche einen Platz, wo ich Arbeitspakete abladen kann und wo jemand sich ev. auch drum kümmert. Triage ist ein doppeltes Wortspiel / Anspielung, hier auf die Daten in Wikidata angewandt. Triage im militärischen Sinn heißt, die die eh sterben, sterben zu lassen. Triage hier auf Wikidata hieße, die sinnlosen Daten, die falschen Daten, die inkonsistenten Daten einfach liegen zu lassen. Auch beim Militär sterben ja nicht alle, die dafür aussortiert werden. Der zweite Teil der Anspielung bezieht sich auf Love it, change it or leave it (Liebe es, ändere es oder geh). Was dann heißen soll: Wunderbare Daten, alles ok - ich liebe die Daten; fehlerhafte Daten, widersprüchliche Daten, unvollständige Daten - dann krempel ich die Ärmel hoch und ändere sie halt; oder ich geh (es gibt unfeinere Formulierungen dafür) bzw. lasse sie wie sie sind, schau nicht hin, .... lg --Herzi Pinki (talk) 08:38, 17 September 2023 (UTC)Reply[reply]

Alright, then I would like to know what sort of data (subject matter, format) you are referring to and how you imagine it could be integrated with Wikidata. In principle the community can help with curating data donations of this sort, but we need to know what we would be dealing with more specifically. Jonathan Groß (talk) 09:10, 17 September 2023 (UTC)Reply[reply]

Or maybe I again misunderstood, and by "Arbeitspakete" you mean tasks that need to be done on existing data in Wikidata? We have Wikidata:Bot requests, but that is mostly inactive. Why don't you just use this venue and list your jobs here? Jonathan Groß (talk) 09:21, 17 September 2023 (UTC)Reply[reply]

see below --Herzi Pinki (talk) 09:26, 17 September 2023 (UTC)Reply[reply]

Most people on Wikidata have certain data they are interested on and work on that data. There are few people on Wikidata who want to do large packages of work that someone else believes should be done. One place where you can discuss data problems is in the relevant Wikiprojects. ChristianKl❫ 14:47, 21 September 2023 (UTC)Reply[reply]

@ChristianKl: thanks for the hint, got your point. My problem is the other way round. What if people are not really interested in the data they have imported from somewhere else? What if they do not care and curate and complete the data? Am I one of those few people on Wikidata who want to do large packages of work that someone else caused and nobody is eager to fix? Is quality a common goal of the community here on Wikidata? Or is it just number of items? As an example see the mountains in the state of Washington assigned to country Canada (Q16) (just below). The assignment of Canada (Q16) has been done by User:&beer&love, not active any more. I just stumbled across. Sure, I can try to fix it and not bother someone else. And this is only Canada in the US. What about Kenia in Tansania, or Chile in Argentina? I also stumbled across data imported from peakware (discontinued for more than 2 years now) with a lot of duplicates. I asked the importing user to fix this or to help me to fix it. He gave me the advice that I will waste my time. I'm working on Austrian geographical objects imported via geonames / Ljsbot / WP:ceb / EmausBot for years now: Existence is not always clear, spelling often is incorrect, location is cut to minutes mostly, elevation is derived from this inaccurate location info and there are a lot of duplicates (and I did not dare to have a look across the Austrian borders). I'm also working on the protected monuments of Austria for just a while, those have been imported by some bot in 2017 using wrong heuristics. There was no will (=no reaction to my complaints) from the bot operator to care for the necessary fixes. These are only some examples. Thanks for listening. --Herzi Pinki (talk) 11:18, 22 September 2023 (UTC)Reply[reply]
Different people have different goals. Generally, finding some place where data quality can be improved on Wikidata isn't that hard. The bottleneck is not in having people point to problems.
People who do care about data usually care about a certain subset of the data. When it comes to data quality of mountains, that should probably be talked about in some Wikiproject about geography. ChristianKl❫ 00:24, 23 September 2023 (UTC)Reply[reply]
ok. I did not find a suitable wikiproject here on wikidata (only items linking to various projects in various languages in various wikipedias), can you give me an example here in wikidata? The other point is how do we get a feeling about the overall quality of wikidata in the large, if we not even do collect flaws? Is quality a goal of Wikidata? (different people have different goals)? How is quality organized and created? best --Herzi Pinki (talk) 11:12, 24 September 2023 (UTC)Reply[reply]
Also see
M2k~dewiki (talk) 12:31, 24 September 2023 (UTC)Reply[reply]

fallback language behavior?[edit]

Hi, is there a definition somewhere how language fallback will work on wikidata and adjacent wikipedias? Consider language pairs (en & en-ca, en-br; de & de-at; pt & pt-br etc.) I often find copies of labels / descriptions / aliases here in wikidata, which are just copies (redundant copies) from the main language to the variant language. In worse cases those labels / descriptions are meant as copies, but have been modified on one side only and give different values, not because of the underlying language, but due to edits not done on all ends. My assumption is that if there is no label for en-ca, the label for en will be taken automatically (eventually depending on some user config). This is the fallback from en-ca to en. Does my assumption hold true for

  • page title of items
  • page titles and descriptions shown in auto-completion when typing ahead
  • page titles and descriptions shown in search results
  • page titles and descriptions in Sparql results
  • labels and descriptions shown on wikipedias?

Concluding: Is there any need to have labels, descriptions & aliases for en-ca, if they are all identical to en? best --Herzi Pinki (talk) 11:31, 22 September 2023 (UTC)Reply[reply]

We've had discussions of duplicated labels before, see Help_talk:Mul_language_code for how a 'mull' language code might help. Most people believe there is no need to duplicate a value if a more general language class has it, or if the label in the home country language of an item would apply in equally well in other languages. [[ describes how a query can return values in a preference list of languages. Vicarage (talk) 12:49, 22 September 2023 (UTC)Reply[reply]
my question was not about discussions and not about beliefs, but about specification and implementation. best --Herzi Pinki (talk) 11:03, 24 September 2023 (UTC)Reply[reply]
M2k~dewiki (talk) 11:06, 24 September 2023 (UTC)Reply[reply]
if I understand it correctly, this will not be of any use for users without language specific settings here, for flyby users not logged in, for anonymous users, for users in other Wikipedias. It is editor-centered, not reader-centered? --Herzi Pinki (talk) 15:55, 25 September 2023 (UTC)Reply[reply]
If you are not logged in, you get some mix of browser language settings for content pages and english as standard language, e.g. for user interface elements. For SPARQL-Queries you have to specify the language for the labels in the query. M2k~dewiki (talk) 16:06, 25 September 2023 (UTC)Reply[reply]

Interested in a little image-cropping project?[edit]

Normally I would contact the user about this, but the account is appears to be stale. This pertains to a user who have added P18 statements which I consider to be unsuitable, partly because they are PDFs but mostly because they they do not serve as illustrations or photos of the subject in question, as exemplified on Edvard Grieg (Q80621) and Giuseppe Verdi (Q7317). Now it should be possible to crop an image from some of the PDFs if you are interested in a little project. If there is no interest I plan to mass-deprecate all the statements, that is every P18 statement made by the user that links to a PDF. Infrastruktur (talk) 18:06, 22 September 2023 (UTC)Reply[reply]

Careful, though - some of them do serve as illustrations of photos of the subject in question, even if they are PDFs.
As for stale - the user appears to have been Wikipedian in Residence in Belgrade, so if you really want, you could contact Wikimedia Serbia. DS (talk) 19:38, 22 September 2023 (UTC)Reply[reply]
IMO if the illustration is on the first page, it might be better than nothing, if it's on page two it's useless in it's uncropped state. Infrastruktur (talk) 21:01, 23 September 2023 (UTC)Reply[reply]
I think it would be useful to start with a page using Wikidata:Listeria query to track the usage. -- William Graham (talk) 15:37, 23 September 2023 (UTC)Reply[reply]
You can install the Crop Tool ( at c: and easily crop out PNGs from PDFs. If someone needs help with that, let me know. —Justin (koavf)TCM 23:26, 23 September 2023 (UTC)Reply[reply]
I tested it on Lazar Jovanović (Q31186075) just now. It looks much better. Infrastruktur (talk) 00:11, 24 September 2023 (UTC)Reply[reply]
Nice. Teamwork makes the dream work. <3 —Justin (koavf)TCM 00:13, 24 September 2023 (UTC)Reply[reply]

Limit of 400 readings from Wikidata per page[edit]

On Slovene Wikipedia we have List of populated places in Slovenia (S) that consists of 694 entries. We are trying to add Wikidata image (Property:P18) to every entry, but there is a limit of 400 readings from Wikidata. Is there any possibility to set this limit at 700? Janezdrilc (talk) 13:57, 24 September 2023 (UTC)Reply[reply]

@Janezdrilc: Changing that limit alone won’t be enough, I’m afraid – as the change that introduced it points out, above 500 accesses you also start to run into the general expensive parser function limit, which is 500 for all wikis. You can file a Phabricator task (that might be a better place to discuss this than project chat), but to me it doesn’t feel likely that this limit will be lifted significantly :/ Lucas Werkmeister (WMDE) (talk) 10:59, 25 September 2023 (UTC)Reply[reply]

Ok, no problem. Thanks for answer. --Janezdrilc (talk) 11:57, 25 September 2023 (UTC)Reply[reply]

Page deletion[edit]

This page should be deleted, I accidentally created it when the page already exists

Q122804948 21:40, 24 September 2023 (UTC)Reply[reply]

Which is the correct page? Michgrig (talk) 21:50, 24 September 2023 (UTC)Reply[reply]
Q117634166 is the correct page 21:59, 24 September 2023 (UTC)Reply[reply]

One item about two lakes is a mess.[edit]

Lake Akdoğan (Q30681483) states that it is instance of (P31) lake (Q23397). But the linked Wikipedia articles are about two neighboring lakes with the same name (distinguished as the "bigger" and the "smaller"). The numbers in the item for height, depth and area are for the bigger lake, but the image pictures the smaller lake. This is confusing and misleading. How to fix it? Dipsacus fullonum (talk) 00:13, 25 September 2023 (UTC)Reply[reply]

It seems that the only reason that there is currently one item is because the Wikipedias each represent two lakes in a single article. There should be nothing preventing us from dividing this up into two items, Akdoğan Gölü and Küçük Akdoğan Gölü, because if I am to believe the English-language article, they are distinct, neighboring lakes, and they are separated by a strip of land, and as you describe, they each have separate characteristics. Elizium23 (talk) 02:16, 25 September 2023 (UTC)Reply[reply]
This is known as the "Bonnie and Clyde" problem (see intro paragraphs at Wikidata:Sitelinks to redirects). There are two options: 1) Keep it a an item about the pair of lakes, and create new items (unless they already exist?) about each lake or 2) Make it about one lake, and create a new item about the other, and another item about the pair. In either case, move the relevant values and interwiki links to the correct items. Since the item was originally created to match the en.Wikipedia article, option 1 would eb best in this case. Link the tree items with "has part"/ "part of" statements. Finally, check all inbound links and update if needed. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:25, 25 September 2023 (UTC)Reply[reply]
Bonnie and Clyde (Q219937) is instance of (P31) duo (Q10648343). What should an item for two lakes be an instance of? Dipsacus fullonum (talk) 13:03, 25 September 2023 (UTC)Reply[reply]
group of lakes (Q5926864), possibly also dyad (Q29431432) but it isn't used much. Peter James (talk) 16:28, 25 September 2023 (UTC)Reply[reply]
Or simply Wikipedia article covering multiple topics (Q21484471) if those two lakes aren't described together as one (two-part) object by reliable sources. --2A02:810B:580:11D4:88F5:DCA0:EAAF:636D 12:49, 26 September 2023 (UTC)Reply[reply]
Yes, I agree; the "Bonnie and Clyde" conundrum is for two entities inextricably linked yet independent in small ways. These are two separate lakes. They happen to share a name, but that hasn't hindered the Turkish from distinguishing them.
Lazy Wikipedians simply mashed them into one article so they'd survive AFD and there's hardly anything to write about both together. Elizium23 (talk) 14:50, 26 September 2023 (UTC)Reply[reply]
Three of the four references in the English article ( mention both, so I think "generally described together" (in Q5926864) applies; the other (reference 2) just mentions one. Peter James (talk) 21:36, 26 September 2023 (UTC)Reply[reply]

Suggestion list ordering[edit]

How can we fix the ordering in the suggestion list for sex or gender (P21), so that the most frequently used values appear first? Discussion on its talk page has proved fruitless. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:16, 25 September 2023 (UTC)Reply[reply]

Wikidata weekly summary #595[edit]

Basic question. How do I mark that a drug has a en:boxed warning?[edit]

Basic question. How do I mark that a drug has a en:boxed warning? I notice that boxed warning (Q879952)exists, but I just can't figure out what statements I should add. I don't see that binary values are supported (0/1 or No/Yes...). I'm not sure if how to do the source property stuff. Looks like I have to look up the Q#s for each thing? Or no? Use the UNII? Even once I figure out what to do, I need help. Can't use - I'm not autoconfirmed. See . RudolfoMD (talk) 21:31, 25 September 2023 (UTC)Reply[reply]

@RudolfoMD Basically, the item would be attached to the drug by a property - "X [has effect] Y" or "X [is a type of] Z" or whatever. The underlying problem is deciding what property to use for modelling it, since these have to be taken from a specific list (or created) rather than just being a yes/no value linked to a specific item. Once you've worked out the model, actually adding the statements would be relatively straightforward.
I'm not an expert here, but after a bit of reading it seems that maybe the property legal status (medicine) (P3493) would seem appropriate, using a qualifier to indicate the boxed requirement - eg in Q5515069#P3493 you have a qualifier saying the Schedule 10 ranking applies in Australia, and there could be another qualifier to indicate it needs a boxed warning. (I'm not immediately sure what that qualifier property would be, but we could definitely figure something out.) That approach would make sense if the boxed warning only applies in the specific context of the national licensing for the drug - which I think is true? The US might require a boxed warning, and France use an additional pictogram, while Germany has nothing, or whatever.
The odd thing here is that legal status (medicine) (P3493) is very rarely used for some reason - but I'm guessing that whatever source you're importing this data from would let you also import the underlying licensing status. If not, an alternative approach would be to use legal status (medicine) (P3493):boxed warning (Q879952) on its own, seperate from the existing licensing statements if they exist.
Finally, the ultimate fallback is to create a new property ("boxed warning requirement" or something) which would then have a "requires boxed warning" or "does not require boxed warning" value - there's a process for property creation but if it's needed it could certainly be done. However, it feels like going down the qualifier route might make most sense. Andrew Gray (talk) 13:06, 29 September 2023 (UTC)Reply[reply]

Q item of related concepts[edit]

There are tons of items, such as arg max and min (Q287353), directly copied from wikipedia pages, which are pairs of separate but related concepts. (Is this called umbrella/collective item?) I think these concepts should be created as separate items, such as arg max (Q122827897) and `arg min`. My question is how to interlink (by which properties) between collective and individual ones? Thanks JuguangXiao (talk) 06:25, 26 September 2023 (UTC)Reply[reply]

I've seen the use of part of (P361) and has part(s) (P527). Then make the item for both an instance of Wikipedia article page (Q50081413). And if possible add intentional sitelinks to redirects covering the individual concepts for wikis that only have the combined articles. As examples see short and long titles (Q118210315), short title (Q2470803), and long title (Q6673543). -- William Graham (talk) 16:19, 26 September 2023 (UTC)Reply[reply]
Wikipedia article covering multiple topics (Q21484471) would be better, with main subject (P921) being used for the "subtopics" instead of has part(s) (P527). —Xezbeth (talk) 16:53, 26 September 2023 (UTC)Reply[reply]
That's what I have used before, but there is also pair of concepts (Q814383), which is probably more accurate for most items of this type. Peter James (talk) 15:14, 28 September 2023 (UTC)Reply[reply]

Merge Mullan Road entities[edit]

Mullan Road (Q106042441) and Mullan Road (Q1952200) should be unified. The Commons link should be Category:Mullan Road. Bri (talk) 18:36, 26 September 2023 (UTC)Reply[reply]

WikidataCon 2023: first version of the program is out[edit]

Hello all,

On behalf of the WikidataCon 2023 organizing team, I’m happy to share that we released the program of the conference dedicated to the Wikidata community. Co-organized by Wikimedia Deutschland and Wikimedia Taiwan, it will take place on October 28-29 with a hybrid format. In this email, you will find more information about the format, the program and registration for the conference.


While most of the world will attend the sessions and discussions online, people living in Taiwan and the neighbouring countries can decide to join the onsite part of the event taking place in Taipei. Online participants will be able to join the conference, follow the sessions and interact with other participants on the platform GatherTown. Onsite participants will be hosted at the National Taipei University. You will find more information about the onsite and online aspects of the conference onwiki.

Part of the program (Day 1 and Day 2) will be delivered from the onsite event in Taipei, on their day time. The other part of the program (Day 1.5 and Day 2.5) will be produced online, at times that will be more friendly to participants located on the European, African and American continents. In any case, the contents of the conference will be available online during or after the conference.


Together with volunteers involved in the program committees, we elaborated an ambitious program that will run across time zones. Connected to our theme “Collaboration across boundaries”, the sessions will cover a huge range of topics (data quality, data reuse, Wikidata and the Wikimedia Projects, open data initiatives). In particular, we want to highlight the Wikidata and open data projects and initiatives in the ESEAP region.You can find the first version of the schedule here. You can switch time zones, languages, and mark sessions as favourites. Some changes may still happen in the upcoming weeks.

English and Mandarin are the official languages of the event. The sessions will be available in these two languages, either via captions or interpretation. If you are interested in helping us translate contents (program, wikipages, sessions) please get in touch!


You can already sign up to the event by registering here on the Pretix platform. The registration form includes both online participation and onsite participation in Taiwan. Registration will be open until the start of the event. Some useful information will be sent in the upcoming weeks to people who registered, including the link to the event platform for online participants. Please note that registration is free of charge, but no scholarship to attend the online or onsite event will be provided.

Questions ? Feel free to contact the organizing team by writing on this talk page or at

We are looking forward to seeing you at the WikidataCon 2023!


主題:2023 年維基數據大會:第一版活動議程上線!


很高興我代表2023年維基數據大會組織團隊,發佈本次 Wikidatacon 2023 的第一版議程。這是由德國維基媒體協會和台灣維基媒體協會共同主辦,會議將於 10 月 28 日至 29 日以混合形式舉行。在此電子郵件中,你會看到更多有關會議形式,專案以及註冊的資訊。


儘管世界上大多數人將參與網上會議以及線上討論,但居住在台灣和鄰近國家的人,可以決定參加在台北舉行的活動的實體活動。在線參與者可加入會議,關注會議並在Gather Town 平台上,與其他線上參與者互動。實體參與者將在國立臺北大學舉行。你可以在 wikidata page上找到有關會議現場和在線方面的更多資訊。

部份議程(第 1 天和第 2 天)將在白天的臺北現場活動進行。議程的另一部份(第 1.5 天和第 2.5 天)將在線上呈現,該時段更適合對位於歐洲,非洲和美洲大陸的參與者將。會議內容將在會議期間或在這之後,以線上提供。


我們與參與專案的委員會的義工一起制定一項具有企圖心的計劃,該計劃將跨時區運行。與我們的主題“:鏈結資料 Koh(再)發現:Wikidata 的跨領域匯流"”相關,會議將涵蓋廣泛的主題(資料品質,資料重用,維基數據和維基媒體專案,開放資料倡議)。




你已經可以透過在 Pretix 平台上報名該活動。報名表包含線上參與者,和臺灣現場參與者。報名將開放到活動日開始。相關有用的資訊將在未來幾周內寄送給已完成報名者,包括線上參與者的活動平台連結。請注意,報名是免費的,但不提供參加線上或現場活動參與的獎助金。

有問題?請隨時透過在此討論頁或 上寫信與組織團隊聯絡。

我們期待在2023年維基數據大會上與你見面! Lea Lacroix (WMDE) (talk) 06:08, 27 September 2023 (UTC)Reply[reply]

Изтрита Уикипедия[edit]

Здравейте казвам се Пламена Червенкова забелязах,че страницата ми е изтрита без причина от анонимен.Моля да бъде възстановена! Благодаря 10:04, 28 September 2023 (UTC)Reply[reply]

Page deletion[edit]

This page Q122854647 isn't notable 19:40, 28 September 2023 (UTC) Reply[reply]

I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. --Matěj Suchánek (talk) 08:17, 30 September 2023 (UTC)Reply[reply]

Help to resolve triggering an abuse filter via OpenRefine upload[edit]

Hi, I'm trying to upload new items to Wikidata via OpenRefine. The load will create new items for each issue of a journal title. I'll then do another load of article metadata to link to each issue Qid. I've successfully done this before for other journal titles, but this time I discovered my load was failing due to abuse filter 36: non notable subpages. I'm unsure what it is that Wikidata isn't liking, or where to ask for help as I can't see a discussion page on the abuse filter page itself Any advice or help much appreciated. Thanks HelsKRW (talk) 10:12, 29 September 2023 (UTC)Reply[reply]

I think what's happening here is that the filter is meant to stop creating items with a sitelink (to another wiki) that contain specific strings, eg /archive, to stop things like people creating items for discussion subpages. Your upload has one url with /archive in it, and I wonder if somehow that's triggering things?
I'm not very experienced with these so it's not obvious to me if the filter is actually only stopping sitelinks or if it's stopping any new lines at all. pinging @Matěj Suchánek: who was the most recent editor there, and will understand it better... Andrew Gray (talk) 12:39, 29 September 2023 (UTC)Reply[reply]
Thanks for your help @Andrew Gray I can replace the /archive URL with something else, which I've just done and am trying to upload, but OpenRefine is currently sitting at 0% complete! I'll have another go on Monday morning and report back! HelsKRW (talk) 14:55, 29 September 2023 (UTC)Reply[reply]
I updated the edit filter not to check URL addresses. You should now be able to create those items. Sorry for the disruption. --Matěj Suchánek (talk) 08:16, 30 September 2023 (UTC)Reply[reply]

typo Univeristy --> University[edit]

Hello, anybody hanging around to take a look at the spelling of university? "Univeristy" should be corrected. Thank you so much for your time. Lotje (talk) 13:28, 29 September 2023 (UTC)Reply[reply]

Also see M2k~dewiki (talk) 20:52, 29 September 2023 (UTC)Reply[reply]
Also see de:Benutzer:M2k~dewiki/FAQ#Wie_kann_ich_massenweise_Tippfehler_in_Beschreibungen_korrigieren? M2k~dewiki (talk) 20:57, 29 September 2023 (UTC)Reply[reply]
Thank you M2k~dewiki for coming back on this. Do you mean you or Aka could change that "in one go"? Lotje (talk) 03:46, 30 September 2023 (UTC)Reply[reply]

Describing a page range for an article within a PDF file[edit]

I want to document that Fan Bay Deep Shelter (Q20711474) is described by pages 49-62 of issue 38 of magazine Subterranea (Q115785692), which I know is in the Internet Archive as a PDF at I know that the magazine has a cover and inside cover without page numbers, it numbers from 1 on the third page. Should I refer to the visible page number, 49, or the PDF page 51, so that the #page= extension to the URL loads at the right point. I could quote the URL itself with the correct PDF page, but that feels inflexible. Is there any merit in documenting this page offset in the magazine description? Vicarage (talk) 13:53, 29 September 2023 (UTC)Reply[reply]

I think file page (P7668) was created for cases like that. See the property examples there. --2A02:810B:580:11D4:2F4:6FFF:FE72:D65F 07:11, 30 September 2023 (UTC)Reply[reply]
Thanks. I see there is also file page offset (P8330), though that's currently restricted to being a qualifier. In Property_talk:P8330#Use_for_magazines_and_journals I propose it being allowed at the higher level. Vicarage (talk) 08:30, 30 September 2023 (UTC)Reply[reply]

Where to gain traction for proposed technical improvements?[edit]

I know how to create Phabricator tickets, but not how to get them prioritized. I think there is some recurring event where the community votes on improvements for the developers to prioritize, but I can't find it. Does anyone know what I'm talking about? Swpb (talk) 16:29, 29 September 2023 (UTC)Reply[reply]

M2k~dewiki (talk) 20:40, 29 September 2023 (UTC)Reply[reply]
There's the wishlist survery. Besides that you can also talk to Lydia Pintscher (WMDE) who's responsible for the main priorisitation around Wikidata. ChristianKl❫ 08:33, 30 September 2023 (UTC)Reply[reply]

Radio frequencies[edit]

There are radio-frequencies which are quite stable : famous radio emitters, official frequency bands, maritime or airlines frequencies used for security or communication purpose etc...

Examples: Q209421 or Q4698543.

The frequency as a number of Hertz very rarely used (possibly never), and also it lacks the modulation which is a kind of qualifier of the frequency: AM (Q183755), MFSK (Q3867059), plus the location of the emitter (also a qualifier) etc...

This would allow to upload list of public frequencies such as [1] or [2] and would be interesting for radio-amateurs : these databases are often created and contributed by hobbyists. What do you think ? Thanks. Rc1959 (talk) 20:02, 29 September 2023 (UTC)Reply[reply]

I don't know. There is nothing inherently special about a single frequency, and they are often allocated temporarily to a service provider or operator. The temporary nature of radio frequencies makes them a bad fit for Wikidata.
In some cases their use is internationally standardized, at which point it might deserve it's own item, such as 121.5 MHz. At this point there is probably a wikipedia page about it too. Some of those old and large radio transmitters, especially those transmitting on lower than shortwave might have wikipedia pages, or at least there is a good chance they are considered notable. Officially recognized frequency bands should be notable too as they are described by serious sources.
It might be more feasible to link to an external source in the case of radio transmitters, that describes things like their frequency and modulation. If you know a good source for this you can propose it as an external identifier, see WD:PP. Infrastruktur (talk) 17:18, 30 September 2023 (UTC)Reply[reply]

Temporary accounts for unregistered editors[edit]

Read this in your languagePlease help translate to your language • Please tell other users about these changes

Mock up of history page showing old and new username styles. The IP address changes to the temporary account ~2024-23126-086, with an icon for revealing the underlying IP address
Next year, unregistered editors will start using temporary accounts.

In 2024, editors who have not registered an account will automatically begin using temporary accounts. These editors are sometimes called "IP editors" because the IP address is displayed in the page history.

The Trust and Safety Product team gave a presentation at Wikimania about this change. You can watch it on YouTube.

There is more information at m:IP Editing: Privacy Enhancement and Abuse Mitigation.

SGrabarczuk (WMF) (talk) 02:05, 30 September 2023 (UTC)Reply[reply]

Property for conveying accessibility information about venues and events[edit]

We currently use the property wheelchair accessibility (P2846) to convey information about the accessibility of venues and events.

The property is presently used in about 7100 instances. In most of the cases, it is used to indicate the "wheelchair" accessibility of the place or event, as the label would suggest. In about 60 cases, however, it is used to indicate whether a place/event is accessible for people with visual impairment.

I am currently cooperating with the Sitios Association, which is the result of a recent merger of several initiatives in the field of accessibility information in Switzerland. The goal of the organization is to further build up and maintain a database with accessibility information about places/venues in Switzerland. The database, which can be found at (name change pending), currently contains about 12'000 entries - a number that is expected to double over the coming two years. We are about to start (1) linking the Sitios database to Wikidata (bidirectional linking via IDs), (2) complementing Wikidata entries based on the Sitios database, (3) adding high-level accessibility information directly to the respective Wikidata items, with links to detail pages of the Sitios database.

The dimensions for high-level accessibility information we would like to insert into Wikidata comprise:

  • Wheelchair accessibility (in general; toilets; parking place)
  • Availability of a magnetic loop (for the hearing impaired)
  • Availability of an offer in sign language
  • Accessibility for people with visual impairments
  • Accessibility for people with cognitive impairments

I think it would make sense to use the same property (wheelchair accessibility (P2846)) to convey the different types of accessibility information. In this case, I would however suggest to change its label to "accessibility of location or event". -- Are there any thoughts about this?

--Beat Estermann (talk) 10:08, 30 September 2023 (UTC)Reply[reply]

I agree with the proposal to change the label as indicated and allow values that are subclasses of accessibility (Q555097) or inaccessibility (Q73019188). -- William Graham (talk) 14:07, 30 September 2023 (UTC)Reply[reply]
Is there any international standard body that defines categories for this? ChristianKl❫ 14:51, 30 September 2023 (UTC)Reply[reply]