Wikidata:Project chat/Archive/2018/04

From Wikidata
Jump to navigation Jump to search

"Search" does not return the obvious one

When I Search, WD does not return the most obvious hit (i.e., the one I am expecting). Searching "Horse" (lang=English) [1], the top-3 results are:

Vale of White Horse (Q1540110)
Dark Horse Comics (Q373933)
horse breed (Q3745054)

Why is not horse (Q726) in top?

(I'm sorry if this is a repeated point, but ay me searching this page makes it recursive ;-). And also consider the date). - DePiep (talk) 22:33, 1 April 2018 (UTC)

It is a known issue - there's a tiny bit more at Wikidata:Suggester ranking input which explains a few odd results, and an echo of your query, with a response, here. --Tagishsimon (talk) 22:46, 1 April 2018 (UTC)
This section was archived on a request by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:08, 3 April 2018 (UTC)

Problem with a property

Nominis given name ID (P4284) in format as a regular expression (P1793) say \d{2,3}/([a-z]+) but Tommy (Q1272055) has 6305/Tommy as ID. It is valid but does not validate the regex. What should be done in a case like that? Is it simply modified? --Metrónomo (talk) 05:40, 3 April 2018 (UTC)

@Metrónomo: I've updated the regex. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:07, 3 April 2018 (UTC)
This section was archived on a request by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:07, 3 April 2018 (UTC)

Linking human to a related wikisource category

Easiest way I found: https://www.wikidata.org/w/index.php?title=Q51187097&diff=659303150&oldid=659296753, but probably not correct- What is the correct way? 78.53.143.236 02:38, 1 April 2018 (UTC)

That looks OK to me. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:08, 3 April 2018 (UTC)
This section was archived on a request by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:54, 5 April 2018 (UTC)

How to define the location a report is about?

If a report or paper is about a geographic area, how to define that? I've looked at narrative location (P840), geography of topic (P2633) and applies to jurisdiction (P1001) but they all seem too specific. Is there any other property that could be used for this? Pauljmackay (talk) 07:36, 2 April 2018 (UTC)

@Pauljmackay: How about main subject (P921)? --Okkn (talk) 10:26, 2 April 2018 (UTC)
@Okkn: Yes that was the most obvious candidate, except I dont think its ideal because its more about defining a restricting scope, rather than defining what the subject of a report is about. Pauljmackay (talk) 16:24, 2 April 2018 (UTC)
You can use of (P642), as in "Birds of Portugal", "Snails of the Severn Estuary". See, for example, Fauna of New Zealand (Q21385461). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:16, 2 April 2018 (UTC)
This section was archived on a request by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:53, 5 April 2018 (UTC)

Merging IRC client list pages

Hi ! I'm quite a newbie on wikidata, so please, can someone merge together comparison of Internet Relay Chat clients (Q2378978) and list of IRC clients (Q2622988) ? Both are on the same subject. Thanks ! --Evachan39 (talk) 16:15, 4 April 2018 (UTC)

This cannot be done, as the former links to pt:Comparação dos clientes de IRC and the latter to pt:Cliente de IRC. That's not to say that someone couldn't propose merging those articles on the Portuguese Wikipedia. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:22, 4 April 2018 (UTC)
This section was archived on a request by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:52, 5 April 2018 (UTC)
@ User:Pigsonthewing -> Je suppose que vous parlez francais au vu du message que vous m'avez posté sur mon mur ([2]) je vous écrit donc ce message en francais (la chose est pour moi plus aisée). Je ne connais pas exactement les principes de fonctionnement de fonctionnement détaillés de wikidata, donc excusez-moi si je me trompe, mais les liens que vous avez donné se référent à des choses différentes. Ces histoires de liens meta m'ont l'air relativement peu organisés sur le sujet... [3] (cf [4]; liste de clients IRC=list of IRC clients) et [5] (confer [6]; comparaison de clients IRC=comparison of IRC clients) sont véritablement très proches au niveau de leur signification.

Je pense que vous confondez probablement (???) les deux précédents, qui sont des listes de clients IRC, avec la page dédiée au client IRC, par exemple : [7] (soit [8]). Post Scriptum ː D'ailleurs, mon clavier fait des siennes sur ce site, je n'ai pas l'impression qu'il supporte les clavier "de type francais" (à défaut de savoir les nommer correctement ?) ou je m'y suis évidemment mal pris (ce qui est très vraisemblablement le cas à mon humble avis ǃ). --Evachan39 (talk) 16:29, 6 April 2018 (UTC)

Undo blanking

I wanna undo this blanking edit. When I try to do so I get this errors: 

Errors:

    The text you wanted to save was blocked by the spam filter.

This is probably caused by a link to a blacklisted external site.

    The following text is what triggered our spam filter: twitter.com/search

Can you help me? --Jobu0101 (talk) 22:07, 4 April 2018 (UTC)

I have just responded to your request for help with this issue, at Wikidata:Administrators' noticeboard#‎Undo edit:. Please don't post such things in multiple locations. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:50, 5 April 2018 (UTC)
✓ Done Matěj Suchánek (talk) 14:53, 5 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 14:53, 5 April 2018 (UTC)

Merge

Please merge Q51514156 (source http://www.bbl-digital.de/eintrag/Schiemann-Carl-Christian-1763-1835/ physician in Mitau/Jelgava, studied in Göttingen) into Q21607917 (source http://www.ipni.org/ipni/idAuthorSearch.do?id=9076-1 -> biographical details: also mention "Dr. med" from Mitau who studied in Göttingen) Year of birth and of death for both items are the same. 78.55.48.88 04:38, 6 April 2018 (UTC)

✓ Done. --Liuxinyu970226 (talk) 06:37, 6 April 2018 (UTC)
This section was archived on a request by: Liuxinyu970226 (talk) 11:54, 6 April 2018 (UTC)

Merge Yuta Nakamoto and Nakamoto Yuta

Please merge Yuta Nakamoto to Nakamoto Yuta; w:es:Yuta Nakamoto redirects to w:es:NCT (d:Q23725822). Regards, ----Omotecho (talk) 17:43, 6 April 2018 (UTC)

→ ← Merged Matěj Suchánek (talk) 17:59, 6 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 17:59, 6 April 2018 (UTC)

Classifiying knots

I checked some knot items and all used instance of (P31) to classify them. Based on one of my earlier questions about use of subclass vs instance of, is it correct to suggest that knots should use subclass of (P279) instead? Because for any knot there could be many "instances". My understanding is that there is a preference for instance of to only be used for things that only have a single instance. Pauljmackay (talk) 19:29, 31 March 2018 (UTC)

I think knots should be subclasses (except for things like Gordian knot (Q193373). - PKM (talk) 03:06, 1 April 2018 (UTC)

Referencing works in other works

Hi, how can I reference a song that is used in a TV show episode or a movie? If a novel mentions another novel, or a song or movie, is there a way to include that?--37.201.30.50 08:25, 1 April 2018 (UTC)

Need help in translation of 4 descriptions to EU languages

I'd appreciate some help in translating 4 descriptions to EU languages: in the table here (other languages are welcome too). These are about elements of Globally Harmonized System of Classification and Labelling of Chemicals (Q899146) and will be used in a few hundred items I'm planning to create for safety classification and labelling (P4952) (and proposed qualifiers for this property). Wostr (talk) 16:13, 1 April 2018 (UTC)

Category:Devşirme

The people in this category were taken from their homes as kids and educated to be high level officials in the Ottoman Empire. How do I indicate on the person that this happened to him? Thanks, GerardM (talk) 07:45, 1 April 2018 (UTC)

Slaves are recorded as Sojourner Truth (Q105180) social classification (P3716) enslaved person (Q12773225), but it lacks the link with devşirme (Q815841). Ghouston (talk) 11:38, 1 April 2018 (UTC)
They were not exactly slaves. Some of them became the highest functionaries of the realm, married the sister of the Sultan. So this is too simple. Thanks, GerardM (talk) 12:26, 1 April 2018 (UTC)
Still, a kind of social class, just lacking an item for the person that corresponds to the practice. Ghouston (talk) 12:43, 1 April 2018 (UTC)
What about Mesih Pasha (Q2480875)subject has role (P2868)devşirme (Q815841) provided that devşirme (Q815841) is a designation of a person and not a social class i.e slave contra slavery. Pmt (talk) 00:18, 2 April 2018 (UTC)
I think that would be like saying Mesih Pasha (Q2480875)subject has role (P2868)slavery (Q8463), it'd still need an item for the class of people who are subjected to it. Ghouston (talk) 01:55, 2 April 2018 (UTC)
I made it a "significant event". GerardM (talk) 05:35, 2 April 2018 (UTC)

Art+Feminism workplan + QuickSheets proof of concept

I wanted to share our plan for Art+Feminism related Wikidata work, which includes a proof of concept for QuickSheets, a set of python scripts which semi-automate the addition of occupation (P106) data. Please see our work plan here. —Theredproject (talk) 16:35, 1 April 2018 (UTC)

@Theredproject: (and others), I'll give feedback on User talk:Theredproject/2018 Workplan, that seems to be the best place. Multichill (talk) 19:25, 1 April 2018 (UTC)

Conflict with separate Wikidata items vs. single, combined Wikipedia articles about distinct works

Hi. Is there an established way how Wikidata and Wikipedia handle cases where the Wikipedia article covers several distinct works while standard practice over here is to create separate items for each of those? How are Wikipedia templates on such articles able to pull data about these works from separate Wikidata items?

The reason I'm asking: User:InMontreal alerted me to a problem here. I was adding external IDs for movies and TV shows and came across some anime shows that didn't have their own items yet. When it comes to mangas/light novels and their anime/film/game adaptations, standard practice on Wikipedia (all/most language versions) seems to be to cover them all in one big article. Which means that currently our items in that field are usually a wild mix of statements pertaining to mangas, light novels, animes, video games, etc, and also include external IDs like IMDb ID (P345), AlloCiné series ID (P1267) and Anime News Network anime ID (P1985) for specific works. So I created several new items for anime series and moved all respective IDs and statements for these shows to the new items.

But it seems on various Wikipedias such IDs might be removed after they have been transferred to Wikidata, so the link templates no longer contain them and simply pull the ID from Wikidata. So when you split them into separate items for each work (as is usual on Wikidata), those IDs are also moved to those items and are no longer accessible for Wikipedia templates. Infoboxes also suffer the same problem of course. I hope there's a solution besides just telling Wikipedians to deal with it themselves.

== Active users == ValterVB LydiaPintscher Ermanon Mushroom Queryzo Danrok Rogi Mbch331 Jobu0101 putnik AmaryllisGardener Andreasmperu U+1F350 Bodhisattwa Shisma Wolverène Tris T7 Antoine2711 CptViraj ʂɤɲ Trivialist 2le2im-bdc Sotiale Wallacegromit1, mostly focus on media historiography and works from the Global South M2k~dewiki Rockpeterson Mathieu Kappler Sidohayder Spinster Gnoeee Ranjithsiji Ontogon Supaplex Carlinmack Haseeb Demadrend Jakeob9000 RealityBites Sriveenkat Keplersj dseomn Fuzheado BeLucky

Notified participants of WikiProject Movies --Kam Solusar (talk) 01:56, 24 March 2018 (UTC)

Here's the advice: Help:Handling sitelinks overlapping multiple items --Tagishsimon (talk) 02:03, 24 March 2018 (UTC)
It's curious that the Help article doesn't mention Wikipedia article covering multiple topics (Q21484471) at all. - PKM (talk) 18:53, 24 March 2018 (UTC)
So it seems the "solution" is indeed "just telling Wikipedians to deal with it themselves" :-/ . I mean, it probably isn't a bad idea for Wikipedias to modify their infoboxes to pull their data from other items if needed. But who's going to tell them, since it doesn't seem like the infoboxes on the affected articles are able to do that at the moment?
So, as the guy who created the new items and moved the statements over, I guess I have two options: a) tell two or three dozen Wikipedias they now have to change their link and infobox templates to deal with the changes to a few items I chose to make on Wikidata - although to them, there wasn't really a problem with the status quo that needed fixing. And then I'd probably have to justify the changes in many languages that I don't even speak. Or b) undo my edits and give the whole topic a wide berth in the future - in the hopes that someone with a lot more time, energy and patience will stumble upon this problem and is willing to deal with it. --Kam Solusar (talk) 12:04, 26 March 2018 (UTC)
I don't think the infoboxes on any of these Wikipedia articles worked correctly. If users there are interested, eventually they will figure out how to get the content. Thanks for starting to tackle these few remaining ones ..
--- Jura 12:13, 26 March 2018 (UTC)
I'm not sure that's the best way to handle such situations, since Wikipedians might see it as Wikidata changing/breaking things without giving Wikipedia a warning. And especially on smaller Wikipedias, it might break stuff for a long time until people figure out what happened and find someone to fix the templates. One of the problems User:InMontreal pointed out on my talk page is, that when you move all the statements and IDs to a separate item, there's no indication on the main item why they were removed and where to find them now. Or people might just assume that they are missing and try to add them back (which happens often with external IDs harvested from link templates). I didn't know a good way to indicate "for information about the anime that is also described in the linked articles, see QXXXX", though I just found derivative work (P4969) which looks like it might do the job of at least having visible links on the main item.
Though I'd prefer if there was some kind of outreach to point out the problem to affected Wikidia communities and how to handle it, so not every single community has to stumble on the problem and try to figure it out for themselves. --Kam Solusar (talk) 00:36, 29 March 2018 (UTC)
Sorry but it is not as if Wikidata is new or that practices like these are new. There are plenty of things Wikipedia does that does not make sense from a Wikidata perspective. The problem is also that communication is often broken by Wikipedians insisting on no Wikidata. So outreach sure.. But who is to do this? Thanks, GerardM (talk) 05:53, 29 March 2018 (UTC)
@Kam Solusar: The correct way to do this is to create an item for the franchise like Naruto (Q642), link the article to it and add the anime, manga, light novel, OVA etc items with has part(s) (P527).
Then Wikipedias can create modules to retrieve the correct identifier. --Thibaut120094 (talk) 03:14, 31 March 2018 (UTC)
Thibaut120094: Seems more like a workaround (as the articles' primary topic seems to usually be the manga/light novel series, not a franchise), but if that's the current solution for these cases, guess I'll use it. Though I'm not really looking to take on the task of creating items for all the works covered by those artices/items myself (especially since I don't know all that much about about manga franchises)... --Kam Solusar (talk) 11:58, 3 April 2018 (UTC)
  • If there were cases where this seemed to work before, it might worth looking into them in detail. It's likely that they were probably broken in one way or the other, but people just didn't notice (e.g. publication date of a film displayed in the infobox about a print version). To work correctly, infoboxes that are used in articles combining several topics need to do a check on P31 of items they attempt to use data from.
    --- Jura 06:33, 29 March 2018 (UTC)
Yeah, some might have shown false information for quite a while without anyone realizing it (or they realized it and didn't find someone who could help fix the problem). Checking the P31 statements helps, but I imagine there could still be some problems with items that have multiple P31 statements and in cases where there's more than 1 infobox for the same type of work on one page (i.e. a franchise with more than 1 tv show or videogame). --Kam Solusar (talk) 11:58, 3 April 2018 (UTC)

Edinburgh Research Archive - thesis collection data

Hi all, after a hiatus we are returning to look again at importing data on the Edinburgh Research Archive's thesis collection into Wikidata. The ERA has over 20,000 thesis records so we are looking to get the modelling correct - once agreed our Library Digital Scholarship Developer @ChaoticReality: can proceed with the import reasonably quickly. For a refresher on what we originally discussed in November 2017 in Wikidata's Project Chat is here and the Google document with the data model we originally proposed is here.
3 questions arose from our discussions last time:

  1. There are 24 (approx) thesis types in the collection. If I create a thesis' item for each type as was suggested e.g. ChM Master of Surgery thesis then is it to be described as P31 Instance of a masters thesis? OR P279 subclass of a masters thesis?
  2. Prize essays are a whole other kettle of fish. They seem to be described on Wikidata as scientific article but they can be on any subject I believe and sometimes become thesis dissertations and sometimes not. I wonder if it would not be better for prize essay to have its own Q number? What do you think? A reference to one with some explanation here: here. Gweduni has more info on the Prize essay distinction if you need further info.
  3. Finally, Should we also update existing items that don't yet have links? For example Doctor of Philosophy (Q752297) is a subclass of Doctoral Degree (Q849697) but Doctor of Education (Q837184) is not. So should we include statements to make that connection?


For info, all the below are theses.
Subsets of thesis: doctoral, masters, ‘other’.
Doctoral (Level 12 on the Scottish Credit and Qualifications Framework – the highest level)

  • DClinPsychol Doctor of Clinical Psychology
  • DD Doctor of Divinity
  • DDS Doctor of Dental Surgery
  • DLitt Doctor of Literature
  • DPhil Doctor of Philosophy
  • DPsychol Doctor of Psychology
  • DSc Doctor of Science
  • DVM&S Doctor of Veterinary Medicine and Surgery
  • EdD Doctor of Education
  • MD Doctor of Medicine
  • PhD Doctor of Philosophy
  • PhD(P) Doctor of Philosophy by Research Publications

Masters (Level 11 on the Scottish Credit and Qualifications Framework)

  • ChM Master of Surgery
  • LLM Master of Laws
  • MA Master of Arts – although in Scotland MA(Hons) is undergraduate, equivalent to English BA (for info MA(Hons) / BA is Level 10 on SCQF)
  • MBA Master of Business Administration
  • MEd Master of Education
  • MLitt Master of Literature
  • MLT Master of Letters
  • MPhil Master of Philosophy
  • MSc Master of Science
  • MSc(R) Master of Science by Research
  • MTh Master of Theology

Other:

  • Prize Essay

Any pointers, gratefully received. Many thanks, Stinglehammer (talk) 18:01, 28 March 2018 (UTC)

'P279 subclass' is correct. Instance should only be used for a specific thesis with a given title and author etc. If you find missing items or statements as you suggest, please add them as long as they make logical sense; wikidata is always going to be incomplete, it's up to us to help improve it! ArthurPSmith (talk) 19:27, 28 March 2018 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ I've numbered your questions:

  1. Subclass of; per Arthur
  2. "would not be better for prize essay to have its own Q number?" Yes, the maxim "be bold" applies
  3. yes.

-- Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:51, 28 March 2018 (UTC)

Well that was easy. Thanks @ArthurPSmith, Pigsonthewing:! I'll let ChaoticReality and Gweduni know so we can move this forward. (Gotta love Wikidata discussion threads. Sooooo much easier). Stinglehammer (talk) 21:25, 28 March 2018 (UTC)
Brill, thanks both. We'll be doing a pilot run of ~50 objects in the first instance, so I'll probably post again closer to the time we're ready to do this. ChaoticReality (talk) 13:12, 29 March 2018 (UTC)
@ChaoticReality, Pigsonthewing, ArthurPSmith: Ok

Have created Wikidata numbers for the thesis types now and added on this Google sheet. Items in red are newly created thesis type items. Items in blue indicate there is a one item being used for two thesis types. i.e. Master of Literature (Master of Letters) and Master of Letters (Master of Letters) are both Q1907865 so the master’s thesis is Q51283362 for both. 3 questions cropped up again:

  1. Checking with Gweduni that he is happy with this or whether we should separate the two. Thoughts?
  2. Prize essay now has its own Q number: Q51282441. Are we happy how it is described? Checking with Gweduni as to what would be an adequate description and statement type for a prize essay.
  3. I’ve added that each degree type (e.g. Master of Surgery degree) is a subclass of master’s or doctoral degree etc. but there might be some extra tidying up needed – seems to be a real mix of instance and subclass statements being used to describe each degree entirely differently at present. What should be the main statement modelling/hierarchy we should be working with on these degrees do you think to keep things logical & consistent? Stinglehammer (talk) 11:38, 3 April 2018 (UTC)

Capturing provenance of food

Is there any property used to capture the region or more specific location where a particular food originates from? For example that a pasty originates from Cornwall? There is country of origin (P495) but thats only for countries. There is location of creation (P1071) but that seems more focussed on manufacturing rather than where a food product is originally from. Does this need a new property? Pauljmackay (talk) 11:11, 31 March 2018 (UTC)

There are some claims like Valtellina Casera (Q782709) part of (P361) Lombard cuisine (Q47176), but maybe that's not quite the same thing. Ghouston (talk) 11:33, 31 March 2018 (UTC)
@Pauljmackay: - indigenous to (P2341) is explicitly scoped to include "cooking style", so this would seem a good approach. Just be careful with baklava (Q187495) and Pavlova (Q1419824) ;-) Andrew Gray (talk) 11:53, 31 March 2018 (UTC)
@Andrew Gray: - that seems like a good option except after adding it to Cumberland sausage (Q586591) it flags an issue because the constraints on that property require it to be a dish (Q746549), which wont be true for all food products. Should the constraint be altered? Pauljmackay (talk) 12:22, 31 March 2018 (UTC)
@Pauljmackay: yeah, I think that's the best approach - broaden the constraint a bit. As well as a cooking style we could have indigenously-identified prepared foods, basic foodstuffs, spice, drinks... Andrew Gray (talk) 14:39, 31 March 2018 (UTC)
Might also be useful to note foods etc that have a formally protected en:Geographical indication. There's been an industry initative to produce a worldwide database of these (report), and it looks like the database is usually live here, though just at the moment it seems to have fallen off the net. But here's Google's cache of the first page.
Everything on the list probably ought to have an item, if it doesn't already. Property named after (P138) may also be useful. Jheald (talk) 19:02, 31 March 2018 (UTC)
There is location of creation (P1071) with alias "location of origin" and "place of origin". Current label is more focused on manufacturing, but it was changed in 2014 from "place made", which was an original label in time of proposal.
We have a lot of items about appellation d'origine contrôlée (Q1565828) and most of them use located in the administrative territorial entity (P131), which is not appropriate in my view.--Jklamo (talk) 12:25, 3 April 2018 (UTC)

Property with unproper type

Hello, I think the property Patientplus ID (P1461) should be of type external identifier. Louperivois (talk) 21:29, 31 March 2018 (UTC)

Indeed. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:09, 3 April 2018 (UTC)

Academic writing: activity or type of work?

academic writing (Q4119870) has been a confusing item since it has been created. Is this about the activity of writing academic works or is it a type of academic work? I can't read all languages. Sjoerd de Bruin (talk) 11:25, 2 April 2018 (UTC)

I think Wikidata needs separate items for the type of work and the activity of writing such work. I would expect a Wikipedia article to cover both the activity and the work it produces (in fact, I think that it would be hard to separate them in an encyclopedic article). For the sake of useful Interwiki links, I'd keep the article links on the existing work item and create a new linked item for the activity. (For an interesting comparison, see technical writing (Q1193158) [activity] and technical documentation (Q1413406) [work].) - PKM (talk) 21:09, 2 April 2018 (UTC)

Wikidata weekly summary #306

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Other Noteworthy Stuff
    • if you're working with Lua modules, you can give input here about new functions
    • you can still fill this page if you encounter a problem with suggester ranking

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.


Improving the definitions of crafts

The Heritage Craft Assocation defines a "heritage craft" as "a practice which employs manual dexterity and skill and an understanding of traditional materials, design and techniques, and which has been practised for two or more successive generations" (see their Red List). IMHO most of the items that link to craft (Q2207288) are occupations, not the craft (skill, practice) itself. So stonemasonry (Q19794820) is a craft, a stonemason (Q328325) is an occupation or artisan that practice that craft.

Also these crafts and occupations should be subclasses, not instances, along similar lines to my other recent questions :-) What do others think? Pauljmackay (talk) 16:41, 2 April 2018 (UTC)

I agree on separation of crafts and craftspeople. See pottery (Q11642) and potter (Q3400050), tapestry weaving (Q29167534) and tapestry weaver (Q21820569). These can be connected using practiced by (P3095). I am less certain about instances vs. subclasses. I think <instance of> is perfectly appropriate for things like crafts and occupations. If stonemason (Q328325) isn't an instance of an occupation, what would be? - PKM (talk) 20:54, 2 April 2018 (UTC)
Maybe this overview helps? Multichill (talk) 21:52, 2 April 2018 (UTC)
On instance of (P31) versus subclass of (P279), a question I find is often quite useful is: can one imagine a specialist form of this activity or thing. So one could imagine "Chinese fan-making" or "18th-century fanmaking" or "paper fanmaking", suggesting the activity should be a class. As for the occupation, I am less convinced that one would identify "18th-century fanmaker" (ie a maker of 18th-century fans) as a distinct occupation, so I would be minded to keep it as an instance, unless/until a more specific instance gets created. Jheald (talk) 09:41, 3 April 2018 (UTC)
@Jheald: thinking more on that aspect, I wondered if a sensible structure is that a stonemason (or whatever) is an instance of an "occupation" and a subclass of "artisan". So Joe Bloggs is a stonemason or artisan, but has an occupation of stonemason. Pauljmackay (talk) 10:33, 3 April 2018 (UTC)
Except that, on Wikidata, Joe Bloggs is a human (Q5), and stonemason is not a subclass of person. artisan (Q1294787) we identify as a subclass of profession (Q28640), in turn a subclass of occupation (Q12737077); so making stonemason a subclass of artisan (Q1294787) makes it a subclass of occupation (Q12737077) as well. Jheald (talk) 11:20, 3 April 2018 (UTC)
@Jheald: currently I see artisan (Q1294787) is an instance of profession (Q28640), so does that need changing to subclass? Pauljmackay (talk) 11:47, 3 April 2018 (UTC)
@Pauljmackay: Perhaps not: Reasonator says we have 489 instances of profession (Q28640), only 8 subclasses. Passing this to the crowd, to see what anybody else thinks. Jheald (talk) 12:52, 3 April 2018 (UTC)
There’s discussion on this point on the talk page of profession (Q28640)! which I would characterize as “unresolved”. Personally, I like the distinction filmmaking occupation (Q4220920) = subclass and cinematographer (Q222344) = instance. - PKM (talk) 20:38, 3 April 2018 (UTC)

various langauges configuration

Hello, Is there a language set up for various languages? in particular bts and map-bms? Map-bms has its own wiki. I tried adding info to these languages, yet its not supported on Wikidata. Artix Kreiger (talk) 18:57, 3 April 2018 (UTC)

A sitelink to https://map-bms.wikipedia.org is on Q13149114.
--- Jura 19:08, 3 April 2018 (UTC)

2 communities merge to 1

Before 2005 two communities (villages): Kato Amiantos (Q48413729), Pano Amiantos (Q47004149). After that the two communities merged to one, new community with the name Amiantos (Q6377735). I used

⟨ Amiantos (Q6377735)  View with Reasonator View with SQID ⟩ has part(s) (P527) View with SQID ⟨ Kato Amiantos (Q48413729)  View with Reasonator View with SQID ⟩

and

⟨ Amiantos (Q6377735)  View with Reasonator View with SQID ⟩ has part(s) (P527) View with SQID ⟨ Pano Amiantos (Q47004149)  View with Reasonator View with SQID ⟩

to show this. Is that correct? The new community is not split to two areas...

And how to show that Kato Amiantos (Q48413729) merge with Pano Amiantos (Q47004149) and create Amiantos (Q6377735)? I means in Kato Amiantos (Q48413729) item.

Xaris333 (talk) 19:01, 3 April 2018 (UTC)

Use replaced by (P1366) and replaces (P1365). Sjoerd de Bruin (talk) 19:07, 3 April 2018 (UTC)

OK.

⟨ Kato Amiantos (Q48413729)  View with Reasonator View with SQID ⟩ replaced by (P1366) View with SQID ⟨ Amiantos (Q6377735)  View with Reasonator View with SQID ⟩

But how to show that it was merged with Pano Amiantos (Q47004149)? Xaris333 (talk) 19:50, 3 April 2018 (UTC)

together with (P1706) as qualifier? Sjoerd de Bruin (talk) 20:42, 3 April 2018 (UTC)

Can I use dissolved, abolished or demolished date (P576)?

⟨ Kato Amiantos (Q48413729)  View with Reasonator View with SQID ⟩ dissolved, abolished or demolished date (P576) View with SQID ⟨ 2005 ⟩
has cause (P828) View with SQID ⟨ union (Q17853087)  View with Reasonator View with SQID ⟩
together with (P1706) View with SQID ⟨ Pano Amiantos (Q47004149)  View with Reasonator View with SQID ⟩
replaced by (P1366) View with SQID ⟨ Amiantos (Q6377735)  View with Reasonator View with SQID ⟩

Xaris333 (talk) 21:38, 3 April 2018 (UTC)

Notification of Rapid Grant proposal for Wikidata QuickSheets expansion for P172

FYI, I wanted to let you all know that I have submitted a Rapid Grant proposal to further develop software that will semi-automate the process of moving data from Wikipedia to Wikidata. The tool is designed to be accessible to those without programming experience by using simple article lists to generate spreadsheets for human evaluation. This builds on work done as part the Art+Feminism campaign, which is detailed at our workplan but this grant is not for A+F related work. I welcome your feedback on the proposal. --Theredproject (talk) 21:25, 3 April 2018 (UTC)

Are empty categories allowed on Wikidata?

Hi all

I'm working on a way to record datasets that are being and have been imported into Wikidata and it will involve some use of categories. However some categories will be empty when the tool is published and then be populated later, and could possibly at some points be empty again. Are empty categories allowed on Wikidata? Is there a way to signify that even though the categories are currently empty they will be used in future? If someone comes and deletes the categories thinking they are helping tidy up it will break the tool and this is something I would very much like to avoid.

Thanks

--John Cummings (talk) 12:35, 31 March 2018 (UTC)

Unless if they are registered in Special:TrackingCategories or as Babel categories, I  Strong oppose including them, because they are just e.g. "Category:Articles for deletion since 2009", who will ever dig histories of them via this site? --Liuxinyu970226 (talk) 13:22, 31 March 2018 (UTC)
@Liuxinyu970226: can you tell me how to register them in Special:TrackingCategories? It would be around 15 categories (including all sub categories) in total. --John Cummings (talk) 14:49, 31 March 2018 (UTC)
@John Cummings: That depends on developer level, you can see phab:T35033 (tracking task) on ideas of it. --Liuxinyu970226 (talk) 14:55, 31 March 2018 (UTC)
Liuxinyu970226's objection does not make much sense to me. "I oppose because they'll be eg Category:Articles for deletion since 2009 and who will look at them?" What? 1. They won't have such a name, and the name is largely irrelevant 2. no-one is being invited to dig through them: categories have other uses than being lists that people look through 3. no harm is adduced in your objection 4. we can make the assumption that JohnC is trying to do good things with them. What is not to like? Meanwhile they are not tracking categories - if I understand the rubric of that page correctly, nor are they Babel categories. Registering them as such seems pointless and damaging. I don't know the category system in wikidata well; I recall en.wiki has a "do not throw me away if empty" template designed for exactly the situation outlined by JC. So, perhaps I might ask for an articulation of credible issues that might arise from JC's proposal? --Tagishsimon (talk) 18:19, 31 March 2018 (UTC)
@Tagishsimon: Your "support" makes nonsense either, if a category page was just created by LTA e.g. Kagemusha (影武者), then how can investigation of their spam history must also likely to be a main work of this site? --Liuxinyu970226 (talk) 23:56, 31 March 2018 (UTC)
JC is talking about creating categories, and is not talking about crearing items denoting a category. "how can investigation of their spam history"... what on earth has this to do with the question in hand. You might as well ask, what if JC creates chocolate categories and they melt. We've established there is precedent for empty categories in wikidata, and have yet to hear any sensible or comprehensible issues from you. If you have anything practical to say, please do so. --Tagishsimon (talk) 00:02, 1 April 2018 (UTC)
  • I am the one who deletes most of the empty category items these days, mostly based on Wikidata:Database reports/to delete/empty category items. We have a couple of hundred new cases each week, due to Wikipedia category deletions that leave back empty category items here at Wikidata. Deletion of those items is not very complicated in any way, but it consumes quite some time if done properly.
    Since the most important purpose to have categories linked to Wikidata at all is to provide interwiki links, empty category items appear indeed somewhat unnecessary and I'd prefer not to start with such exceptions. If however the community now comes to agreement that we should accepte empty category items for a while, I could try to avoid running into them (provided I have a list of affected items) while working down the queue. However, these kinds of exceptions do not scale at all, and I'd probably stay away from category deletions if a second editor requested something similar.
    An alternative would be to create the categories already at Wikipedia and link them to the items; however, there is a comparable risk that some editor nominated them for deletion "because they are empty". Same problem technically, just that it had to be dealt with somewhere else... Not sure what to think of this :-) --MisterSynergy (talk) 20:35, 31 March 2018 (UTC)


@Tagishsimon: @MisterSynergy:, thanks very much, does the 'don't delete me, I'm useful' template (I think on Commons its called a maintenance category?) exist already?

So to explain what I want to use them for, I want to create an updated version of Wikidata:Data_Import_Hub where each dataset has its own page because it is becoming popular and very crowded. The pages for each dataset would be organised by using some broad categories around if the dataset import is 'in progress', 'needs help' or is 'complete' and then split by some broad topics (copying the Mix n' Match topics) and then a few categories for things like data type (e.g spreadsheet or list). The user can then explore the datasets by running searches for more than one category at a time e.g all the 'in progress' datasets about 'art'. These searches would all be premade using cirrussearch and the user would just click the button.

When the process first gets up and running a small number of the categories will be empty and its possible that some categories will be empty from time to time because there is nothing fitting that description e.g imports that need help. I just don't want categories to get deleted during this window which would break the system. In total the system would have maybe 30 categories and my guess is perhaps 2 or 3 would be empty from time to time.

There's a rough draft of the user interface at User:John_Cummings/Wikidata_Import_Hub_draft which may explain things a little better.

Thanks

--John Cummings (talk) 20:50, 31 March 2018 (UTC)

After reading your reply I am not sure whether I completely understood your opening comment; so in case my previous comment appears offtopic, don't worry any longer.
Do you speak about categories in Wikidata? If so, there is {{Administrative category}}, and there is Special:UnusedCategories to have an overview of currently empty categories. I don't think that there are many admins looking for those, but maybe it would be useful to secure them with a permanent marker, or a link to a page that explains why this category could be (temporarily) empty. --MisterSynergy (talk) 21:07, 31 March 2018 (UTC)
I hope you'll be bold and go ahead, John, with {{Administrative category}} and a pointer to an explanation of the cats as your friends. --Tagishsimon (talk) 21:55, 31 March 2018 (UTC)
Thanks @Tagishsimon: and @MisterSynergy: for taking the time to read my explanation and come up with a solution :) --John Cummings (talk) 21:58, 31 March 2018 (UTC)
@John Cummings: Wikidata:Property proposal/Wikidata focus list should also go live in a couple of days, and your items would seem natural candidates to mark with such a property. Jheald (talk) 23:56, 31 March 2018 (UTC)
@Jheald: 👍 --John Cummings (talk) 15:09, 4 April 2018 (UTC)

Standardizing historic patents

I work on historic patents and would like to standardize our ontology for them here. There are millions of them, and for those before 1900 no standard global source to search for them. Wikidata can do that. Is anyone else interested in working together on this?

Probably a patent item should have:

  • instance of patent (Q253623)
  • country where filed -- but not necessarily a currently-existing country; e.g. in the German states before 1870 there were patent systems. Presumably the WIPO code is appropriate to mark the country of a patent, but that would require some kind of property proposal. Property talk:P3068
  • discoverer or inventor (Property:P61) -- there can be multiple inventors
  • applicant(s) -- not always the same as the inventor(s), notably if a company filed the application, or the inventor died and someone else is filing
  • title -- a string
  • filing date is not the same as grant date; it would be good to store these distinctly because they have different historic meanings. The filing date is when the technological claim was made by the patent-filer, and the grant date is the date the government accepts and approves it officially. Both have legal importance.
  • Wikisource id should be used when a patent is transcribed there. it's infinite work but a good idea in some cases. Good text is not always available and Wikimedia can help.
  • patent number -- a Wikidata standard for this exists: (Property:P1246) -- however format is restrictive, requiring format like US891393. Usually this works but in some historical cases there are letters in the 'number' (1BB##### in 19th century France, though possibly the 1BB can be left out), and there may be cases in which there is no proper number, so this item is optional but will be there in >99% of cases. Note that patent number in this format does not include year and so is not unique e.g. in Great Britain numbers were reused year after year.
  • publication date, optionally -- historically this was sometimes later than grant date
  • A standard naming scheme for the items is desirable in those cases that they are not named distinctively, e.g. by the title. Something of this kind has been unambiguous in my research so far: Patent plus country plus year plus number, e.g. "Patent US 1906 821393"

-- econterms (talk) 04:28, 4 April 2018 (UTC) WikiProject ShEx has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. --Daniel Mietchen (talk) 13:01, 4 April 2018 (UTC)

Filter 41

This section was archived on a request by: Metrónomo (talk) 17:01, 10 April 2018 (UTC)

Is it possible to configure the abuse filter #41 so that it does not detect Korean characters? The words of a single character are common in Chinese, Korean and Japanese (Special:AbuseLog/4316266). --Metrónomo (talk) 05:58, 10 April 2018 (UTC)

It is. ✓ Done Matěj Suchánek (talk) 14:45, 10 April 2018 (UTC)
Thank you! --Metrónomo (talk) 17:01, 10 April 2018 (UTC)

Merge Q1684610 and Q6221722

I get the following error when I try to merge these two pages using Special:Merge: "Failed to merge items, please resolve any conflicts first. Error: Conflicting descriptions for language en." I don't see any guidance for dealing with this problem at Help:Merge. 1, can someone perform this merge. 2, if it is simple and useful, perhaps some guidance for this issue could be added either at Help:Merge or more detail in the error message at Special:Merge. Thank you, 16:51, 10 April 2018 (UTC)

Merged, I got no error. JAn Dudík (talk) 19:26, 10 April 2018 (UTC)
Thanks! Smmurphy (talk) 20:59, 10 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 14:53, 11 April 2018 (UTC)

Documentation of edit validation at input?

Several kinds of edits are rejected by the UI or the API, e.g. certain kinds of formats or values of dates are not allowed as values in Wikidata statements involving date properties. I am looking for documentation on how this works, for what kinds of formats/ edits etc. this exists. Thanks. --Daniel Mietchen (talk) 01:02, 4 April 2018 (UTC)

@Daniel Mietchen: We do have Help:Dates if that helps with your particular question? ArthurPSmith (talk) 16:18, 5 April 2018 (UTC)

What to do if a company sell different divisions to different companies?

Should I fill in followed by (P156) with the names of all the various buyer companies?
Or should I create one item for every department of the company and fill in the respective followed by (P156) properties with the organization that bought that division?--Malore (talk) 15:16, 5 April 2018 (UTC)

I think creating items for the divisions that were sold, if they pre-existed the sale, is a good idea. There may be other aspects of this to consider - for example you could create an item for the event itself, if it was notable enough. ArthurPSmith (talk) 16:16, 5 April 2018 (UTC)
If you mean subsidiary (Q658255) the best is to:
--Jklamo (talk) 17:42, 5 April 2018 (UTC)
@ArthurPSmith, Jklamo: Thank you for your precious advice. In my case, I'm working on E. Remington and Sons (Q3045750), which sold his typewriter division and then the rest of the organization to another company. Should I create a "E. Redmington and Sons typewriter division" item?--Malore (talk) 20:16, 5 April 2018 (UTC)

Merge needed

Q20816698 should be merged with Q79182. Sorry if I ask in the incorrect place. Żyrafał (talk) 20:30, 5 April 2018 (UTC)

Can't merge, they both have a different article on gawiki. And they are not about the same thing. The other is about an island and the second about a castle. Stryn (talk) 20:36, 5 April 2018 (UTC)
@Chaco: Thoughts? --Liuxinyu970226 (talk) 06:39, 6 April 2018 (UTC)

Individual tree

In General Sherman (Q152482), there is instance of (P31) = Sequoiadendron giganteum (Q149851), but it has a notice that indicates that there is some kind of problem with this statement. If we have an remarkable tree (Q811534), how do we correctly indicate the species of that tree? Thanks in advance, Paucabot (talk) 08:15, 5 April 2018 (UTC)

This looks fine to me. Perhaps we need to change the constraints on instance of (P31) to also allow for the taxonomic hierarchies? ArthurPSmith (talk) 16:49, 5 April 2018 (UTC)
I made Sequoiadendron giganteum (Q149851) a subclass of tree (Q10884), now it doesn't complain. It's no worse than Homo sapiens (Q15978631) being a subclass of omnivore (Q164509). Ghouston (talk) 05:15, 6 April 2018 (UTC)
Technically some humans aren't omnivores, and some Sequoiadendron giganteum (Q149851) (like seeds and saplings) aren't trees. Ghouston (talk) 05:18, 6 April 2018 (UTC)
There would be accurate items they could be sub-classed to, organism (Q7239) at least. Ghouston (talk) 05:23, 6 April 2018 (UTC)
I've been wondering about that too. It makes sense that target of instance of (P31) for every item should be a subclass of something, i.e. it uses subclass of (P279) (see here), and it makes sense that every individual organism is an instance of some taxon, which is a class of organisms. Taxon items however generally avoid using P279, they use parent taxon (P171) to build the class hierarchy. P279 could be used in addition to "parent taxon", but then for many cases one would seem redundant to another, e.g. why would we indicate every species in some genus as subclass of "tree" (or some other organism), while this applies to all species in this genus. I see that using P279 instead of P171 was considered in the past, but P171 stayed in use, in my opinion, judging by the RfD, without much of a good reason. While P171 is a well-defined property, I don't quite get it how is this different from any other classification that works out more or less fine using P279, including for cases when an item is part of multiple classifications. And property label ('subclass') being homonymous seems to be of little importance, say, if English label isn't homonymous then it may be in some other language or vice versa, and I assume users would generally check the description before applying a property that they are unfamiliar with. Perhaps, in light of this tree issue and data model model in general, using P171 should be finally reconsidered? 90.191.81.65 06:23, 6 April 2018 (UTC)
That discussion about P171 provides the right thing to set subclass of (P279) to: the item for the nearest parent taxon. So Homo sapiens (Q15978631) can be a subclass of Homo (Q171283) and Sequoiadendron giganteum (Q149851) a subclass of Sequoiadendron (Q14707792). That does make parent taxon (P171) seem redundant. You can check taxon rank (P105) to work out if a particular subclass or superclass is a taxon, or something random like omnivore (Q164509). Ghouston (talk) 06:52, 6 April 2018 (UTC)
Or check instance of (P31) taxon (Q16521). Ghouston (talk) 06:56, 6 April 2018 (UTC)

I would like to add our business information to WikiData

I could use some help with this. Still new to Wiki and reading tutorials but can't learn it over night. Any chance someone can help?

Bryanhill10 (talk) 18:06, 5 April 2018 (UTC)

@Bryanhill10: On what basis does your business meet our notability criteria? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:29, 5 April 2018 (UTC)
@Bryanhill10: Welcome to the wiki. If you want, feel free to can ask me whatever you want. I myself am not an expert - I've been contributing for about a month - but I remember at the beginning I was very confused.--Malore (talk) 08:57, 6 April 2018 (UTC)

Proper properties for literary awards to novels

While some literary awards are to an author for their entire body of work, many are specifically given to an author for a single book. In the latter case, which is the appropriate item to tag? Should the item for the author have award received (P166) together with the qualifier for work (P1686)? Or should it be the item for the book itself that has award received (P166)? Or should the two items (author and book) both have award received (P166)? Gabbe (talk) 11:53, 6 April 2018 (UTC)

Piano, per favore

Hello. I can not revise my edits all the time. Slow down, please. Compromises are always needed...

MinDat blocks, if you use it as a reference for valid minerals more than 2 times/day.
Misocam blocks [Mineralogical Society of America (MSA), Handbook of Minerals (HOM)], if you use it as a reference more than half a dozen times/day during ten days.
Nowadays, I use webmineral.com, rruff.info/ima, mineralienatlas.de and many things are implied on 'The IMA List of Minerals'.
Property 'as' got deprecated, but 'object has role' is not acceptable as a qualifier of 'instance of'. Many valid minerals were discovered after they got synthesised first.
All valid minerals need to be an 'instance of' 'mineral', otherwise 'chemical formula' is not acceptable.
Property 'named after' (eponym), qualifier 'subclass' is not acceptable. But Wikipedia labels are substantives, sometimes adjectives and verbs are needed too.
And when you have the paid trolls who try to rewrite history...
Just an observation, a warning shot. Regards --Chris.urs-o (talk) 13:52, 6 April 2018 (UTC)

Political activism?

How should one associate a person with a poltical campaign, movement or organisation if founded by (P112) or member of (P463) are not suitable? thanks MassiveEartha (talk) 16:49, 7 April 2018 (UTC)

@MassiveEartha: affiliation (P1416) exists for organisation; and it might be acceptable to use it for movements & political campaigns. --Tagishsimon (talk) 05:08, 8 April 2018 (UTC)

Military badges and insignia

I've sketched in a class tree between breast badge (Q799000) and Marine Aerial Navigator insignia (Q6763987). There is work needed in this area to build out all countries and all branches of service. There are many items and categories already in Wikidata, but they don't seem to have subclass statements or be linked to each other. It's possible cleaning this up could be automated using Wikipedia categories as a starting place. But that's not part of my skill set, and military insignia aren't my area of expertise or particular interest. So far I have not distinguished between the concepts "badge" and "insignia", since they seem to be interchangeable on Commons. My understanding is badge=physical object and insignia=symbol, but I don't know that the military use the terms in this way. Anyway, does anyone want to take on expanding this structure? - PKM (talk) 19:39, 7 April 2018 (UTC)

Add new data to Wikidata

I would like to add the term "WoeKO" into the database. I have the history of the individual and links. How can this process work?  – The preceding unsigned comment was added by Woeko (talk • contribs) at 8. 4. 2018, 05:49‎ (UTC).

@Woeko : please check out Wikidata:Notability and let us know if "WoeKO" meets the test. If so, an item can be added via the "create a new item" left-side menu link. --Tagishsimon (talk) 06:14, 8 April 2018 (UTC)

Should website and company be separated?

There are a number of items like Internet Archive (Q461) and YouTube (Q866) that represent both the website and the organization. Should this items be split?--Malore (talk) 20:08, 2 April 2018 (UTC)

Yes. An interesting conundrum is which one the existing item should be used for. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:10, 2 April 2018 (UTC)
We also have the same problem with streaming services like Netflix (Q907311), Hulu (Q1630304), Amazon Prime Video (Q4740856) or Crunchyroll (Q1142035). --Kam Solusar (talk) 12:03, 3 April 2018 (UTC)
Definitely yes. Some items were already splitted (like Twitter (Q918) vs Twitter, Inc. (Q1390577), Google (Q95) vs Google Search (Q9366) or Facebook (Q355) vs Meta Platforms (Q380)). But it is not easy to determine where sitelinks belong, as most of the sites are covering both concepts (website/service and company). In a case of Twitter (Q918) and Facebook (Q355) most of sitelinks are on "service" item, while for Google (Q95) on "company" item.--Jklamo (talk) 12:39, 3 April 2018 (UTC)
@Pigsonthewing, Kam Solusar, Jklamo:I noted a problem: if both the items have an official website (P856) property linking to the same URL, there is a property constraint violation. What to do?
Create a new property (renaming official website (P856) in "official website URL") that links to the item representing the official website instead of the website itself?
Or add statement is subject of (P805) filled with the item representing the website as value of the official website (P856) property inside the company item and change the property constraint so that it is allowed more than one item linking to the same URL, but only one without a statement is subject of (P805) qualifier? --Malore (talk) 11:43, 5 April 2018 (UTC)
I would say yes as YouTube, LLC. And YouTube.com are two very different things, in many cases the company could also have separate products unrelated to the service that gave the company its name or vice versa (E.G. Google mentioned above). -- 徵國單  (討論 🀄) (方孔錢 💴) 09:09, 8 April 2018 (UTC)

How can I express search results as an external link?

Hi all

I'm trying to create something that uses links to search results with include and exclude the results of multiple categories, is it possible to express these links as internal links? E.g

https://www.wikidata.org/w/index.php?search=&search=incategory%3A%22Archives+datasets%22+incategory%3A%22Datasets+by+progress%22&title=Special:Search&go=Go&searchToken=4on3a24isms4ii04x8nmz7zqj

Thanks

John Cummings (talk) 15:40, 4 April 2018 (UTC)

Given the complexity of such a linking, I would say no. But a template can help with it. Matěj Suchánek (talk) 07:21, 5 April 2018 (UTC)
@Matěj Suchánek:, do you know if this template exists on any Wikimedia projects? I think I can make it work with just the links but it will complex for other people to edit. --John Cummings (talk) 13:00, 5 April 2018 (UTC)
Something like w:en:Template:Search link. But it will likely need to be enhanced. Matěj Suchánek (talk) 14:42, 5 April 2018 (UTC)
Thanks @Matěj Suchánek:, --John Cummings (talk) 13:50, 8 April 2018 (UTC)

Prime factors?

I wonder whether there is any appropriate way to represent prime factors of a number in Wikidata, e.g., 89 681 and 96 079 for Jevons' number (Q1100748) or 37 975 227 936 943 673 922 808 872 755 445 627 854 565 536 638 199 and 40 094 690 950 920 881 030 683 735 292 761 468 389 214 899 724 061 for RSA-100 (Q15990637). Alternatively, maybe a property can be created for the identifier in the factordb (Q1391719)? -- IvanP (talk) 10:18, 8 April 2018 (UTC)

See: Wikidata:Property proposal/KIT Linked Open Numbers ID. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:46, 8 April 2018 (UTC)
You know that this project (Linked Open Numbers) is an April fools joke? But you still would like Wikidata to link to it? It does not even provide a prime factorization to all composites in it and 1 is displayed in prime decompositions (e.g., “Prime factors: 1 * 2 * 3 * 7” for n42, though in Semantic forms and Graphite, we only see n2, n3 and n7). -- IvanP (talk) 12:09, 8 April 2018 (UTC)

Fashion or style "based on" another

I have sources that state that the homburg (Q697666) is "based on" the Tyrolean hat (Q694050). (I have also run into this for other types of clothing.) Using based on (P144) for this relationship throws a constraint violation since clothing and its subclasses aren't currently works. I think the best way to resolve this is to make clothing (Q11460) a subclass of work (Q386724). Thoughts? - PKM (talk) 01:55, 6 April 2018 (UTC)

 Support Sounds fine to me. --Marsupium (talk) 21:06, 6 April 2018 (UTC)
✓ Done - PKM (talk) 22:51, 8 April 2018 (UTC)

chiarimenti su voce giambattista callegari

buon giorno, poche settimane fa ho cercato di aggiungare la Voce Giambattista Callegari, inventore che ha realizzato dei circuiti oscillanti nel campo della radiotecnica radio biologica, quindi attinenti la fisica moderna ... di lui diversi libri, studi scientifici, tesi universitarie, citazioni e riconoscimenti pubblici ... tuttavia è stato negato per mancanza di interesse enciclopedico, ... non sono molto esperto con wikipedia e devo aver commesso errori. una parte del testo l'ho preso dal sito di "RadionicaCallegari", ma dispongo di tutte le autorizzazioni. non ho capito l'errore ed ho letto il regolamento. grazie tante Piediperlaterra  – The preceding unsigned comment was added by Piediperlaterra (talk • contribs) at 09:13, 6 April 2018‎ (UTC).

Google translates the above as:

good morning, a few weeks ago I tried to add the Voice Giambattista Callegari, inventor who has made oscillating circuits in the field of radio radiotechnics biological, so relevant to modern physics ... of him several books, scientific studies, university theses, citations and public acknowledgments ... however it has been denied due to lack of encyclopedic interest, ... I am not very expert with wikipedia and I must have made mistakes. I took part of the text from the site of "RadionicaCallegari", but I have all the permissions. I did not understand the error and I read the rules. Thank you very much

-- Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:51, 6 April 2018 (UTC)

@Epìdosis, Rippitippi, Sannita, ValterVB:, as our it-N admins who can respond to this query. Mahir256 (talk) 22:05, 8 April 2018 (UTC)
@Piediperlaterra: Questa è Wikidata, non Wikipedia, leggi la tua pagina di discussione su Wikipedia. Problemi di copyright e di enciclopedicità. Thanks @Pigsonthewing: It's a problem on deleted page on it.wikipedia. --ValterVB (talk) 06:20, 9 April 2018 (UTC)

Decorated by

How can I show that an interior have been decorated by an artist?. One possibillity is to use the same solution as done with the Sistine Chapel (Q2943) where the item Sistine Chapel ceiling (Q844675) has been created as an item with Michelangelo (Q5592) as creator (P170). But doing that I would have to create an new interior item for each building. Pmt (talk) 17:25, 8 April 2018 (UTC)

Is this out of the question? "Of" is designed as a "qualifier stating that a statement applies within the scope of a particular item".
⟨ Sistine Chapel (Q2943)  View with Reasonator View with SQID ⟩ creator (P170) View with SQID ⟨ Michelangelo (Q5592)  View with Reasonator View with SQID ⟩
of (P642) View with SQID ⟨ interior (Q2998430)  View with Reasonator View with SQID ⟩
--Tagishsimon (talk) 19:38, 8 April 2018 (UTC)
Could we use the qualifier applies to part (P518) here: <creator> [name] <applies to part> interior decoration (Q5875045)? Or possibly significant person (P3342) and object has role (P3831): <significant person> [name] <object has role> interior designer (Q2133309)? - PKM (talk) 20:39, 8 April 2018 (UTC)
That's better ... --Tagishsimon (talk) 20:46, 8 April 2018 (UTC)

Thanks Pmt (talk) 21:23, 8 April 2018 (UTC)

That formulation gives us a property constraint issue on creator (P170) not allowing applies to part (P518) as a qualifier. I'm inclined to argue for the constraint to change. --Tagishsimon (talk) 22:00, 8 April 2018 (UTC)

 Support Changing that constraint. - PKM (talk) 22:46, 8 April 2018 (UTC)
And changing the constraint would also allow for paintings where the figures are by one artist and the background landscape is by another. - PKM (talk) 23:44, 8 April 2018 (UTC)
 Support Change needed Pmt (talk) 23:53, 8 April 2018 (UTC)

Is there a limit to the number of references?

Is it possible to have too many references for a statement? I ask it because I was thinking about searching for all information available in a given publication or book - in particular textbooks - in order to use it in a Spaced Repetition Software like Anki. If it would be possible to add as many references as possible, we could ideally represent in Wikidata all the information available in a textbook and export it in a flashcard software if someone need to study that textbook.--Malore (talk) 21:37, 8 April 2018 (UTC)

My perception is that the purpose of references is to help readers and other data consumers to verify that a statement is true, or at least, that the statement correctly reports the position of the referenced source. Creating large numbers of references for a single statement has at least two problems.
  1. It is hard for the reader to wade through a long list of references and decide which ones are the best.
  2. It is the responsibility of the editor adding a reference to actually read the reference and be sure that it actually supports the statement in Wikidata. This is a lot of work. Any editor adding massive numbers of references to a single statement in rapid succession will be suspected of failing to meet this responsibility. Jc3s5h (talk) 14:34, 9 April 2018 (UTC)

Wikidata:Requests for comment/Privacy and Living People

Additional proposals have been made on the RFC, and your comments are welcome. --MediaWiki message delivery (talk) 03:40, 9 April 2018 (UTC) (for Rschen7754)

I had to hunt around for these new proposals, which are not the last in sequence. They were made in this edit. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:19, 9 April 2018 (UTC)

How do I nominate a category for deletion?

Hi all

How do I nominate a category for deletion? I made a duplicate category by accident (Category:Webpage datasets) and I can't find instructions and none of the Commons templates work.

Thanks

--John Cummings (talk) 10:34, 9 April 2018 (UTC)

As simple as {{Delete}}. --Edgars2007 (talk) 10:47, 9 April 2018 (UTC)
Thanks very much @Edgars2007:, --John Cummings (talk) 13:41, 9 April 2018 (UTC)

Wikidata weekly summary #307

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • Differences between range constraints now work only if both properties are present (phab:T185480)
    • Work on a new constraint unit which defines which units are alowed for a given property (phab:T164372)
    • Make the Special:ConstraintReport easier to understand (phab:T188728)
    • Work on a Lua function which checks if an item is a subclass of or instance of another item (phab:T179155)
    • Started deleting old autopatrolled from the loggin table (phab:T190447, phab:T184791)
    • More work on setting up permissions for Lexeme (phab:T189480)
    • Fixing a bug about removing the last Lemma (phab:T189185)
    • Disabling the possibility to move a Lexeme (phab:T189481)

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/where to search through Category:Pages using Wikidata property P#

I found Category:Pages using Wikidata property P566 (Q28039488), now I want to look at all the pages that use P566. WD says it's a Wikimedia category, but it doesn't exist on Wikimedia nor on Wikipedia. — Tom.Reding (talk) 14:46, 9 April 2018 (UTC)

@Tom.Reding: The category represented by the wikidata item is on the ar.wiki here ... quite what or why, I don't know. You can view the items having P566 as a property with the query below; and Property talk:P566 provides some other ways to view items & counts. (The report times out if I try to fetch the values against each P566 ... might try to fix that later.
SELECT ?item ?itemLabel
WHERE {
  ?item wdt:P566 [].
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}
order by ?itemLabel
Try it!
hth --Tagishsimon (talk) 15:10, 9 April 2018 (UTC)
The query that Tom wants and which works (at least for the first 100) is given at en:Template talk:Taxonbar#Basionyms and synonyms. Peter coxhead (talk) 15:41, 9 April 2018 (UTC)
Yup. Avoiding the timeout when returning all values is the problem. This seems to work; download the results into a spreadsheet if you want to sort them:
SELECT ?item ?itemLabel ?value ?valueLabel
WHERE {
  ?item wdt:P566 ?value.
  SERVICE wikibase:label { bd:serviceParam wikibase:language "en" . }
}
Try it!
--Tagishsimon (talk) 16:30, 9 April 2018 (UTC)

w:User:Charles Matthews/Facto Post/Issue 11 – 9 April 2018

The latest issue of the Facto Post newsletter, which has things to say about citations, but more about open access downloading. Charles Matthews (talk) 16:35, 9 April 2018 (UTC)

How to behave if a product or company is renamed?

For example, Mozilla Firefox (Q698) born as Phoenix (Q3327049), was later renamed in Mozilla Firebird (Q3327045) and, finally, became Firefox. What is the best approach to manage this information? Mozilla Firefox (Q698) should represent the follower of Mozilla Firebird (Q3327045) or all the versions starting from Mozilla Firefox (Q698)?--Malore (talk) 14:03, 9 April 2018 (UTC)

Use official name with start and end qualifiers. Snipre (talk) 05:41, 10 April 2018 (UTC)

Move request - Wikidata Query Service

I am requesting a page move at Wikidata_talk:SPARQL_query_service#SPARQL_query_service_versus_Wikidata_Query.

I requested a move at the admin board at Wikidata:Administrators'_noticeboard#Move_request_-_Wikidata_Query_Service. I got the suggestion to seek further community comment. The process for organizing a pagemove here is uncertain to me, but I think posting here to seek comments works. I would appreciate any comments on that user page about the move. Blue Rasberry (talk) 19:37, 9 April 2018 (UTC)

Labels and desrciptions

Hello. Vasa Kellakiou (Q7916291) and Vasa Koilaniou (Q7916293). 2 different villages with the same name. People added a second word after their name and we use to use that words even today. But their official name is the same. I was trying to add the official name in English (and Greek) with the same description but the system don't let me. I understand that in same cases when two items have the same label and same description maybe the same item. But in that case they are not. I had to change the description of one of the item. Xaris333 (talk) 23:04, 9 April 2018 (UTC)

It is the case that "no two items may have both the same label and the same description" per Help:Label so amending the description is indeed the way to go. --Tagishsimon (talk) 23:25, 9 April 2018 (UTC)
It seems useless to me. I just add Cyprus to one and Republic of Cyprus to the other. Same thing. I mean that the description its self may not be the way to show the different between them. Xaris333 (talk) 00:08, 10 April 2018 (UTC)
@Xaris333: We routinely use common names instead of official names in labels; "India" instead of "Republic of India", "Hamlet" instead of "The Tragedy of Hamlet, Prince of Denmark", and so on. It makes sense to me to put "Vasa Kellakiou" and "Vasa Koilaniou" as the English labels, each with an alias "Vasa", and within each item leave the official name (P1448) as you have left it. Mahir256 (talk) 00:12, 10 April 2018 (UTC)
It would also be possible to use geographic descriptions, like "village and communal council in the west of Limassol District, Cyprus" vs the one in the east. Ghouston (talk) 00:19, 10 April 2018 (UTC)
It is not useless. It serves to prevent duplicate items being created. The description is expressely "a short phrase designed to disambiguate items with the same or similar labels" (see Help:Description) and thus it is incumbent on us to concoct descriptions that effect that disambiguation - as Ghouston's example does. It's deliberately not okay to to seek a same label same description, and force the user to rely on other claims to disambuguate, since only the label and description are supplied in search results and so discovery of items by search is frustrated. --Tagishsimon (talk) 01:22, 10 April 2018 (UTC)

human (Q5) violates subclass constraint

On human (Q5), the "subclass of: person" statement has an issue attached that I found rather confusing. It says "conflicts-with constraint: An entity should not have a statement for subclass of if it also has a statement for instance of (P31) with value common name (Q502895)." --Krinkle (talk) 02:26, 10 April 2018 (UTC)

I read this as saying, inter alia, that because "human" is a common name, it is the thing that it is a common name of (i.e. homo sapiens) that should have the subclass of person claim, and not the common name item. --Tagishsimon (talk) 04:38, 10 April 2018 (UTC)
Such issues have been discussed previously on Talk:Q5 and elsewhere, but it's hard to convince anybody to change the status quo. Ghouston (talk) 05:10, 10 April 2018 (UTC)

anonymous (Q4233718) violates gender constraint

At anonymous (Q4233718), the "sex or gender: unknown value" statements shows the following issue: "conflicts-with constraint: An entity should not have statements for both sex or gender (P21) and subclass of (P279)}}." Can someone explain why that constraint exists? And how should we solve it? --Krinkle (talk) 02:25, 10 April 2018 (UTC)

Because sex or gender applies to an individual thing, which will be an instance of something, not a subclass of something. How to fix? anonymous (Q4233718) is quite sucky right now. The en description somewhat nails it: the item is a term (Q1969448) which represents an unsupplied creator name. The item is not a human, nor is it a subclass of a humen, creator, etc etc. It is a term. The item should not, therefore, have lots of 'unknown values' for a series of attributes of humans. It should instead have claims that are suited to the description of the term, such as
⟨ anonymous (Q4233718)  View with Reasonator View with SQID ⟩ subclass of (P279) View with SQID ⟨ pseudonym (Q61002)  View with Reasonator View with SQID ⟩
. --Tagishsimon (talk) 04:04, 10 April 2018 (UTC)
If that was changed it would probably introduce other constraint violations when it's used for an author etc., like "author should have a gender". Ghouston (talk) 05:15, 10 April 2018 (UTC)
Agreed. Though a quick look at items with a P50 of Q4233718 suggests that many should have P50='unknown value' and a very few should have a pseudonym (P742) of Q4233718. --Tagishsimon (talk) 05:4710 April 2018 (UTC)
This is a particularly egregious error, because it is quite possible that the creator of a specific work is both anonymous and known to be male, or known to be female. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 08:28, 10 April 2018 (UTC)

Japan properties

I have just created {{Japan properties}}. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:39, 10 April 2018 (UTC)

Help with Czech

Can a Czech speaker help me understand the difference between tow (Q18193532) and tow (Q14789578)? Both have the same English translation "tow" in their cswiki articles, and both seem to be about the bundles of fibers used in machine spinning. I can't tell how they are different. Thanks! - PKM (talk) 02:03, 15 April 2018 (UTC)

(cs-N) They look like duplicates to me, proposed them to merge. Matěj Suchánek (talk) 08:43, 15 April 2018 (UTC)
Thank you very much! - PKM (talk) 19:40, 15 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 07:25, 16 April 2018 (UTC)

Merging problem

I'm trying to merge the two wikidata pages on When Calls the Heart that refer to the same TV series, but I keep getting an error message (A conflict detected on enwiki: Q16886768 with enwiki:When Calls the Heart, Q15728550 with enwiki:When Calls the Heart (TV series)), presumedly because the en-wiki page for one of them redirects to the other one. I'm not sure what to do to fix it, so I'd appreciate any and all help. Thanks. -Yupik (talk) 19:31, 15 April 2018 (UTC)

@Yupik: If there's a redirect and you need to merge the items, the best approach is to remove the sitelink for the redirect and then merge - this should go through without problems and leave you with one item which has the correct sitelink. Andrew Gray (talk) 19:33, 15 April 2018 (UTC)
Sorted. Was a link to an en redirect lingering on one of the items. Now merged. --Tagishsimon (talk) 19:35, 15 April 2018 (UTC)
4 minutes. Not too shabby. --Tagishsimon (talk) 19:36, 15 April 2018 (UTC)
That was quick :D Thanks! -Yupik (talk) 19:37, 15 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 07:23, 16 April 2018 (UTC)

Translate

  • Near statements a magnifying glass is showing. The word more is showing when the cursor is on it. I want to translate that but I can't find it in translatewiki.net.
  • I also want to translate Query Service, Concept URI, Mark as duplicate, VIP's labels, Browse Primary Sources, ▲ back to top ▲, Random Primary Sources item, Primary Sources list, Check sitelink from the left column of Wikidata. (Maybe some are from gadgets).

Xaris333 (talk) 02:05, 6 April 2018 (UTC)

@Xaris333: here is the way to translate WQS. --Liuxinyu970226 (talk) 11:55, 6 April 2018 (UTC)
Thanks. But that didn't help me. Xaris333 (talk) 15:10, 6 April 2018 (UTC)
May I help you, then? What's up? Matěj Suchánek (talk) 18:01, 6 April 2018 (UTC)
@Xaris333: The sources for those phrases seem to be here:
The first two seem to have built-in i18n, the later two not. Good luck with providing translations! --Marsupium (talk) 22:31, 10 April 2018 (UTC)
Thanks! Xaris333 (talk) 23:11, 10 April 2018 (UTC)

More than 50% of my edits get reverted. And "Days of The Week" are separated into 2 items; undiscussed.

  1. Everything starts with one day I find that the Chinese, English, Deutsch version of the "Days of the Week" are not connected. 42 languages are in one group and 11 languages are in the other group, and I found no discussion about the split in Wikidata or the English Wiki, so I merged them. And then I got reverted by User:Andreasmperu. So I merged them with a message "No one single language separates this exact same topic. No separation discussion found in wikidata or english wiki of the topic 'Days of the week'", and leave a comment on the Talk:Q41825 "Separation will create unnecessary difficulties to compare difficult languages of the 'Days in the Week'.". And then I got reverted by the same admin again with just two words "wrong item" with no reply on the discussion page. [9] [10]
  2. As a result, more than 50% of my edits in Wikidata get reverted without a reason. If you do Google translate, you will know these two are the same topic "a ship destroyed and sank".
  3. G translate will tell you that these two(w:en:Chuanyue) are about the same topic of "Time Travel Fiction".
  4. Click in to look at the map and coordination and you know one is about the same town, and the other is not a town.
  5. You don't even need Google translate to know these two are the exact same group of Baptist Universities. Just spend 5 seconds to click in.

What an irony that an IP user is more willing to explain and communicate than an administrator. 118.143.147.130 22:01, 9 April 2018 (UTC)

In the first case, you seem to have conflated "Names of the days of the week" with "Day of the week". That's an error, since those are different concepts, and was rightly reverted. Similarly, not all time travel in fiction is a time travel novel. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:24, 9 April 2018 (UTC)
Per Andy, you in fact get 5/5 wrong.
  1. Two different concepts, names of days of the week versus day of the week.
  2. Disambiguation page versus article
  3. As noted, not all time travel in fiction is a time travel novel; different concepts.
  4. zh.wiki has articles attached to both items. You'd need to sort that out before they could be merged.
  5. Disambiguation page versus list article. You can argue that zh.wiki should change their article to a list, but until they do, it's a dab, and we need to reflect that.
So. Sorry about that. --Tagishsimon (talk) 22:54, 9 April 2018 (UTC)
  1. The English Wiki "Names of the days of the week" is actually the same article moved from "Days of the week". What topic of contents is different before the en-wiki article moved from "Days of the Week" and after moved to the current "w:Names of the days of the week"? They are still all about the exact same topic: how to count the seven days, their astronomy origins and their names in different languages. "No one single language Wiki separates this exact same topic into two articles" is the strongest evidence that none of the 53 languages thinks that they are about different topics.
  2. I don't think good idea to separate the exact same topic of different languages simply because one has less content. This will break the cross-reference and cross-improvement connections among different languages.
  3. Of course, fiction includes novels and films. It is wrong to say that "小說" in Chinese only includes written books. 小說 does include films and other forms of art, so 穿越小說 is same as "Time travel in fiction" and both of them include "Time travel novel". Google translate tells you the same 小說=Fiction Fiction=小說.
  4. I wonder whether you guys did click in for 5 seconds. w:zh:金字牌鎮 and w:sv:Jinzipai (köpinghuvudort i Kina, Anhui Sheng, lat 29,85, long 117,81) are the same location at (29.85 N, 117.81 E). w:zh:金字牌 is a "message certificate of royal orders", which is definitely not a location.
  5. Bad idea to separate the exact same topic of different languages; not to mention even their content are the same. So if I change the w:Baptist University to disambiguation, this one is done? 118.143.147.130 12:24, 10 April 2018 (UTC)
  1. (All IMO...) This one is something of a clusterfuck. The two wikidata items are distinct and for that reason should not be merged. However I concede that there are claims and wikilinks on day of the week (Q41825) that belong instead on names of the days of the week (Q42962347) or on another item (e.g. the GND ID seems to point to a concept of 'Working Days'). I concede, too, that not much may be left on day of the week (Q41825) by the time we finish.
  2. These remain two distinct items, one pointing to articles on concept of shipwreck/ing, the other a zh disambiguation page. zh's article on shipwrecks was pointed at the wikidata item maritime disaster (Q2620513) ... I've moved it to shipwrecking (Q906512). I think that's now sorted?
  3. So for this one ... chuanyue (Q10453828) is a distinct subclass of time travel in fiction (Q253732) and the two should not be merged. The zh. wikilink from chuanyue (Q10453828) looks appropriate; is about tt novels. The en. wikilink was to an article on Chuanyue which is a distinct subclass of chuanyue (Q10453828) ... I've created a new item, chuanyue (Q51733250) and moved the en. wikilink to it. I'm hoping we're good on this now.
  4. This one should be sorted now.
  5. Yes; I've done that. With luck, also sorted.
More generally, you'll not get much dispute that there is a lot wrong or that can be improved in many wikidata items. Exactly how to effect improvement can be a more thorny issue. --Tagishsimon (talk) 00:47, 11 April 2018 (UTC)
@Tagishsimon: I'm afraid that I have to bump WD:RFP/R#Andreasmperu again here. --Liuxinyu970226 (talk) 10:11, 11 April 2018 (UTC)

Linking templates to categories and lists without a main topic item

I'm wondering how to link an item for a list-like navigational template to the corresponding list and category items, when there is no corresponding main topic item. A specific example is Template:Major National Historical and Cultural Sites (Jiangsu) (Q20691005), which corresponds to the list Major National Historical and Cultural Sites (Jiangsu) (Q1188621) and category Category:Major National Historical and Cultural Sites in Jiangsu (Q17392112). The category and list can be linked to each other without using a main topic, using category related to list (P1754) and list related to category (P1753). Are there similar properties for templates, and should there be? Rigadoun (talk) 14:32, 10 April 2018 (UTC)

This seems like a good use case for a ‘template combines topics’ property (by analogy with category combines topics (P971)). Mahir256 (talk) 14:43, 10 April 2018 (UTC)
Agree. --Edgars2007 (talk) 10:33, 11 April 2018 (UTC)
template has topic (P1423) and topic's main template (P1424) are what you're looking for, with the list as the main topic. Thanks. Mike Peel (talk) 14:32, 11 April 2018 (UTC)

Organizations with more than one official website

official website (P856) says there can only be one per item but some organizations have two (or more) official domains. For example, canoe.ca (Q902502) (Canadian Online Explorer) has two portals: canoe.ca (French) and canoe.com (English). Caras (Q9696470) has three portals: caras.perfil.com (Spanish), caras.uol.br (Portuguese of Brazil) and caras.sapo.pt (Portuguese of Portugal). --Metrónomo (talk) 08:25, 11 April 2018 (UTC)

Note: be careful when reading Caras (Q9696470) because I am still developing it, I need to investigate a little more. The information presented in the Wikipedia in English and Portuguese is partially true and incomplete. Here (page 38) say that Caras was founded by Grupo Perfil in Buenos Aires in 1992 and it started as a spanish magazine. Then it began to be published in Brazil (1993), Portugal (1995), Angola (2004) and Uruguay (2007). In 1993, Grupo Perfil was associated with Grupo Abril and in 1995 it was associated with Grupo Impresa. In 2014, Grupo Perfil bought the part from Grupo Abril and today is the only owner of the franchise (Argentina, Brasil, Angola and Uruguay) except for Caras Portugal who continues to share with Grupo Impresa. I'm not quite sure how to reflect that on Wikidata (co-owners or local partners). --Metrónomo (talk) 08:46, 11 April 2018 (UTC)
Pick the main one like how it's done on Google Search (Q9366) otherwise you'll end up with long lists. Multichill (talk) 09:56, 11 April 2018 (UTC)
In the case of Canoë, none is the main one (French/English duality in Canada). In the case of Caras can be considered (Spanish?) Even if you choose the one you choose, someone will always disagree (Portuguese speakers will say: why Spanish?) The difference with Google is that we all agree which is the main or most representative site. --Metrónomo (talk) 10:38, 11 April 2018 (UTC)
A discussion from 2016 at Property_talk:P856 suggests allowing multiple values if they are qualified somehow, like language of work or name (P407), applies to jurisdiction (P1001) or intended public (P2360). I'm not sure if doing that fixes the constraint. Ghouston (talk) 11:41, 11 April 2018 (UTC)
The constraint statement already has separator (P4155) parameters, which should allow multiple values as long as they have different qualifiers for e. g. language of work or name (P407) – but unfortunately we haven’t implemented support for that parameter in the WikibaseQualityConstraints extension yet, see phabricator:T173594. --Lucas Werkmeister (WMDE) (talk) 13:18, 11 April 2018 (UTC)

Formatter URLs for GET request method

Data from an ORCID record can be fetched using the relevant ORCID iD (P496) value and HTTP GET (Q30542074), thus:

https://pub.orcid.org/v2.0/0000-0001-5882-6823/person

I have stored that here using formatter URI for RDF resource (P1921) - is there a better way to record this? As formatter URL (P1630)? Do we need a new property, "GET formatter URL"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:15, 11 April 2018 (UTC)

@Pigsonthewing: I don’t think that’s correct – an URL with a v2.0 version number in it doesn’t look very much like a canonical URI to me.
And I don’t think there’s any need to store this URL separately, in fact – if you request the real canonical URI, e. g. https://orcid.org/0000-0001-5882-6823, with the right Accept header (e. g. Accept: application/xml), it redirects you to the pub.orcid.org/v2.0/… URL. We do the same thing in Wikidata – see WD:Data access#Linked Data interface. --Lucas Werkmeister (talk) 15:30, 11 April 2018 (UTC)

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

It is true that:

https://pub.orcid.org/0000-0001-5882-6823/person

will also work. However my question is how we can best store such URLs for cases where your solution either does not work, or or is not available to the user. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:55, 11 April 2018 (UTC)

Improvements on the search results

Hello all,

Some improvements have been made by the Search team at WMF on Special:Search (the page that is loaded when you type something in the search field and press enter, for example).

  • The string that you typed is highlighted in bold in the label of the results
  • The number of bytes is no longer displayed, instead you can see the number of sitelinks and statements
  • The code is now using the same indexing as completion search, which should improve relevancy and better control over ranking

If you have any question or find a bug, feel free to ping user:Smalyshev (WMF). Lea Lacroix (WMDE) (talk) 07:33, 5 April 2018 (UTC)

Good improvement. Thanks! Paucabot (talk) 08:10, 5 April 2018 (UTC)
Good news, I like the new display. BTW, what happened with the related birthday present from last year (Wikidata:Status_updates/2017_10_30#Wikidata's_birthday).
--- Jura 09:24, 5 April 2018 (UTC)
@Jura1: Which one do you mean? All the search-related ones are still there. Smalyshev (WMF) (talk) 18:08, 5 April 2018 (UTC)
@Smalyshev (WMF): The first one. Somehow I thought it was about all items, but apparently it's just for those with Q4167410 (and Q13442814 added later).
--- Jura 07:18, 6 April 2018 (UTC)
Great! Hmm, when I search "US" (capitalization doesn't matter, probably), in completion search I see United States of America (Q30) (which is expected), but in Special:Search the country is outside of first page. --Edgars2007 (talk) 07:24, 6 April 2018 (UTC)
For me, searching for "US" ranks United States of America (Q30) first in both contexts. But for "USA", United States of America (Q30) only comes out second in Special:Search. I would be interested in understanding the difference between the two endpoints. − Pintoch (talk) 08:56, 8 April 2018 (UTC)
@Smalyshev (WMF): is there a way to get the Special:Search results via the API? It seems to me that the "srsearch" action still uses the old index (example: https://www.wikidata.org/w/api.php?action=query&list=search&srsearch=united+states&srnamespace=0 ranks United States of America (Q30) in 7th position instead of 1st for Special:Search). − Pintoch (talk) 19:30, 12 April 2018 (UTC)
This is strange, the code should use the same code path. I'll check that. Smalyshev (WMF) (talk) 19:33, 12 April 2018 (UTC)

Property talk:P2281

Properity has reported issue with links since 2015. Anyone could fix this?  – The preceding unsigned comment was added by Eurohunter (talk • contribs) at 6. 4. 2018, 18:29‎ (UTC).

Please add an explanation there how you'd think this should be done. In the meantime, you may de-activate or ignore the formatter url.
--- Jura 09:26, 7 April 2018 (UTC)
@Jura1: In current form it force us to linking American version of iTunes via https://itunes.apple.com/us/. There should be option to link to other versions of iTunes like https://itunes.apple.com/fr/ https://itunes.apple.com/se/ and other. Eurohunter (talk) 19:56, 11 April 2018 (UTC)

Revamping the list of properties

The list of properties is out of date. Not only that, some subpages like this one show errors. I was wondering if someone has the technical skills to write a bot that could automate the list based on the classification of the property tree which is always updated. Said bot could either write listeria queries for each branch or use {{List of properties/Row}}. What do you think? What would be the best way to do it? Micru (talk) 18:12, 3 April 2018 (UTC)

Wouldn't a category, or category tree, be better? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:30, 4 April 2018 (UTC)
@Pigsonthewing: How would you categorize the properties? AFAIK there is no way to add the properties to any category, unless you put it on the talk page, which is less than ideal... Micru (talk) 20:06, 4 April 2018 (UTC)
Yes, I was thinking of the talk pages. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:53, 5 April 2018 (UTC)

@Pigsonthewing, Micru: - For lists there is listeria. Re-invent the wheel? Not invented here syndrom? 92.231.172.31 15:06, 12 April 2018 (UTC)

If you read my initial comment you will see that I do mention listeria. Anyway, recently a new tool has been introduced that seems to be good enough: https://tools.wmflabs.org/prop-explorer/ Micru (talk) 06:53, 13 April 2018 (UTC)

P2600 geni.com profile ID - Wikidata does store ID plus irrelevant URL part

Now copied to Property talk:P2600#P2600 geni.com profile ID - Wikidata does store ID plus irrelevant URL part. 92.231.172.31 15:20, 12 April 2018 (UTC)

Removed from here as duplicated discussion continues there. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:56, 13 April 2018 (UTC)

Novel

Hello, "novel" is "genre" or "instance of"? Thanks. --Titodutta (talk) 01:47, 11 April 2018 (UTC)

We have
⟨ novel (Q8261)  View with Reasonator View with SQID ⟩ instance of (P31) View with SQID ⟨ literary genre (Q223393)  View with Reasonator View with SQID ⟩
if that helps. --Tagishsimon (talk) 02:01, 11 April 2018 (UTC)
Thank you. --Titodutta (talk) 20:28, 13 April 2018 (UTC)


Feedback requested on educational resources and curricula

I have a goal of creating a series of lessons which a new user could follow to gain the skills to use Wikidata.

I drafted an outline of resources.

My thought is that I will seek to collect any lessons on the "educational resources" page, then include those individual lessons and more into a curriculum on the "curricula" page.

Here is the feedback I am seeking now:

  1. Can anyone inform me if they are aware of any Wikidata page which already serves either of these functions? Is there a catch-all directory of educational resources, and has anyone already outlined how to create a Wikidata class learning plan?
  2. Can anyone post more educational resources to that page?
  3. Can anyone direct me to individuals who are likely to want to collaborate in developing a multi-part lesson plan for new users to learn Wikidata?

Thanks. Blue Rasberry (talk) 19:49, 11 April 2018 (UTC)

@Titodutta: Thanks. This is not stable. I am seeking comments from others. If you are aware of anything equivalent in other languages then share. I have no idea what already exists. Right now I am only collecting anything already published and getting feedback, so I think translation is premature. Blue Rasberry (talk) 20:28, 13 April 2018 (UTC)

How to load a new dataset?

Hello, I'm wondering what is the correct process for loading a dataset. Some week ago, I tried to open up a new purpose here, but now I do not know how to proceed. Floatingpurr (talk) 09:47, 12 April 2018 (UTC)

Seems to be plenty of supply on that page, but hardly any demand. Multichill (talk) 19:28, 12 April 2018 (UTC)
I see. So what how can I do if I want to contribute loading datasets? Floatingpurr (talk) 11:24, 13 April 2018 (UTC)
There are probably three or four main tasks associated with this proposed upload: 1. Working out whether we already have records for any of the schools 2. working out the coding so we can represent e.g. Region or Province to our values for Italian regions & Provinces 3. Actually uploading the data, probably using QuickStatements 2 (Q29032512) and possibly 4. creating a new property for CODICESCUOLA (and, who knows, thinking about CODICEISTITUTORIFERIMENTO and whether we need to support that ID). So it's a non-trivial task, I'm afraid, one which will take a considerable amount of effort to achieve. There is not an automagical solution. So the question is, whether you have the time and aptitude to do the work? I can probably give you pointers. --Tagishsimon (talk) 14:28, 13 April 2018 (UTC)
Thanks for your kind reply. So, I'd like to contribute and to try loading datasets for becoming confident with this process. I go over your points. 1. There are definitely record of such schools, for example this one Liceo Classico Massimo D'Azeglio (Q3268994). I guess the only way to get them all is string matching, right? How do I merge existing data? 2. Again, also in this case string matching is the only way, isn't it? 3. Aren't there APIs for loading data (e.g., the ones the bots harness)? Do I need a bot to load them? 4. I may start loading just basic info. I understand it's not a trivial task and I do not know how much time I can dedicate to this task. Anyway, I still do not find a clear way for contributing with huge data loadings, like this one. Floatingpurr (talk) 23:52, 13 April 2018 (UTC)

Property "Time of the day" by using 1440 items

Since the use case of "time of the day" is different than the use case of "duration in HH:MM:SS", I was wondering if we could create a property "time of the day" with data type item which takes as value one of 1440 items (24h * 60min), each one labeled as "00:00", "00:01", "00:02", and so on. Would it make sense? Are there any use cases?

For opening hours I suppose we would need a property "open on" with item values "Monday", "Tuesday", etc. and qualifiers "start time of the day" and "end time of the day".

Thoughts? --Micru (talk) 21:14, 13 April 2018 (UTC)

I'm also thinking that perhaps there is no need to create so many items, as normally the only items used would be for hours, or half-hours, perhaps quarters in some rare occasions. In that case only 96 items (4*24) would be necessary to have the most used times 0h, 0:15, 0:30, 0:45, and so on. --Micru (talk) 10:36, 15 April 2018 (UTC)

Records and record progressions

On the one hand, the English-language Wikipedia has an article called Flight airspeed record, on the other hand, it has an article called Men's pole vault world record progression.

Now two examples are given for record held (P1000), Renaud Lavillenie (Q1742) → men's pole vault world record (Q1136293) and Augustus Orlebar (Q4821509) → flight airspeed record (Q1038621), but one cannot say that Lavillenie holds the men's pole vault world record progression. What to do about this, changing the Wikidata label into men's pole vault world record but leaving the Wikipedia article linked to it? Note that the French label is record du monde du saut à la perche (OK) and the German label is Stabhochsprungweltrekorde der Männer (plural!).

I am also uncontent with the use of the property record held (P1000) in the item 100 metres (Q164761). It is not like Usain Bolt is a record held by the 100 metres race; rather, the record for the fastest 100 metres race is held by Usain Bolt. -- IvanP (talk) 15:12, 14 April 2018 (UTC)

men's pole vault world record (Q1136293) actually seems like the better target, because it's an item for a subclass of world records, unlike flight airspeed record (Q1038621) which is an item for a Wikimedia list article. I think it would be OK to change the label. record held (P1000) in 100 metres (Q164761) is just backwards. Ghouston (talk) 03:56, 15 April 2018 (UTC)

Translating "Motivation"

Currently, property proposals have the string "Discussion" translated, by inputting {{int:Talk}} which outputs "Discussion". It would be useful to have the header "Motivation" also internationalized. The preloaded property proposal template would also need to be updated. NMaia (talk) 13:18, 15 April 2018 (UTC)

instance of:list and P279

Somehow subclass of (P279) doesn't work well on items for lists. That is items with instance of (P31)=Wikimedia list article (Q13406463).

There are some on Wikidata:WikiProject Lists/reports/P279 cleanup. It includes:

  • 1. Items with labels = "list of" and a subclass that is also labeled "list of"
  • 2. Items with labels = "list of" and a subclass that is not labeled "list of"
  • 3. Items with labels in plural (other than "list of") and a subclass that is not labeled "list of"
  • 4. Items with labels in singular (other than "list of") and a subclass that is not labeled "list of"

What to do?

Merge person Friedrich Bienemann

Friedrich Bienemann (Q12362741) = Friedrich Bienemann (Q19197249) 77.179.147.176 15:29, 21 April 2018 (UTC)

✓ Done Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:55, 21 April 2018 (UTC)
This section was archived on a request by: Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:55, 21 April 2018 (UTC)

Ancient and modern cities

Should the relationship between Londinium (Q927198) and London (Q84) be <followed by/follows> or <replaced by/replaces>? - PKM (talk) 23:38, 8 April 2018 (UTC)

<followed by/follows>, if I read the follows (P155) description right: "Use P1365 (replaces) if the preceding item was replaced, e.g. political offices, states and there is no identity between precedent and following geographic unit". There is an identity in the continuum of London. --Tagishsimon (talk) 23:43, 8 April 2018 (UTC)
Great, thank you. - PKM (talk) 23:46, 8 April 2018 (UTC)
@PKM, Tagishsimon: follows (P155) is supposed to only be used as a qualifier, according to the constraints. --Yair rand (talk) 06:29, 10 April 2018 (UTC)
Yes; sigh. The constraint forces us to define the domain of the observed sequence.
Perhaps
⟨ London (Q84)  View with Reasonator View with SQID ⟩ official name (P1448) View with SQID ⟨ London ⟩
follows (P155) View with SQID ⟨ Londinium (Q927198)  View with Reasonator View with SQID ⟩
or more colloquially
⟨ London (Q84)  View with Reasonator View with SQID ⟩ name (P2561) View with SQID ⟨ London ⟩
follows (P155) View with SQID ⟨ Londinium (Q927198)  View with Reasonator View with SQID ⟩
 ? --Tagishsimon (talk) 06:52, 10 April 2018 (UTC)
Or kill the constraint. I suspect it's used "incorrectly" (i.e. not as a qualifier) at leastr as often as it used "correctly" according to the constraint. - PKM (talk) 20:14, 12 April 2018 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ From the SQUID poperty browser:

  • followed by (P156) used as Statement: 381,522 Qualifier: 77,925
  • follows (P155) used as Statement: 391,115 Qualifier: 78,487

This constraint is a lost cause. - PKM (talk) 21:36, 13 April 2018 (UTC)

  • Sigh. But this constraint is not accidental - using these properties as main properties makes semantics more ambiguous: one cannot tell in which sequence they are "next/previous". The other solution was always to mark sequence as a qualifier with P361 or better part of the series (P179). But this approach seems to be even more rare. --Infovarius (talk) 09:29, 15 April 2018 (UTC)
So how could we fix these ~300k "errors"? It seems we'd need to convene a task force or some such, to make or find a sequence/series that is appropriate and then edit them all. - PKM (talk) 19:35, 15 April 2018 (UTC)

Refound date

We are using inception (P571) and dissolved, abolished or demolished date (P576) for a team foundation date and dissolved date. But, what we do if a team refounded after dissolved? Two different questions I have.

1) We put the refounded date at P571, as a second value? But, it has the constraint single-value constraint (Q19474404). How can we show that the team was refounded a specific date?

2) We made P576 value a deprecated rank? (sorry for my English)

3) If the team dissolved again? P756 has the constraint single-value constraint (Q19474404).

My example,

Pezoporikos Larnaca FC (Q2277220) founded at 1924, then dissolved at 1932 (merged with other team, created a new team), then refounded at 1937 and then dissolved at 1994 (merged with other team). For other teams, they just dissolved (not by merging) and refounded some years later.

Xaris333 (talk) 22:37, 14 April 2018 (UTC)

I'd say first decide whether the refounding is genuinely the same team, and not a new team with the same name. The same thing also happens with other types of organization like political parties and companies. If it a continuation, then I suppose you'd use inception (P571) for the first founding date, and dissolved, abolished or demolished date (P576) for the final dissolved date, with periods of inactivity not recorded in that way. Ghouston (talk) 23:14, 14 April 2018 (UTC)
Its the same team... Ok, but with your way we lose informations. There is a time in period that the team, political party, company, didn't exist. Xaris333 (talk) 23:51, 14 April 2018 (UTC)
Because that first time it dissolved it wasn't a true dissolution, because it continued at a later period. And the refounding isn't a true foundation since it existed previously. Ghouston (talk) 03:32, 15 April 2018 (UTC)
But how can we show that the team, political party, company didn't exist for some years? That period, the dissolution was true. Xaris333 (talk) 04:18, 15 April 2018 (UTC)
Maybe with significant event (P793) with date of dissolution, abolition or demolition (Q29933798) and point in time (Q186408), although this seems confusing. There don't seem to be items for start and end of a period of inactivity, or temporary cessation / restart. Ghouston (talk) 07:16, 15 April 2018 (UTC)
Quite a few such examples can be found by searching for "refounding" in Wikipedia. A lot of them seem to be sports teams, but Saxony-Anhalt (Q1206) is also a good example. Perhaps defining a couple of new items to be used with significant event (P793) would handle it? Ghouston (talk) 07:30, 15 April 2018 (UTC)
Identity is strongly associated to the temporal continuity of existence, which is delimited by a start date and an end date. If something ceases to exist, that will never exist again nor will be recreated, and a future entity with the same name should be considered a different entity. So I would create several entities in Wikidata, since I also consider them different entities in the real world. --abián 13:25, 15 April 2018 (UTC)
Wikidata doesn't really have the privilege of deciding such things: it's determined by whether the entity has multiple Wikipedia entries or a single entry. Perhaps a case will turn up that has multiple entries in one language and a singe entry in another. In reality its messy: there are likely to be aspects in which the entity acts as a continuation (beyond just the name, such as having some of the same people involved) and aspects in which it acts as a new entity (such as a new legal registration.) Did China become a new state in 1949 when it became the People's Republic of China (Q148), or is was it a continuation of the same state with a new government and new name? We have People's Republic of China (Q148) founded in 1949. Ghouston (talk) 23:39, 15 April 2018 (UTC)

Title

Hello. Why title (P1476) should only contain a single value? Some papers and files have two languages on them. For example English and Greek. No one of the language can be called as the original and the other as the translated language. Both are the original languages. Xaris333 (talk) 19:02, 4 April 2018 (UTC)

You can enter both titles, and then edit title (P1476) to have a new "exception to constraint" for this particular item that has two titles. - PKM (talk) 20:39, 4 April 2018 (UTC)
But there are many items. Xaris333 (talk) 22:17, 4 April 2018 (UTC)
@Xaris333: Can you give an example or two, please? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:51, 5 April 2018 (UTC)
@Pigsonthewing: See Amiantos (Q6377735) --> population (P1082) --> 262. See the source. Xaris333 (talk) 13:33, 5 April 2018 (UTC)
Thank you. I was looking for examples of Wikidata items about such works. Incidentally, you can link to a specific property on an item, like this: Q6377735#P1082. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:06, 5 April 2018 (UTC)
there are also all multilingual editions of books or texts in wikisource, with a link on both ws projects, like Q25991162, which is a whole book, or Q25991163, which is a poem. As all poems and individual works in those multilingual texts are to be referenced, there will be a LOT of them, just for latin authors :)
aaannd cases of poems that are known and published under various titles (in the same edition), like medieval Rutebeuf's see here Q19181898, where 3 different titles are mentioned on the very same page of the same edition :/ --Hsarrazin (talk) 15:21, 5 April 2018 (UTC)

Should we remove the constraint? Xaris333 (talk) 01:20, 6 April 2018 (UTC)

  •  Support removing the single-value constraint on "title". - PKM (talk) 19:42, 9 April 2018 (UTC)
  •  Oppose The overwhelming majority of the cases will still have only a single value, and all too often extra values are added that shouldn't be there. It's not a mandatory constraint, but the constraint is useful for tracking these cases. – Máté (talk) 06:19, 11 April 2018 (UTC)
  • ⟨ subject ⟩  Wikidata property  ⟨ object or value ⟩
    as long as it is NOT mandatory, there is no problem for me :) --Hsarrazin (talk) 08:33, 16 April 2018 (UTC)

Change a birth date on a protected page

I am not autoconfirmed since I don't edit wikidata. The issue is with María Félix (Q465189). She was born 4 May not 8 April (that's her death date). Google recently made a doodle of the actress mistakenly giving that date as her birthday. Here is a source for her actual birth from the New York Times obituary: "Ms. Félix was born María de los Ángeles Félix Guereña, on May 4, 1914, in Álamos, Sonora, according to the birth certificate discovered by Paco Ignacio Taibo, the author of La Doña, a 1985 biography of Ms. Félix." (https://www.nytimes.com/2002/04/09/movies/maria-felix-87-feisty-heroine-who-reigned-supreme-in-mexican-cinema-dies.html) AuroralColibri (talk) 23:26, 10 April 2018 (UTC)

I've added the NYT date, and deprecated the data.bnf.fr date ... I think we have to keep both, as we have a source making a claim. --Tagishsimon (talk) 00:13, 11 April 2018 (UTC)
I just notified BnF for correction
next time, you can do it yoursel by clicking the "Signaler une erreur sur cette notice" link on the person's record http://catalogue.bnf.fr/ark:/12148/cb14046928v (with source, of course) -- they will send you notice when they fix it :) --Hsarrazin (talk) 08:27, 16 April 2018 (UTC)

Updating Wikidata:List of properties

Since all of our old lists of properties are outdated, I would like to replace Wikidata:List of properties with a new User:Micru/Wikidata:List of properties.

Could you please take a look and voice your opinion? Thanks! --Micru (talk) 20:00, 13 April 2018 (UTC)

  • Hi, I would agree that we need better ways to explore and visualize properties, and your proposal goes into the right direction. However, some of the tools you are listing are by far not mature (yet). For example, it should definitely be possible to filter out all the identifier properties - in fact they should be explorable separately from the other properties. Also, some of the standard queries of the tools you are referencing run into time-out issues (at least when run from the Mozilla Browser, which has a lower performance on the SPARQL service than for example Google Chrome). Cheers. --Beat Estermann (talk) 09:48, 15 April 2018 (UTC)
@Beat Estermann: Prop explorer has been updated and now you can filter out all the identifier properties, and you can explore them independently if you wish, just select the type of property that you want to show. If you tell me which tools give you problems I can either contact the tool creators or add a notice with the potential issues. So do you think the proposed page is now ready to be used as default? --Micru (talk) 07:35, 16 April 2018 (UTC)

Wikidata Games

Are any of the Wikidata games working for anyone? I've tried two browsers, and various gales, and they're not for me (and I know Magnus is busy elsewhere). for example https://tools.wmflabs.org/wikidata-game/distributed/#game=9 hangs in Firefox and gives an unspecified error in Opera. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:08, 13 April 2018 (UTC)

Game Mode from Mix'n'Match (https://tools.wmflabs.org/mix-n-match/#/random/299) works for me. The Distributed Game (your example link https://tools.wmflabs.org/wikidata-game/distributed/#game=9) hangs. I am using Chrome + Windows 10. - PKM (talk) 21:48, 13 April 2018 (UTC)
Does not work for me, there seems to a problem in the tool that provides suggestion for the game. Matěj Suchánek (talk) 13:25, 14 April 2018 (UTC)
The "error" was that you had played through all' of the tiles (~8400) I had created back in the day. The update function was still using WDQ, which has been deactivated quite a while ago, so no new candidates. I patched it up a bit, running now. Should already work again. --Magnus Manske (talk) 08:32, 16 April 2018 (UTC)
https://tools.wmflabs.org/wikidata-game/ certainly stopped working for me months ago : impossible to get a suggestion. waiting forever.
pity because it was very useful to match people and gender quickly :( --Hsarrazin (talk) 08:47, 16 April 2018 (UTC)

Help, PetScan

Hi. I wanna know why I'm not getting labels in Spanish with this petscan-query. strakhov (talk) 21:34, 13 April 2018 (UTC)

I think it's Topic:Ub9l5yd1u0m8t92q. --Edgars2007 (talk) 16:27, 14 April 2018 (UTC)
Thanks! Well, I'll wait to see what happens. :) strakhov (talk) 17:30, 15 April 2018 (UTC)
Fixed. --Magnus Manske (talk) 09:37, 16 April 2018 (UTC)

about project tiger

how can i be a perticipent of this competition? should i only need to log in by my wiki account and edit pages? or do i need to do things first?

@Mahmudul hasan topu: (Whenever you write a message on this page or any other talk page, please sign your posts using four tildes—like "~~~~", but without any quotation marks or <nowiki> tags around it.) At present Wikidata is not within the scope of the Project Tiger Editathon. You may wish to find out more here. Mahir256 (talk) 16:18, 16 April 2018 (UTC)

Wikidata weekly summary #308

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • Disable Senses in the API for the initial version of the WikibaseLexeme extension (phab:T187198)
    • Create a new data type to allow linking to Forms (phab:T165579)
    • Fix a bug happening when editing a Form (phab:T187215)
    • Enforce permission checks when editing Lexeme (phab:T189480)
    • Make not possible to remove the last lemma (phab:T189185)
    • Make not possible to move a Lexeme (phab:T189481)
    • Implement “Units” constraint (phab:T164372)
    • More work on the possibility to pick an item's parent in Lua (phab:T179155)
    • Show a thumbnail for image/video/audio properties in statements (phab:T46727)
    • Add “no bounds” constraint (phab:T170610)
    • Make the “difference within range” constraint not requiring “item requires claim” constraint (phab:T185480)
    • New Lua function to test for validity of Wikidata entity IDs (phab:T143970)
    • Include checksums in dumps (phab:T190457)

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.


New constraint to enforce that two properties have the same value on a given item

There is an ongoing discussion at Wikidata:Property proposal/Directory of Open Access Journals ID and Wikidata:Property proposal/Plants of the World online (among others) about how to link to two different databases which use the same identifier. The simplest solution seems to consist in creating two distinct properties which would generally have the same value when they are present on the same item (but different URLs generated by the formatters). It can be a burden to deal with this redundancy so some editors wonder if a new type of constraint could be created to alleviate that. Would that be an appropriate solution? @Lucas Werkmeister (WMDE): would it be technically feasible? − Pintoch (talk) 08:51, 8 April 2018 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ As I've noted elsewhere, "distinct properties which would generally have the same value" is far from simple, and is in fact a harmful way to proceed. Consider:

for which we currently have Flora of North America taxon ID (P1727) and Flora of China ID (P1747); with a further eighteen potential properties to follow. Sadly, only four people responded when I raised this here and at Wikidata:Requests for deletions/Archive/2018/Properties/1#eFlora properties. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:52, 8 April 2018 (UTC)

This seems to be a slightly different issue as all these URLs come from the same database. I understand that you vocally oppose the creation of these properties but I still don't understand what you are proposing to do instead (specifically for the two proposals linked above)? Third-party formatter URLs are not a solution for this particular problem. − Pintoch (talk) 12:14, 8 April 2018 (UTC)
"Third-party formatter URLs are not a solution for this particular problem" Yes, they are; and that is the solution I have - quite clearly - proposed. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:34, 8 April 2018 (UTC)
@Pintoch: I tend to agree with Andy here: we should be using a third-party formatter URL in these cases. Unless there is at least one instance where the identifier differs between, say, P1727 and P1747 for the same item (which judging from Andy's links is in the same database), the two properties should be unified. What should be worked on is a way to control which formatter URL appears on a given item—in this case, based on some sort of geographical information on the item itself. Mahir256 (talk) 18:46, 8 April 2018 (UTC)
@Pigsonthewing, Mahir256: Sorry if it is obvious but I still do not see how you would use a third-party formatter URL to indicate that a given identifier is available in database X but not in database Y? For instance, not all ISSNs can be found in DOAJ. Also I don't see in which context the third-party formatter URL is actually used to generate a link for the identifier in the Wikibase UI? It would be great if you could give a very concrete example of this, because I think that I am not the only one who does not get it. − Pintoch (talk) 19:13, 8 April 2018 (UTC)
Couldn't we create a property for journals like IndexedIn:DOAJ ? --Gerwoman (talk) 19:17, 8 April 2018 (UTC)
To piggyback off Gerwoman's suggestion, why not have the formatter URL change based on a reference "stated in (P248) Flora of China (Q5460442)" or "stated in (P248) Directory of Open Access Journals (Q1227538)" as appropriate? Mahir256 (talk) 19:22, 8 April 2018 (UTC)
That's intuitively harder to implement, but why not. As far as I understand Andy claims that somehow no change is needed and third party formatters already provide a solution. − Pintoch (talk) 21:36, 8 April 2018 (UTC)
While it is not necessary to "indicate that a given identifier is available in database X but not in database Y" (that's what the 404 response code is for), it is possible to do so, using referencing, as I have demonstrated previously. You were apparently aware of the example I gave then. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:36, 9 April 2018 (UTC)
As I understand it, people just want to have a hypertext link to the said databases. Manually constructing the link from a third-party formatter URL and checking if it resolves correctly might be an option for bots, but not for people. So even if I like the idea, I don't really see how this method is a solution: that involves statically instantiating the third-party formatter URL for the given ID (which makes it hard to update the URLs if the format of the website changes). And I guess it is not very convenient for users either, as they have to expand references to check if there is a link there. I am not really speaking for myself here: I just notice that there are quite a lot of support votes for these properties, and I am worried that by refusing their creation we are going to make their life harder. − Pintoch (talk) 14:22, 9 April 2018 (UTC)
Hello all,
Feel free to create a ticket on Phabricator, with the tags Wikidata and Wikibase-Quality-Constraints. Don't forget to provide examples of what you need. Is this related to this specific property, or would you need it for others as well? We'll have a look as soon as possible.
Cheers, Lea Lacroix (WMDE) (talk) 15:20, 9 April 2018 (UTC)
Thanks! I have created a task here: https://phabricator.wikimedia.org/T191963 − Pintoch (talk) 09:49, 11 April 2018 (UTC)
"people just want to have a hypertext link to the said databases" Perhaps. But given Wikidata's purpose, is that the right thing for us to do? "I want" is not a very compelling use-case. "that involves statically instantiating the third-party formatter URL for the given ID ". No, it involves giving a valid citation for the data; and happily also meets your requirement to "indicate that a given identifier is available in database X but not in database Y". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:43, 9 April 2018 (UTC)
The ID overlap between FNA and FOC is around 2,500. This will produce more than 40,000 404 errors. --Succu (talk) 19:33, 13 April 2018 (UTC)

Despite this ongoing discussion, and despite the lack of examples of the type requested in its property proposal, Plants of the World Online ID (P5037) has just ben created. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:34, 11 April 2018 (UTC)

One example: Byttneriaceae (Q2548700) in POWO has no entry in IPNI. --Succu (talk) 18:08, 13 April 2018 (UTC)

IAFD person Ambiguity properties

There is an ambiguity with the IAFD person properties on Wikidata, IAFD male performer ID (P4505) and IAFD female performer ID (P3869), the correct one would be to only exist an "IAFD person ID" as like Adult Film Database person ID (P3351) . I believe that this ambiguity occurred because of the second parameter gender. It was mistaken for a secondary ID, in fact it is just a layout parameter for the IAFD site.

All links below point to the same element on the IAFD site.

http://www.iafd.com/person.rme/perfid=glynn
http://www.iafd.com/person.rme/perfid=glynn/gender=f/ginger-lynn.htm
http://www.iafd.com/person.rme/perfid=glynn/gender=m/ginger-lynn.htm
http://www.iafd.com/person.rme/perfid=glynn/gender=d/ginger-lynn.htm

This ambiguity makes it difficult to integrate with templates on Wikipedia.

Guilherme Burn (talk) 13:34, 10 April 2018 (UTC)

Note: IAFD is an "adult film" database and the above links may be considered "not safe for work". Having learned that the, er, difficult, way: your first two links return different data. For example, the first link has no picture, while the second does; and the former only lists two films; the latter has 342. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:12, 10 April 2018 (UTC)
Yes @Pigsonthewing:. But that does not justify treating people of masculine, feminine, and "director" genres as different elements. If you repair the first link, without this parameter already have the correct information.Guilherme Burn (talk) 15:32, 10 April 2018 (UTC)

Is it possible to rename a property? If yes, it would be enough to exclude the less used one and renomer another. Of course, if there is consensus.Guilherme Burn (talk) 19:33, 16 April 2018 (UTC)

Dye plants

Do people prefer Reseda luteola (Q157927) <instance of> dye plant (Q907341) or Reseda luteola (Q157927) <use> dye plant (Q907341)? I prefer <instance of> but not enough to argue about it if others disagree, as long as we're consistent. There are only two instances so far, so I can change them easily if that's the consensus. - PKM (talk) 22:56, 16 April 2018 (UTC)

Or perhaps <subclass of> dye plant (Q907341)? - PKM (talk) 23:07, 16 April 2018 (UTC)
Instance or subclass. dye plant (Q907341) has a description "plant from which a dye can be extracted" and a has use (P366) with the value dyeing (Q1164991). I could do with more guidance on instance versus subclass; it's not always clear to me which is appropriate. --Tagishsimon (talk) 23:11, 16 April 2018 (UTC)
An individual plant would be an instance, Reseda luteola (Q157927) is a class of many plants. Ghouston (talk) 01:12, 17 April 2018 (UTC)
I think generally it's not a good idea to go overboard with subclassing. It's nice that all humans are instances of the same item (even if that item is somewhat screwed up.) I think I'd just make each species a subclass of its parent taxon and identify its characteristics with other properties. Ghouston (talk) 01:20, 17 April 2018 (UTC)

defining formula (P2534)

Is really a defining formula for monoid (Q208237)? (from semigroup (Q207348)) is at least an equation, but is not a notation specifically for a semigroup, rather, it can be used to define what a semigroup is by saying that is a semigroup iff is an associative operation and is a set closed under .

Furthermore, has characteristic (P1552) currently has two values in monoid (Q208237), identity element (Q185813) and associativity (Q177251). I am fine with identity element (Q185813) – there are monoids with different identity elements –, but think that associativity (Q177251) is inappropriate. -- IvanP (talk) 18:57, 16 April 2018 (UTC)

@IvanP: On defining formula, a lot of these were imported wholesale from enwiki in a careless manner, and I've deleted hundreds that were just inappropriate. However the two you mention were added more recently by @Bigbossfarin: - might be something else going on here? ArthurPSmith (talk) 12:37, 17 April 2018 (UTC)
Hi, I was importing the formula from de-wiki. In this sense a monoid (Q208237) is defined as a triple (Q20088882) , where is a set (Q36161), is a associative internal binary operation (Q47407371) on and a identity element (Q185813) in . The input box is simply to short to get all this information in it. Maybe would be an improvement? And I agree with you that associativity (Q177251) is very inaccurate. --Bigbossfarin (talk) 13:30, 17 April 2018 (UTC)

2 years

Dear administrators,

Mr Javier Gomá (https://www.wikidata.org/wiki/Q6165536) was hired by the Juan March Foundation in 1996. In 2003, he was named executive director. How can I express it in Wikidata? I am not able to put 2 years.

Regards.

Soleil222 (talk) 08:24, 17 April 2018 (UTC)

I think you have to add the employer twice, and only once with "position held" with the 2003 date. --Anvilaquarius (talk) 08:28, 17 April 2018 (UTC)

Translation not updated

Tracked in Phabricator
Task T192379
Resolved

Can someone tell me why this translation has not been yet replicated to MediaWiki:Valueview-expertextender-languageselector-label/ca? I translated it last december. Thanks, Paucabot (talk) 14:47, 17 April 2018 (UTC)

@Paucabot: the translation was imported correctly (Gerrit change Idd2f67762d), but that message is not directly part of Wikibase (the MediaWiki extension that powers Wikidata), but of a library which Wikibase uses. Wikibase imports the latest released version of that library, and we haven’t released a new version of it since November, so all translation changes since then aren’t effective on Wikidata yet. We’ll try to get this fixed, it should be deployed either tomorrow or next Wednesday. Thanks for translating, and for bringing this to our attention! --Lucas Werkmeister (WMDE) (talk) 16:17, 17 April 2018 (UTC)

Commons Cultural heritage monuments with known IDs

Hi guys,

The question is simple, there's some effort to make this: Category:Cultural heritage monuments with known IDs some how useful here, there's some integration possible to both worlds? Rodrigo Tetsuo Argenton (talk) 20:04, 15 April 2018 (UTC)

@Rodrigo Tetsuo Argenton: Sorry, I don't understand your question. Can you please rephrase it? You're probably interested in Wikidata:WikiProject WLM. Multichill (talk) 21:17, 15 April 2018 (UTC)
@Multichill:, if there is any kind of integration between the categories/templates at Wikimedia Commons and here. We have there a number to identify the monument, maybe we could use some how here. Because, we have there thousands of images identified, and as we categorised it by template, we could simply change the templates and created a bridge there easily by bot, not manually as now.
Rodrigo Tetsuo Argenton (talk) 21:48, 15 April 2018 (UTC)
@Rodrigo Tetsuo Argenton: If the ID template is being used in categories, then it would be good to migrate the info to here (if not already here) and we can use commons:Template:Wikidata Infobox to show it there. For files, a lot of the work will become easier once structured commons is here, but for IDs that are present both here and there then we could try to match them up, and add image (P18) here, and "wikidata=" tags there. Thanks. Mike Peel (talk) 11:51, 16 April 2018 (UTC)
@Rodrigo Tetsuo Argenton, Mike Peel: before you start doing double work:
The monuments database indexes about 1,5 million monuments (stats). A lot of these id's are cross referenced with Wikidata already or that can easily be done. The monuments database used to index images too, last time it ran it was about 2,6 million images (more info). I haven't invested any time yet on removing the intermediate step. To explain that, now it's file -> id, id -> Wikidata, that could be file -> Wikidata. The reason I haven't done anything yet is Structured Data on Wikimedia Commons. This is a good set to convert once we have the initial release of that deployed. It's a bit of double work to add Wikidata id's now in template pseudo structured data and doing everything again a bit later.
I would wait a bit and in the meantime focus on improving items with heritage designation (P1435) and making sure all the items about monuments have at least one of these properties. Also the cross referencing effort probably could use a hand, but I don't know the latest status. Multichill (talk) 16:23, 16 April 2018 (UTC)
@Multichill: For files, I'd agree, better to wait for structured commons (although the latest changes to commons:Template:Artwork to auto-fetch info from Wikidata when given the qid are a huge step forward). For categories, structured commons won't impact those, so moving the IDs in those over sooner rather than later makes more sense. Thanks. Mike Peel (talk) 18:12, 16 April 2018 (UTC)

Model a primary source and maybe also quality

In Sweden the National archives released 100 million documents for free last month ==> we can now in WD have Primary sources like birth and death records as a source (example query)

Question: is there a best practice to model a primary source so that I can ask for 

all birth dates for Swedish people not confirmed by a primary source born more than 70 years ago?

If not how do we model sources in the best way

  1. type of source e.g.primary/secondary/compiled/....
  2. quality of source
    1. High quality
      1. they have a well-documented quality system
      2. track record of delivering quality the last 10 years...
      3. by professionals mentioned as having "hallmark" quality and have very very high trustability?
    2. ???

- Salgo60 (talk) 07:30, 18 April 2018 (UTC)

Query coords

I want to create a map of all Items, having a LfDS object ID (P1708). My try is:

https://query.wikidata.org/#%23Liste%20aller%20Kulturdenkmale%20in%20Sachsen%0A%23defaultView%3AMap%0ASELECT%20%3FitemLabel%20%3FitemDescription%20%3Fimage%20%3Fcoord%20WHERE%20%7B%0A%20%20%3Fitem%20%28wdt%3AP31%29%20wd%3AQ11691318.%0A%20%20%3Fitem%20wdt%3AP17%20wd%3AQ183.%0A%20%20%3Fitem%20wdt%3AP625%20%3Fcoord.%0A%20%20OPTIONAL%20%7B%20%3Fitem%20wdt%3AP18%20%3Fimage.%20%7D%0A%20%20SERVICE%20wikibase%3Alabel%20%7B%20bd%3AserviceParam%20wikibase%3Alanguage%20%22de%22.%20%7D%0A%7D

But it does not work, can you help me? Regards, Conny (talk) 17:28, 18 April 2018 (UTC).

@Conny: I think this would be the query: 
#Liste aller Kulturdenkmale in Sachsen
#defaultView:Map
SELECT ?item ?itemLabel ?itemDescription ?id ?image ?coord WHERE {
  ?item wdt:P1708 ?id;
        wdt:P625 ?coord.
  OPTIONAL { ?item wdt:P18 ?image. }
  SERVICE wikibase:label { bd:serviceParam wikibase:language "de". }
}
Try it!
– warning: there are currently 80644 results, showing a map for them can be quite taxing for your browser. (In my case, the page hung for about 60 seconds until the map was displayed). --TweetsFactsAndQueries (talk) 17:38, 18 April 2018 (UTC)

printing industry (Q1261092) and printing industry (Q26897133)

What to do with these two?

The scope of the two German articles looks essentially the same. Could a German speaker either merge them, or distinguish them?

Alternatively, is there a distinction to be made between a broader industrial sector that includes all printing-related activities (eg ink manufacturing, book handling, etc), and the strict business of printing itself ? Jheald (talk) 18:30, 18 April 2018 (UTC)

Is there way to see real "related changes" with certain property

There is "related changes" link (like this). This link shows "edit on label", "edit on description", "addition of sitelink" and so on. Is there way to see only "edits on P1323 value". I want to see such changes for maintenance purpose, if it's possible. --Was a bee (talk) 10:35, 13 April 2018 (UTC)

@Was a bee: User:Yair rand/DiffLists.js allows filtering changes by property, which can be used in conjunction with Special:RelatedChanges to do this. Demo: Recent additions and changes for P1323. --Yair rand (talk) 02:50, 16 April 2018 (UTC)
@Yair rand: Wow, this is what exactly I hoped. Although currently not fully working (interface is shown, filtering doesn't happen), I wait implementation of this functionalities through T121361. Thanks a lot! --Was a bee (talk) 21:37, 16 April 2018 (UTC)
@Was a bee: Ah, I forgot to mention that it doesn't work unless "Hide the improved version of Recent Changes" is checked in Special:Preferences#mw-prefsection-rc. --Yair rand (talk) 22:35, 16 April 2018 (UTC)
@Yair rand: Oh, after resetting preference, it worked!. Although mysteriously filtering doesn't work in watchlist and recent changes, it works in history and contribution page. Hmm, this is interesting software. Thanks for let me know!--Was a bee (talk) 20:40, 18 April 2018 (UTC)

P31=Norse human

When working on property:P2600 I found that one item is instanceOf Norse human Q24451723. There are many more humans linked via P31 Special:WhatLinksHere/Q24451723. Is this desired? 78.55.76.113 18:43, 17 April 2018 (UTC)

I don't think Q24451723 is needed or desirable, it can be done with ethnic group (P172) Norsemen (Q1211290). Subclassing humans would lead to the madness that is Commons categories and intersection categories, since humans can have a lot of different characteristics. Ghouston (talk) 21:56, 17 April 2018 (UTC)
Agreed. Not desired. Looks like a one-person wikiproject - Wikidata:WikiProject Norse - intent on turning all sorts of stuff (well, runestones, humans) into Norse Runestones and Norse People. [11] [12]. @Anders Feder: For Ghouston's reasons, not a sensible way to proceed. --Tagishsimon (talk) 22:07, 17 April 2018 (UTC)
Strongly agreed. Let's get rid of this property. --Anvilaquarius (talk) 08:43, 19 April 2018 (UTC)

Books and Authors

I want some opinions. I created yesterday for litteraturbanken.se a proposal link...

As it looks technical possible to have one property for both books and authors I start feeling its better to split this property into

  • Swedish_Literature_AuthorID
  • Swedish_Literature_BookID

Any comments/suggestions?

My argument is that its conceptual easier implementing two properties- Salgo60 (talk) 05:29, 18 April 2018 (UTC)

@Salgo60: Two properties are required mainly to be able to add some constraints "the item using Swedish_Literature_AuthorID has to have instance of human". Snipre (talk) 13:13, 19 April 2018 (UTC)

Help needed to migrate "unit symbol" from string to monolingual text

unit symbol (P5061) has been created and P558 (P558) has been deprecated. Now there is the issue of the migration of the data from one property to another. Can it be done by bot? --Micru (talk) 14:57, 18 April 2018 (UTC)

Oh, and now we should have 6000 unit symbol (P5061) in each unit?? --Infovarius (talk) 08:50, 19 April 2018 (UTC)
I undid the deprecation as it hadn't gone through a deletion request. In the meantime one was made at Wikidata:Properties_for_deletion#P2237.
--- Jura 12:54, 19 April 2018 (UTC)

"Better source needed" ?

Is there a way to indicate "better source needed" on a statement ?

For example, suppose I have a date and place of birth for somebody, from a genealogical website -- but it doesn't give primary sources. The family relationships are right, confirmed by multiple primary sources; the year of birth might be right - it matches the age given on the death certificate; though the date on the tombstone is a year earlier; and ages given on other official documents, such as census returns, immigration forms etc are all over the place; there's also a place of birth confidently stated on the genealogical website, that seems quite a long way from the family's long-term residence, but doesn't seem to have come from any other source I can see online.

The subject died in 1908, so there's no BLP issue.

I'd like to include the stated birth-place, but note that its ultimate sourcing is not clear. Is there an appropriate way to do this? On some wikis one can tag a statement with Template:Better source needed (Q6716717). Is there anything analogous here that one can do to mark a statement that could use better verification? Jheald (talk) 12:41, 19 April 2018 (UTC)

Found a line in an old newspaper. Turns out that the tombstone was right, the place was right, the death certificate was one year and one day out. But it would still be useful to think how to mark statements where improved sourcing is needed. Jheald (talk) 13:33, 19 April 2018 (UTC)

4 242 000 instances of human - SPARQL for Q5 externalid statistics

SELECT (COUNT(?item) AS ?count)
WHERE {?item wdt:P31/wdt:P279* wd:Q5}
Try it!

Only two Wikipedias have more than 4 million articles, namely enwiki 5.6 mio, and cebwiki 5.3 mio [13]. In the list de:Liste genealogischer Datenbanken Wikidata would be ranked 26 by number of human items.

Are there a sparql queries that could show

or even better

  • lists quantities for each externalID that is used on any instance of Q5

According to Wikidata:Database reports/List of properties/Top100 total VIAF ID (P214) = 1181059, GND ID (P227) = 606095, ISNI (P213) = 541276, BnF ID (P268) = 401272, but these numbers are not restricted to instances of Q5.

92.229.165.74 13:02, 19 April 2018 (UTC) 

SELECT (COUNT(DISTINCT(?item)) AS ?count)
WHERE {
  ?item wdt:P31 wd:Q5 .
  ?item wdt:P214 [] .
}
Try it!
finds 1,018,671 distinct humans with VIAFs. (1,027,083 uses of VIAF ID (P214) on humans as a whole). Jheald (talk) 13:13, 19 April 2018 (UTC)

Jheald, thanks a lot! I did for ISNI P213 and got 476,549. The SPARQL for both on one item gives 474,573, so there are ~2000 items that have ISNI but no VIAF: 

SELECT (COUNT(DISTINCT(?item)) AS ?count)
WHERE {
  ?item wdt:P31 wd:Q5 .
  ?item wdt:P213 [] .
  ?item wdt:P214 [] .
}
Try it!

Do you know how to count two properties in one query, so it should give the VIAF = 1,018,671 and ISNI = 476,549 in one run? 92.229.165.74 13:24, 19 April 2018 (UTC)

This gets close, but it counts the number of different ISNIs and VIAFs rather than the number of different humans.
SELECT (COUNT(?isni) AS ?count_isni) (COUNT(?viaf) AS ?count_viaf)
WHERE {
  ?item wdt:P31 wd:Q5 .
  OPTIONAL {
    ?item wdt:P213 ?isni .
  }
  OPTIONAL {
    ?item wdt:P214 ?viaf .
  }
}
Try it!
Note that the run time is now over 40 seconds -- the maximum allowed for a query is 60 seconds, so that may limit how much more can be built in.
It would probably be a good idea to move this thread to Wikidata:Request a query, which is a forum specifically for the writing and improving of queries. Jheald (talk) 13:38, 19 April 2018 (UTC)
Here's an attempt to count the number of actual distinct items with ISNIs and VIAFs, but it times out: tinyurl.com/ycf5g829 Jheald (talk) 13:42, 19 April 2018 (UTC)

Jheald, thanks a lot, I started : Wikidata:Request a query#SPARQL for Q5 externalid statistics 92.229.165.74 13:50, 19 April 2018 (UTC)

How to split Q6145461 in movie and book version?

Yoru wa mijikashi aruke yo otome: the data here are mixed up, most of them refer to the 2017 anime movie, and a few (chinese, cantonese and japanese) to the 2006 novel book . How can I split the informations and move some of them to a new data page? --Artafinde (talk) 11:14, 19 April 2018 (UTC)

This section was archived on a request by: Matěj Suchánek (talk) 09:24, 25 April 2018 (UTC)

Can an item be an instance of a disambiguation page and also a unit of measure?

These edits assert that a year (Q577) remains an instance of a unit of time (Q1790144) as claimed before the edits, but is also an instance of a Wikimedia disambiguation page (Q4167410). That is counterintuitive. Is it allowed at Wikidata? Jc3s5h (talk) 20:03, 22 April 2018 (UTC)

Of course not. Sjoerd de Bruin (talk) 20:10, 22 April 2018 (UTC)
And now undone. —MisterSynergy (talk) 20:27, 22 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 09:24, 25 April 2018 (UTC)

Mix'n'match - wrong description

https://tools.wmflabs.org/mix-n-match/#/catalog/177 it says 

BBL // person's ID at Baltisches Biographisches Lexikon digital encyclopedia

correct would be 

BBLd // ID at Baltisches Biographisches Lexikon digital encyclopedia

since it is the ID for "Baltisches Biographisches Lexikon digital" not "Baltisches Biographisches Lexikon" and it is not restricted to persons. 77.180.35.91 13:47, 23 April 2018 (UTC)

Also @Magnus Manske: --Tagishsimon (talk) 13:59, 23 April 2018 (UTC)
Found this here by sheer accident. Fixed. --Magnus Manske (talk) 15:30, 23 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 09:24, 25 April 2018 (UTC)

Merge 2x Hermann Amadeus Adolphi 1841-1924

Q52084590 = Q16357635 77.180.35.91 14:47, 23 April 2018 (UTC)

Done --Tagishsimon (talk) 14:53, 23 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 09:24, 25 April 2018 (UTC)

CirrusSearch issue returning pages in multiple categories

Hi

So I'm having some issues with CirrusSearch which I'm 80% sure I'm searching for the right thing and following the instructions correctly but am not getting the results I expect.

I'm working on a new version of Wikidata:Dataset Imports which has each dataset recorded on an individual page and using categories to organise the pages. I'm nearly done with this new version and once I've solved this issue people can use the page.

What I'm trying to do is run a search to find items that exist in both of two categories using the instructions here.

I've created a test page Wikidata:Dataset_Imports/test which is in both Category:Medical datasets and Category:Updated datasets and I'm trying to create a query in the search bar to find the page. I think it should be:

incategory:"Text document datasets" incategory:"Medical datasets"

However this doesn't give me any results..... Can someone tell me what I'm doing wrong?

Thanks

--John Cummings (talk) 09:53, 24 April 2018 (UTC)

Update: I worked it out, I needed to tick only the Wikidata project space in the advanced settings. --John Cummings (talk) 11:53, 24 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 09:24, 25 April 2018 (UTC)

Merge problems

Can someone merge en:Volonté (Q4016278) with de:Volonte (Q2532835) ? 88.70.215.195 14:27, 24 April 2018 (UTC)

Done. --Anvilaquarius (talk) 15:56, 24 April 2018 (UTC)

This section was archived on a request by: Matěj Suchánek (talk) 09:24, 25 April 2018 (UTC)

Merge 2x Alberto Castellano

Q16488303 = q35533344 77.14.46.178 18:53, 24 April 2018 (UTC)

done --Tagishsimon (talk) 19:02, 24 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 09:24, 25 April 2018 (UTC)

Merge 2x Manuel Gregorio Aróztegui

q40878117=q26255200 77.14.46.178 18:58, 24 April 2018 (UTC)

done --Tagishsimon (talk) 19:03, 24 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 09:24, 25 April 2018 (UTC)

Hemp fiber in English and Korean Wikipedias

Not sure what to do with this; 삼베 on kowiki links to nothing on enwiki when there is an article for it, sambe on enwiki. But the English article is about the specifically Korean cultural usage. So I'm not sure hemp fiber (Q13414920) with links to other languages should be broken (e.g. de:Hanffaser). Help! Bri (talk) 19:26, 19 April 2018 (UTC)

English Wikipedia does not have an article on (generic) hemp fiber (Q13414920), so there is no link there to 삼베. Hemp fiber is discussed in Hemp which is linked to industrial hemp (Q7150699) (edited: and in Fiber crop). Many languages have both articles (and it looks at a quick glance like some of them are not linked to the best WD item ... Catalan Fibra de cànem pretty clearly is about the fiber, not the industrial uses of the plant). As you point out, Enwiki sambe is specifically about cultural Korean usage, and I think its WD item Sambe (Q48724197) should probably be a subclass of hemp fiber (Q13414920). If the RFC to allow links to redirects is approved, the ENwiki redirect Hemp fiber would be pointed to hemp fiber (Q13414920) which would solve this problem.
If KOwiki allows manual interwiki links, you could link KO 삼베 to EN Hemp or sambe as you please. But as things stand, I don't know of a way to make this link via Wikidata. (Other than the manual hack that is sometimes used to link redirects to WD items).- PKM (talk) 19:53, 19 April 2018 (UTC)
Someone could also write a new, separate article on Hemp fiber for ENwiki - the current redirect goes to Fiber crop. - PKM (talk) 19:59, 19 April 2018 (UTC)
And it turns out we had two WD items for hemp fiber. I have merged them now. - PKM (talk) 20:33, 19 April 2018 (UTC)
Thanks everyone. I'll probably go ahead and write Hemp fiber for enwp soon. Bri (talk) 02:41, 20 April 2018 (UTC)

Xantus's Murrelet

See also User:Pigsonthewing @ Xantus's Murrelet (Q46338167) --Succu (talk) 21:38, 25 February 2018 (UTC)

We seem to be having a problem at Xantus’s murrelet (Q46338167), which User:Succu persists in repeatedly (five times, so far) trying to merge into one or another item about patently different concepts; or from which he removes cited statements. Given previous difficulties I and other editors have experienced when attempting to discuss similar matters with that user, I'm raising it here, and not on the item's talk page which presumably has no other watchers. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:16, 26 December 2017 (UTC)

And a sixth. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:34, 26 December 2017 (UTC)
Mind to count your own reverts too? The item was originally created for the eBird entry xanmur. This is about two species called en:Xantus's murrelet (= Scripps's Murrelet (Q3120531) and Xantus’s murrelet (Q1276043)). Then Mr. Mabbett added ABA bird ID (P4526) = xanmur, witch is referring only to the common name „Xantus's murrelet“ and a duplication of the value ARKive ID (archived) (P2833)=xantuss-murrelet/synthliboramphus-hypoleucus. Finally (after some reverts) he claimed taxon name (P225) = Synthliboramphus hypoleucus (=Xantus’s murrelet (Q1276043)) about this item. Maybe he could explain here, why and on what base he thinks this is a „patently different concept“. --Succu (talk) 21:00, 26 December 2017 (UTC)
I'm glad that Succu has confirmed that the item in question is about a different concept to the items to which he has variously redirected it (albeit he is confused as to why this is so; and about the edits I have made to the item). Perhaps he will now cease doing so? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:14, 26 December 2017 (UTC)
I'm confirming nothing. I asked for an explaination. --Succu (talk) 21:17, 26 December 2017 (UTC)
"The item ... about two species called en:Xantus's murrelet (= Scripps's Murrelet (Q3120531) and Xantus’s murrelet (Q1276043))". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:51, 26 December 2017 (UTC)
Hence the first merge. Is this item about two species? Would be nice if you could explain your viewpoint to other readers of this topic. --Succu (talk) 21:59, 26 December 2017 (UTC)
Your first merge was to an instance of Wikimedia disambiguation page (Q4167410). My viewpoint is that Q46338167 represents a different concept to any of those with which you have tried to merge it. I'm also sure "other readers" can read both the item's description, and the sources used. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:24, 26 December 2017 (UTC)
So no explaination at all, why your new item is a „patently different concept“. Different from other species? Repeat: Is this item about two species? Would be nice if you could explain your viewpoint to other readers of this topic. Looks like are unwilling to do so, Mr. Mabbett. --Succu (talk) 22:33, 26 December 2017 (UTC)
To make it easier fr you, is your new item Xantus’s murrelet (Q46338167) about:
  1. the two species Scripps's Murrelet (Q3120531) and Xantus’s murrelet (Q1276043)) supported by xanmur
  2. the common name „Xantus's murrelet“ supported by ABA bird ID (P4526) = xanmur
  3. the species name Xantus’s murrelet (Q1276043) supported by ARKive ID (archived) (P2833)=xantuss-murrelet/synthliboramphus-hypoleucus
If your answer is "all of them" (=current status) then please explain it to us. Thanks in advance. --Succu (talk) 22:58, 26 December 2017 (UTC)
No Succu, there's explanation aplenty. My reason for raising the matter here is to solicit third-party input. I won't be answering questions such as yours, based on false premises. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:24, 27 December 2017 (UTC)
Then please list my „false premises“ and explain me why they are wrong. But I don't think you have some real arguments. Otherwise it would be easy to you to give them. By the way: Do you think giving only a ISBN like 0198540329 is a sufficient source? --Succu (talk) 21:01, 29 December 2017 (UTC)
OK, I did another merge. --Succu (talk) 22:44, 30 December 2017 (UTC)
Mr. Mabbett? ISBN 0198540329 stands for what book? On which page does this ISBN supports your view? --Succu (talk) 21:14, 23 January 2018 (UTC)
It's a simple question, Mr. Mabbett. Mind to respond? ---Succu (talk) 21:54, 26 January 2018 (UTC)
I don't see a good reason to reply here and generally think it makes more sense to have such a discussion on the talk page by pinging relevant Wikiprojects. ChristianKl ❪✉❫ 12:59, 27 December 2017 (UTC)
I agree with ChristianKl. I must admit I am completely mystified with what concept Andy Mabbett has in mind. Certainly the item as it now is, seems inconsistent with any way of expressing any concept ever included in Wikidata so far. - Brya (talk) 05:36, 28 December 2017 (UTC)
You're "completely mystified" and - according to your comment on the item's talk page, are "guessing" what it represents; yet you see fit to make changes to the item, which are unsupported by the sources used (and you offer no new sources). That's not a healthy way to proceed. I have again fixed your broken indenting. Wilfully mis-indenting your comments, having been told that doing so is harmful, and having been given advice on how to do so correctly, is disruptive. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:34, 29 December 2017 (UTC)
The problem is that there is very little offered in the way of sources. I see just the ARKive ID link, and given how much junk we already suffered from that source, it is a frail reed to lean anything on.
        And please don't "fix" my comments: you should restrict your religious [?] beliefs to your own comments. - Brya (talk) 11:38, 29 December 2017 (UTC)
There are at least three sources used on the item; none of which are from ARKive. Please stop posting falsehoods. And like I said; disruptive. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:48, 29 December 2017 (UTC)
ARKive ID (archived) (P2833)=xantuss-murrelet/synthliboramphus-hypoleucus states this is Xantus’s murrelet (Q1276043). ---Succu (talk) 21:01, 29 December 2017 (UTC)
BTW, same is true for your weblink to the entry at US ECOS. --Succu (talk) 21:10, 29 December 2017 (UTC)

I've just undone a seventh attempt by Succu to delete this item through a merger to an inappropriate target. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:28, 30 December 2017 (UTC)

Please argue here and do not revert blindly. --Succu (talk) 07:05, 31 December 2017 (UTC)
And an eighth... Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:02, 31 December 2017 (UTC)
 Comment Perhaps it's time to find a source that they are actually different... Matěj Suchánek (talk) 21:10, 31 December 2017 (UTC)
You can find one on Xantus’s murrelet (Q46338167). HTH. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:18, 1 January 2018 (UTC)
Then the easiest way to settle the matter is to cite your "reference" here. --Succu (talk) 13:56, 1 January 2018 (UTC)
The American Ornithological Union checklists "are the official source on the taxonomy and nomenclature of birds found in this region, including adjacent islands." see here: http://www.americanornithology.org/content/checklist-north-and-middle-american-birds . Looking at the current checklist here: http://checklist.aou.org/taxa/ we have Synthliboramphus scrippsi (Scripps's Murrelet) and Synthliboramphus hypoleucus (Guadalupe Murrelet). Therefore I believe, officially Xantus's Murrelet has been split, I don't think what other authorities say is relevant. A species with name Synthliboramphus hypoleucus (Xantus's Murrelet) was deleted from the AOU list as per the 53rd supplement in 2012 http://americanornithologypubs.org/doi/pdf/10.1525/auk.2012.129.3.573 I'm not really a wikidata expert, but I would suggest the best course of action is to retain Xantus’s murrelet (Q46338167) but change the instance of (P31) from taxon (Q16521) to something which indicates that this is a formerly recognised taxon, but which has been deleted. I had a quick look but couldn't find an item that would describe that, but this must have happened before. Species are split all the time. I don't really think that Xantus’s murrelet (Q46338167) should be merged into Xantus’s murrelet (Q1276043) they are different. Just my twopenneth. JerryL2017 (talk) 15:23, 1 January 2018 (UTC)
Wikidata follows a NPoV policy, not a Single Point of View policy; for that try Wikispecies. So, the American Ornithological Union checklists are only one source, not THE source. Of course, it may be possible to start creating items based only on American Ornithological Union concepts, but this would be a fairly big departure from existing practice. - Brya (talk) 05:37, 2 January 2018 (UTC)
We do not model different taxon concepts this way. Thats why I merged the items several and was asking for a good reference to proceed. None was given. --Succu (talk) 16:04, 2 January 2018 (UTC)
Here is the defining reference that concludes Xantus’s murrelet (Q46338167) is 2 species: http://www.bioone.org/doi/full/10.1525/auk.2011.11011 based on that paper the AOU adopted that taxonomy as detailed in the 53rd supplement, http://americanornithologypubs.org/doi/pdf/10.1525/auk.2012.129.3.573 which I had already given above. However, given that not all sources have yet adopted this taxonomy, and based on what others have said here and what is stated in the wikidata taxonomy project guidance it would seem sensible to retain Xantus’s murrelet (Q46338167) for the time being, with the correct links to sources that are still using the former taxonomy. That said, there are issues with Xantus’s murrelet (Q1276043). This item refers to the "split" Guadaloupe Murrelet but has links to sources that do not recognise the split. It also includes the alternative name of Xantus's Murrelet, which is confusing. JerryL2017 (talk) 17:44, 2 January 2018 (UTC)
Rangewide population genetic structure of Xantus's Murrelet (Synthliboramphus hypoleucus) (Q29541111) is proposing a taxonomic opinion about elevating the two subspecies Synthliboramphus hypoleucus hypoleucus (Q47012916) and Synthliboramphus hypoleucus scrippsi (Q47012925) of Xantus’s murrelet (Q1276043) to species level. The American Ornithological Society (Q465985) was following the recommendation. I do not see Xantus’s murrelet (Q46338167) is expessing this. --Succu (talk) 18:37, 2 January 2018 (UTC)
Yes, whatever the intent is, execution seems sloppy. - Brya (talk) 04:01, 3 January 2018 (UTC)
Since Mr. Mabbett refuses to argue here I will merge both items once again. --Succu (talk) 19:27, 5 January 2018 (UTC)
And if you do, absent a consensus here, you will be reverted again; for the reasons already given. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:01, 5 January 2018 (UTC)
I expected this. Other people try to argue. Your are not. What a pitty for you. Hopefully you do not miscount your reverts. --Succu (talk) 22:10, 5 January 2018 (UTC)
And this revert of him. --Succu (talk) 07:21, 6 January 2018 (UTC)
And this revert of him. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:35, 6 January 2018 (UTC)
Obviously you are unwilling to give a reasonable answer here. Probably you can't and are defending the item only because you've created it. --Succu (talk) 21:02, 6 January 2018 (UTC)
No progress here from Mr. Mabbetts side. --Succu (talk) 21:01, 8 January 2018 (UTC)
Now at AN. --Succu (talk) 20:59, 9 January 2018 (UTC)
ok from the point of view of a taxonomist. Reading the paper from 2011 linked above by @JerryL2017: the two taxa in question were considered subspecies, however overlap and do not interbreed in sympatry. By the definition of a subspecies this is not possible, hence they should be species and have been recommended as such by the paper also. As such from this viewpoint you have two species and should have two items one for each. Any other refs, unless you find one that refutes this primary ref with data not opinion, are irrelevant. I see no reason for any further argument. The nomenclatural act has been made, follow it. Where the common names go whatever, they are vernacular names and not relevant to the concept of the species. That is my view on this so I would suggest fixing the pages to reflect this and as for the IOC, ummm they are not a primary taxonomic reference so why would you be adamant about it. Cheers Scott Thomson (Faendalimas) talk 21:46, 13 January 2018 (UTC)
All major bird checklist (including IOC) followed this viewpoint. The "official" english common name of Synthliboramphus hypoleucus was changed from „Xantus’s Murrelet“ to „Guadalupe Murrelet“. --Succu (talk) 22:57, 13 January 2018 (UTC)
Xantus’s murrelet (Q46338167) represents a concept, described by the three reliable source used on that item, which we can refer to, for the sake of brevity as "A". You are saying that a different source refers to the concept "B". The Wikidata model, as I understand it, is that to concepts should be represented by different items, (with, if applicable, mutual "said to be the same as" properties). However, If your contention is that "A" and "B" are the same concept, but with different attributes, then the Wikidata model is to include properties with values stating both attributes, cited to their respective sources. What the Wikidata model does not do, is to pretend that the (reliably-cited) concept "A" does not exist. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:17, 13 January 2018 (UTC)
Name it (the concept)! --Succu (talk) 23:23, 13 January 2018 (UTC)
Umm..... I am seriously not getting this. Ok three reliable sources, I looked on the page I did not see anything I would consider as reliable. Please correct me if I am wrong, do you have separate pages for the scientific name and the vernacular names?? I do not understand why you would do that, but no matter. The vernacular name needs to follow the scientific name according to the most recent primary sources. No I do not consider the collective political opinions of eBird or the IOC nd simalar checklists as primary sources. When the species was split I imagine this did require modification of the accepted common names. As such you would use the primary refs that state this coupled with the research paper that did the nomenclatural act to move the common names accordingly. By the way, a taxon is not a concept it is an hypothesis, the concept is the theory of how species are differentiated, ie the grounds for calling something a species. That is for example the Biological Species Concept. However a species is a hypothesis circumscribed under a concept of the primary authors choosing. So please what are we getting t here cause your paragraphh Andy makes no sense. You are reporting science, it needs to reflect science. Cheers Scott Thomson (Faendalimas) talk 23:48, 13 January 2018 (UTC)
OK, I will try and summarise and make some recommendations. Up until about 2012/2013 the ornithological world recognised 5 species in the genus Synthliboramphus (Q287293) . One of these species had the latin name Synthliboramphus hypoleucus, common name Xantus's Murrelet. This is Xantus’s murrelet (Q46338167). In 2012, in the paper cited above, it was proposed that the 2 recognised sub-species of Xantus’s murrelet (Q46338167) (Synthliboramphus hypoleucus scrippsi (Q47012925) and Synthliboramphus hypoleucus hypoleucus (Q47012916) be considered separate species and that Xantus’s murrelet (Q46338167) be deprecated. This hypothesis was accepted by the American Ornithological Union (they have a scientific committee which decides on such matters) and their taxonomy was updated so that the genus Synthliboramphus has 6 recognised species including: Synthliboramphus hypoleucus, common name Guadalupe Murrelet (Xantus’s murrelet (Q1276043) AND Synthliboramphus scrippsi, common name Scripps's Murrelet (Scripps's Murrelet (Q3120531)). Subsequently a number of other sources have also adopted this new taxonomy (and I make no comment about how reliable these might or might not be) e.g Avibase ID (P2026), IUCN taxon ID (P627), eBird taxon ID (P3444) and others. However a number of others, that are variously cited in wikidata's items have not adopted the new taxonomy. How do I know they have not adopted the new taxonomy? Because when you look at it in detail it does not include Scripps's Murrelet (Q3120531). Examples of these include: NCBI taxonomy ID (P685)), ITIS TSN (P815) and WoRMS-ID for taxa (P850) although there are probably others, I haven't yet reviewed all sources. So, why have these sources not adopted the new taxonomy recommended by experts in the field? I can see 3 reasons: 1) They don't agree with the hypothesis that Xantus's Murrelet is actually 2 separate species - I believe this is unlikely in this case) 2) They haven't updated their taxonomy (highly likely) and over time will probably update their data. 3) They have a scientific requirement to retain Xantus's Murrelet in their taxonomy because of pre-existing data that was linked to Xantus's Murrelet and not either of the sub-species. A good example of this is eBird taxon ID (P3444) where their taxonomy has Guadalupe AND Scripps AND (by their nomenclature) an entity called Scripps/Guadalupe Murrelet (Xantus's Murrelet). So, my conclusion is that wikidata should retain Xantus’s murrelet (Q46338167), some sources are still using the former taxonomy and some bona fide sources WILL ALWAYS have an entity that is best described as Xantus's Murrelet. However, having said that there are some outstanding questions we should resolve 1) For Xantus’s murrelet (Q46338167) what is the best instance of (P31). is it taxon or some other concept that better captures its status. 2) The existing wikidiata items should be updated to ensure each item is linked to sources that best reflect that item i.e. Xantus’s murrelet (Q1276043) should NOT link to sources that have a taxonomy that does not inlcude Scripp's Murrelet. 3) Other data such as images and synomyms should be updated to make it less confusing. (NOTE: if we have agreement on this I am happy to go and make these changes IF I have assurance that they won't just all be reverted). 4) Consideration should be given within the wikidata taxonomy project for additional avian taxonomy properties (some key sources are not available as properties, which isn't helping here) 5) I've seen various comments on here that some sources are "unreliable" if there is a consensus that they are unreliable then why do we retain them? 6) IF we can resolve this case, consideration should be given as to how this is more widely applied within wikidata - for instance it is very common practise across many data recording schemes to "combine" species into groups when they are difficult to identify. I personally think wikidata should be able to reflect that, we are a broad data resource not a wildlife taxonomy. Thanks, I hope this is useful and moves the discussion on a little. JerryL2017 (talk) 09:58, 14 January 2018 (UTC)
Maybe just a slip of a pen, JerryL2017: „the genus Synthliboramphus has 6 recognised species“ but I count only five... --Succu (talk) 23:16, 14 January 2018 (UTC)
Sorry Succu, you are correct.

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ok what I have been trying to say is the taxonomic view of the situation, then it is for wikidata to determine how best to reflect the science. In answer to your questions from my point of view. This is a wonderful example of why common names are a pain, unprofessional and honestly useless. I agree with you @JerryL2017: that the issues of different nomenclatures between sources is most likely lack of updating.
1. There are three common names available for 2 species the names Xantus's Murrelet and Guadalupe Murrelet both apply to the scientific name Synthliboramphus hypoleucus however the former is now considered depreciated, and Scripp's Murrelet which applies to Synthliboramphus scrippsi. When the species was split the current common name goes with the species that retained the original combination. There is no justification in retaining the common name Xantus's Murrelet for Synthliboramphus scrippsi, or honestly at all. The name Xantus's Murrelet does not technically apply to a taxon anymore, it is considered depreciated, it is at best an outdated name of historic value only.
2. Agreed, removal of sources that are outdated in their nomenclature will avoid confusion, or if they must be stated have it as "stated in and as" so the the page is generally set up with the current nomenclature but make note of any departures from it without supporting them.
3. Agreed, all information possible should be updated, including where necessary file names and metadata for images. I would not revert anything. Cannot speak for others. But I think if we hammer out an agreed position I believe people are professional enough to follow it.
4. Avian taxonomy should honestly be following the ICZN code, which they do not. Further they have made recent efforts to dictate to other fields of taxonomy that their viewpoint should be followed, to a massive backlash. However, this is not our problem, we present the science we do not revise it. If you feel the need for further avian properties please elucidate these.
5. Your guess is as good as mine. I think there is a generalised tendancy in projects like this to grab every online reference possible, unfortunately with little consideration of the quality of what is presented and no fact checking. Basically the equivalent of google says this, it must be true. Again I think this is also unprofessional, I think sources that are questionable should be examined by wikidata taxonomy project for validity and if rejected they are removed.
6. The act of combining species into groups is I think beyond the realm of a database. This is done through analysis of the given issue. Wikidata should be presenting the data, with reliable and good sources. As best as possible the primary taxonomic literature, in the scope of this issue, the thing that can come out of this is a better discussion on what is a good resource, the acceptance that complex cases need to be analyzed using only primary references, and that in taxonomic issues the relevant codes are the primary determinant on availability and validity. That is, if a name is published in accordance with the Code it is to be accepted as valid or refuted, a point the avian taxonomists breach the code on repeatedly.
Cheers Scott Thomson (Faendalimas) talk 15:14, 14 January 2018 (UTC)

Yes, it was established early on that in the real world there are/have been two circumscriptions for Synthliboramphus hypoleucus. That is not a problem. There appear to be several problems:
  • Is this wider/older circumscription notable enough to rate a separate item of its own?
  • Is this wider/older circumscription indeed what Andy Mabbett intends with Q46338167, given that he has already denied this. If not, what does he intend?
  • Is it worthwhile discussing if Wikidata should have items for concepts denoted by a standardised common name set by some bird organization? These clearly exist, but are they notable enough?
  • Given that bird organizations use deviant 'scientific names' with rules of their own, should we have a property for that? Something like "Avian scientific name" or more general "deviant taxon name, used by special interest groups" (to include butterflies)? Clearly, it is not a good idea to put non-Code-compliant names in P225.
Brya (talk) 18:10, 14 January 2018 (UTC)
I agree with your analysis @Brya: specifically:
Is this wider/older circumscription notable enough to rate a separate item of its own? I do not think so.
Is this wider/older circumscription indeed what Andy Mabbett intends with Q46338167, given that he has already denied this. If not, what does he intend? My impression was that this is what is being suggested here, I also acknowledge Andy has denied this, but I have no idea what the purpose of this is in that case.
Is it worthwhile discussing if Wikidata should have items for concepts denoted by a standardised common name set by some bird organization? These clearly exist, but are they notable enough? I do not think they are notable in any great degree, unless they are a highly notable list. I would encourage the avoidance of confusion as a priority.
Given that bird organizations use deviant 'scientific names' with rules of their own, should we have a property for that? Something like "Avian scientific name" or more general "deviant taxon name, used by special interest groups" (to include butterflies)? Clearly, it is not a good idea to put non-Code-compliant names in P225. I prefer something along the lines of your second option, since it can be applied outside Aves (Birds) this can apply to Amphibians also. But I definitely agree anything that is non code compliant should be avoided in almost any circumstance. Cheers Scott Thomson (Faendalimas) talk 20:43, 14 January 2018 (UTC)
All concepts are still in use at Wikimedia projects: de:Lummenalk is about the old concept, en:Guadalupe murrelet is about the new concept. What e.g. is species:Synthliboramphus hypoleucus is about remains unclear to me. So how to deal with #2) Where should we place a) outdated Wikimedia articles, b) outdated external identfiers (in case we can judge they are) And yes, this thread needs some insights by Mr. Mabbett. --Succu (talk) 23:01, 14 January 2018 (UTC)
The Wikispecies account is about the species in question as part of this, we do not worry so much about common names there as its not really what we are about. Vernacular names get added by people occasionally as they see fit, I ignore them as best as I can. If someone wants to add the english common name they can. Cheers Scott Thomson (Faendalimas) talk 00:54, 15 January 2018 (UTC)
I do not care much about common names, but I care about references. The Wikispecies article has only a reference to the origninal combination Brachyramphus hypoleucus (not mentioning it at all). So it's hard to know about which taxon concept this entry is. :( --Succu (talk) 19:30, 16 January 2018 (UTC)
Fair enough, sorry I do not do the birds. I would not know the relevant refs. When I do turtles I already have pretty much all the literature, so is easier for me. However, since species:Synthliboramphus scrippsi also exists, then the other species can only be considering the new combination. Cheers Scott Thomson (Faendalimas) talk 00:17, 17 January 2018 (UTC)
But this only a guess. Even the genus species:Synthliboramphus has no reference to a current taxonomic treatment... --Succu (talk) 20:38, 19 January 2018 (UTC)

Any suggestions how to resove this „probem“? Mr. Mabbett is not responding. --Succu (talk) 22:04, 30 January 2018 (UTC)

I will update wikispecies with the appropriate refs to show the two currently accepted species, and with a common name for each, with the relevant references (if anyone has them please send them to me) I need the original descriptions of both species and the treatment that recognises them as currently valid species. If you wish then you can use this to model your database entries on this. That is up to you. My suggestion is to follow what has largely been discussed here, in the absence of any other explanation. So I suggest you have data entries for each of the two species with the currently accepted common names for each taxa, with the original refs. I further suggest that you could delete the entry for the now depreciated common name and list it only as an alternative, older, no longer used name for the species Synthliboramphus hypoleucus in older treatments. Cite the paper that split them as species for justifying this. I would suggest calling the two data entries by the scientific name with the common names as description. This way the taxa are clearly defined and it can be noted the common names are less clear. Just my suggestions, your call. Cheers Scott Thomson (Faendalimas) talk 22:17, 30 January 2018 (UTC)
Please refer to my reply to you (so much for not responding!), above, time-stamped "23:17, 13 January 2018 (UTC)". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:13, 30 January 2018 (UTC)
Your usual gibberish, Mr. Mabbett. You didn't responded to some question directed to you (what concept, give a page). --Succu (talk) 22:57, 31 January 2018 (UTC)
OK, I will merge both items again. --Succu (talk) 19:07, 5 February 2018 (UTC)
If you do, I will revert you, because nothing here refutes my reason for doing so previously. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:16, 5 February 2018 (UTC)
Why would you revert? I thought consensus was what was followed here. Cheers Scott Thomson (Faendalimas) talk 12:56, 6 February 2018 (UTC)
So, why would you revert, Mr. Mabbett? Is nothing here refutes my reason for doing so previously an argument? --Succu (talk) 22:51, 7 February 2018 (UTC)
As you can see, my arguments are laid out above. As a courtesy to our fellow editors, I see no need to repeat them. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 23:16, 7 February 2018 (UTC)
As a courtesy to our fellow editors [...], Mr. Mabbett? An „interesting argument“. I'm one of your fellow editors. I don't think you have an answer to my questions, because you are avoiding to give a unequivocally answer for weeks now. --Succu (talk) 22:10, 8 February 2018 (UTC)
Done. --Succu (talk) 22:55, 8 February 2018 (UTC)
This was again reverted by Mr. Mabbett with the comment „per project chat“. --Succu (talk) 20:21, 14 February 2018 (UTC)
Looks like we have to wait till eternity, to get an explaination by Mr. Mabbett. --Succu (talk) 18:40, 20 February 2018 (UTC)

Despite the above, Succu's latest reason for reverting me was "no argument given". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:10, 24 February 2018 (UTC)

Where above? So what's this item about? You didn't tell us. --Succu (talk) 13:32, 24 February 2018 (UTC)
Still no explaination, by Mr. Mabbett. --Succu (talk) 21:53, 24 February 2018 (UTC)
Still no clarification by Mr. Mabbett. --Succu (talk) 06:11, 2 March 2018 (UTC)

I moved the value of ARKive ID (archived) (P2833). Could you please explain your revert, Mr. Mabbett? --Succu (talk) 19:14, 6 March 2018 (UTC)

Too busy to answer, Mr. Mabbett? --Succu (talk) 16:13, 13 March 2018 (UTC)
ARKive ID (archived) (P2833)=xantuss-murrelet/synthliboramphus-hypoleucus states „Two subspecies of Xantus’s murrelet are recognised“ And this is Xantus’s murrelet (Q1276043). So please undo your revert or give a reasonable explaination, Mr. Mabbett. --Succu (talk) 15:21, 20 March 2018 (UTC)
OK, then I will do it for you. --Succu (talk) 14:47, 27 March 2018 (UTC)
"Mr. As before" reverted this again. No reason was given. When do you start to argue, Mr. Mabbett? --Succu (talk) 21:36, 30 March 2018 (UTC)
As before: no answer was given by Mr. Mabbett! --Succu (talk) 17:30, 5 April 2018 (UTC)

I stumbled into this one (page-searching for a certain word). I am astonished to see that Wikidata has not been able to solve this December 25 thread? Does not look inviting. 19:49, 5 April 2018 (UTC) -  – The preceding unsigned comment was added by DePiep (talk • contribs) at 19:49, 5 April 2018‎ (UTC).

He likes housekeeping (=unsigned). The answer is easy if Mr. Mabbett would cooperate. --Succu (talk) 20:20, 6 April 2018 (UTC)

Nine days ago I asked User:Rschen7754 how to proceed in cases like this. No help at all. --Succu (talk) 14:31, 13 April 2018 (UTC)

Another week: no progress. --Succu (talk) 13:27, 20 April 2018 (UTC)

Standardizing historic patents, post #2

We could start to standardize on an ontology for pre-1920 patents, as I mentioned in a Project Chat post a week ago. Interest was limited. And yet I would like to create a Wikidata WikiProject for this topic, analogous to the ones for companies, organizations, universities, and occupations. Any objections? Or advice?

I'd like to get started now. In two weeks I have a meeting with specialists at WIPO, the UN organization that manages modern-day patent rights, and would like to get their support or a wise critique of the effort. Some of their data could be uploaded eventually, and on the wiki side we might be able to set up automatic links analogous to the authority control templates so as to get immediately from a Wikipedia article about an individual to their patent documents across countries, a service not offered very well elsewhere for historic patents. Modern patents are conceptually the same, but they refer to intellectual property rights that may still be relevant, or maybe they might be copyrighted, and they are longer and more difficult to read, AND there are more services designed to manage that sort of information which might be technologically relevant now. Therefore I think it is reasonable to start with the older ones. -- econterms (talk) 22:53, 16 April 2018 (UTC)

Go ahead and start the wikiproject, at least some of those in the other ones you link are probably interested. Is there any particular reason historic patents would be different from modern patents in terms of data structure/ontology? ArthurPSmith (talk) 12:32, 17 April 2018 (UTC)
Thanks kindly, ArthurPSmith! I will start it. Historic patents can be a little different. (1) They may have been issued by governments that do not now exist, and therefore do not have standard two-letter codes. E.g. Austro-Hungarian Empire, and the German and Italian states that existed before those nations unified to their current form. (2) Before legal standardization there might be more variation in the data in them although I have not noticed much of this. But the earliest ones did not always refer to technologies or inventions but to monopolies granted, e.g. by the British crown, or monopolies on the rights to import some particular kind of thing. (3) they had different term lengths from the modern standard. This is not a long list but if we dig there would be more. -- econterms (talk) 13:22, 20 April 2018 (UTC)
Also, for old patents, we can be sure they are not copyrighted any more and that they have no intellectual property traction any more ; that is, they do not represent legally relevant claims any more so we could treat them a little more informally. -- econterms (talk) 17:48, 20 April 2018 (UTC)

REST API format for wikidata

I'm playing with the REST API for wikidata. The returned results for an item in html form (like http://www.wikidata.org/api/rest_v1/page/html/Q42) just seem to be a dump of the JSON data associated with the item. I presume if people wanted this, they would go to e.g. https://www.wikidata.org/wiki/Special:EntityData/Q42.json. I wonder if, like the wikipedia REST pages (e.g. https://en.wikipedia.org/api/rest_v1/page/html/Albert_Einstein), the data could be arranged in a neater, HTML-like way, perhaps even something like the tabular format seen on a normal "editing" page (i.e. https://www.wikidata.org/wiki/Q42), or perhaps with section headers and breaks inserted. HYanWong (talk) 14:48, 12 April 2018 (UTC)

You should probably talk to those who maintain this API. I'm not sure who is responsible, though. Matěj Suchánek (talk) 13:26, 14 April 2018 (UTC)
I'll try posting a feature request on Phabricator. Thanks! (edit: just done so at https://phabricator.wikimedia.org/T192687) HYanWong (talk) 20:35, 20 April 2018 (UTC)

Synalpheus pinkfloydi (Q29367343) and Synalpheus pinkfloydi sp. nov., a new pistol shrimp from the tropical eastern Pacific (Decapoda: Alpheidae) (Q29390847)

To keep track of this ongoing dispute with Mr. Mabbett:

  1. In Synalpheus pinkfloydi (Q29367343) he readded a constraint violation.
  2. In Synalpheus pinkfloydi sp. nov., a new pistol shrimp from the tropical eastern Pacific (Decapoda: Alpheidae) (Q29390847) he removed heavily used statements.

Any idea what could be done? --Succu (talk) 21:16, 13 April 2018 (UTC)

Bad constraints should be removed fixed. I didn't remove anything from the latter item. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:27, 13 April 2018 (UTC)
Of course you removed statements. Why is the constraint bad? --Succu (talk) 21:30, 13 April 2018 (UTC)
The constraint is bad for the reasons explained, at length, on Wikidata:Project chat/Archive/2017/07#Editwar at Desmopachria barackobamai (Q30434384) and especially its 'named after' sub-section; see the 'Kentish Plover' example therein. Anyone can see that I removed nothing in these edits, but feel free to explain what you believe I did. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:37, 13 April 2018 (UTC)
The widly used statment volume (P478) in Synalpheus pinkfloydi sp. nov., a new pistol shrimp from the tropical eastern Pacific (Decapoda: Alpheidae) (Q29390847) was removed by you. Once again: Why is the constraint bad? Feel free to summarize your arguments for everybody here. Danke. --Succu (talk)
Nope; despite your selective quoting of one diff in a run of several, the volume, 4254, is still shown, quite clearly, on that item. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:10, 13 April 2018 (UTC)
  1. So once again you can't.
  2. Ouf of 12227 atricles we have for Zootaxa (Q220370), only one is lacking the statements you removed.
--Succu (talk) 05:43, 14 April 2018 (UTC)
@User:Rschen7754, am I allowed to restore the correct values? --Succu (talk) 20:45, 15 April 2018 (UTC)
If there is a disagreement between you two, ask others for comment. --Rschen7754 20:56, 15 April 2018 (UTC)
I asked you, User:Rschen7754 as a blocking admin. --Succu (talk) 21:03, 15 April 2018 (UTC)
I have done so; more than once. You'll note, for example, that user:ChristianKl said: "Items can have different names in different languages and also multiple names in the same language. "Named after" is a property of a specific name of an item and therefore I agree that it makes more sense as a qualifier.". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:17, 15 April 2018 (UTC)
As if Andy Mabbett is going to be stopped by a consensus ... - Brya (talk) 04:03, 21 April 2018 (UTC)

I have similarly restored the named after (P138) qualifier on Myrmoteras mcarthuri (Q13871073), which was changed with no explanation and no discussion. My reasons for doing so are as explained Wikidata:Project chat/Archive/2017/07#named after. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:04, 15 April 2018 (UTC)

To cite you: „This still requires a resolution”, but you didn't help to resolve issue #1. --Succu (talk) 21:30, 15 April 2018 (UTC)
I did; I pointed out that "Bad constraints should be fixed". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:11, 15 April 2018 (UTC)
In the thread I cited: no! And reciting yourself again and again without providing a new perspective is useless. --Succu (talk)

Any idea what could be done? --Succu (talk) 21:06, 17 April 2018 (UTC)

Please note this related confusion about Revision of the Neotropical water scavenger beetle genus Quadriops Hansen, 1999 (Coleoptera, Hydrophilidae, Acidocerinae) (Q42255057). --Succu (talk) 21:12, 18 April 2018 (UTC)

Help wanted on humans having ISNI but no VIAF

The query shows humans that have an ISNI but no VIAF. Some are not in the VIAF DB, but those that have a GND ID or LC ID very likely are in the VIAF system. 

SELECT ?item ?itemLabel ?isni ?gnd ?loc
WHERE {
  ?item wdt:P31 wd:Q5 .
  ?item wdt:P213 ?isni .
  MINUS {?item wdt:P214 [] .}
  OPTIONAL {?item wdt:P227 ?gnd.}
  OPTIONAL {?item wdt:P244 ?loc.}
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE]". }
}
ORDER BY DESC(?gnd) DESC(?loc)
Try it!

I already fixed some, but it is a lot of work. The way I did it, is to click on the ISNI and on the ISNI page click on VIAF. Or when isni.org is down, to search on viaf.org 92.229.165.74 14:15, 19 April 2018 (UTC)

Do you know the "Authority control" gadget? It may also help with this task. --Anvilaquarius (talk) 21:20, 20 April 2018 (UTC)
If if helps, this version of the query will apply the formatter URL in the query result so you don’t have to open the item page to open the links: 
SELECT ?item ?itemLabel ?isni ?isniLink ?gnd ?gndLink ?loc ?locLink
WITH {
  SELECT ?item ?itemLabel ?isni ?gnd ?loc WHERE {
    ?item wdt:P31 wd:Q5 .
    ?item wdt:P213 ?isni .
    MINUS {?item wdt:P214 [] .}
    OPTIONAL {?item wdt:P227 ?gnd.}
    OPTIONAL {?item wdt:P244 ?loc.}
    SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE]". }
  }
} AS %results
WITH {
  SELECT * WHERE {
    wd:P213 wdt:P1630 ?isniFormatterUrl.
    wd:P227 wdt:P1630 ?gndFormatterUrl.
    wd:P244 wdt:P1630 ?locFormatterUrl.
  }
} AS %urls WHERE {
  INCLUDE %results.
  INCLUDE %urls.
  BIND(IRI(REPLACE(?isniFormatterUrl, "\\$1", ?isni)) AS ?isniLink)
  BIND(IRI(REPLACE(?gndFormatterUrl, "\\$1", ?gnd)) AS ?gndLink)
  BIND(IRI(REPLACE(?locFormatterUrl, "\\$1", ?loc)) AS ?locLink)
}
ORDER BY DESC(?gnd) DESC(?loc)
Try it!
--TweetsFactsAndQueries (talk) 16:05, 19 April 2018 (UTC)

TweetsFactsAndQueries - Thanks a lot! It makes the process a bit faster. Interesting what can be done with WDQS. 92.229.165.74 16:35, 19 April 2018 (UTC)

Paper size tolerances

Can the value of height (P2048) and width (P2049) in items such as A0 (Q12062327), A1 (Q12062328) etc. also include the permissible tolerance as specified in ISO 216 (±1.5 mm for dimensions ≤ 150 mm, ±2 mm for dimensions > 150 mm and ≤ 600 mm, ±3 mm for dimensions > 600 mm)?

I would state, for instance, that A1 has height (841 ± 3) mm and width (594 ± 2) mm, but is A1 (Q12062328) specifically about the ISO size? DIN 476 used to specify slightly tighter tolerances (±1 mm, ±1.5 mm, ±2 mm) – though the part DIN 476-1 about the A and B series has been withdrawn in favor of DIN EN ISO 216, so perhaps this is obsolete – and there may be other standards with different tolerances, so should I use a qualifier (which one?) or would that be in a separate item? We already distinguish between B series (Q43305767) and JIS B series (Q43351400), but these already differ by the “ideal” sizes. -- IvanP (talk) 23:14, 13 April 2018 (UTC)

@IvanP: yes to the first question, as easily as changing the value from 1,189 to 1,189±3 such as in this (now reverted) edit. See also Help:Data_type#Quantity. I don't have a good answer for ISO vs DIN tolerance diffs; could be done as 2 items, could be done as 1 item with 2 measurements, each referencing the pertinent standard (and there may be an appropriate additional qualifier for the measurement & other claims, pointing to the standard, such as of (P642)?). --Tagishsimon (talk) 23:20, 13 April 2018 (UTC)
Regarding the qualifier, I would not use statement supported by (P3680) since the standard defines the tolerances rather than just supporting the statement that they were set this way. I am unsure about criterion used (P1013); I would use it to make clear which measure is meant exactly (e.g., height (P2048) (which height?) → height to pinnacle (Q26970842)), but in this case, the purpose is to make clear what kind of paper size is meant (namely, A0/A1/… as defined in ISO 216). Is of (P642) the way to go? Would love opinions. -- IvanP (talk) 09:03, 14 April 2018 (UTC)
I suggest creating two items both with label "A1", and a description noting which item is for the ISO variant, and which item is for the DIN variant of the paper size specification. Then use stated in (P248) in references on each item, pointing to the specific ISO or DIN standard. Values of properties such as height (P2048) and width (P2049) would therefore be different on each of the two items due to the variance in tolerances allowed. I would also suggest a third item with label "A1" that is a disambiguation/class item, with the ISO and DIN items being instances of this item. Properties such as height (P2048) and width (P2049) would not be appropriate to use on this third item. Dhx1 (talk) 12:28, 15 April 2018 (UTC)
@Dhx1 Ah, stated in (P248)! I am not happy with of (P642) either, because I would use it in relation to a property value, e.g., position held (P39) mayor of a place in France (Q382617) (of (P642) Rochefourchat (Q323644)), which would mean maire of Rochefourchat. In our case, we have a name which may have multiple meanings. In the end of May, Wikidata should support Lexemes for lexicographical data, but this is not quite what we need here, since we want to describe the name language-independent. Should we make a disambiguation item for kilobyte, megabyte etc. as well, then? conversion to standard unit (P2442) currently has two values in terabyte (Q79741), 1,000,000,000,000 byte and 1,099,511,627,776 byte. The first one concerns the decimal terabyte, the second one the binary terabyte (tebibyte). We already have, e.g., quintillion (Q41650359) for the number name, where it says facet of (P1269) 10^30 (Q3115521) (valid in place (P3005) long scale (Q19202121)) and facet of (P1269) 10^18 (Q319582) (valid in place (P3005) short scale (Q19202120)). However, the long and short scale are not geographical objects (cf. value type constraint!). (@Bigbossfarin) Besides, what to do about sitelinks? The German Wikipedia article Quintillion does mention the short-scale value, but its topic is really the long-scale value, conforming to the meaning of the name in German. The article says:
Die Bezeichnung ist nicht identisch mit quintillion im US-amerikanischen Englisch, das unserer Trillion (1018) entspricht.
The denotation is not identical to quintillion in US-American English, which corresponds to our trillion (1018).
So should the article be linked in quintillion (Q41650359) – as it is now – or rather 10^30 (Q3115521)? In contrast, the Kazakh article Квинтиллион is linked in 10^30 (Q3115521), but is actually about the short-scale value (10^18 (Q319582)), while also mentioning the long-scale one (it says something like [I do not speak Kazakh]: “Quintillion – a one followed by 18 zeros, i.e., 1018. In some countries, a quintillion is 1030.”). I have no concerns with the English article Billion being linked in billion (Q41650356), though. -- IvanP (talk) 12:26, 21 April 2018 (UTC)
Hi, I was correcting the mistake with Квинтиллион. I would propose to replace valid in place (P3005) by applies to part (P518) in number names, any other opinions? Sitelinks about numbers & names have to be checked individually what it belongs to sooner. In the case of Quintillion I would say it is more confortable for reeders & editors as it is right now. Greetings Bigbossfarin (talk) 12:47, 21 April 2018 (UTC)
@Bigbossfarin The long/short scale is not really a part of the name quintillion, it is a system which gives the name a meaning. Maybe we need a new property … “In the case of Quintillion I would say it is more confortable for reeders & editors as it is right now.” – Should we not go by the subject item rather than the name? For instance, kerosene (Q76904) has a sitelink to the German article Petroleum rather than Kerosin. -- IvanP (talk) 14:37, 21 April 2018 (UTC)

What property is best to use for a "seat number" qualifier?

Is there any property that could be used as "seat number" to be used as a qualifier for the property member of (P463) (and maybe position held (P39))? I want to put that as a qualifier for the statement that a human (Q5) is member of (P463) of Swedish Academy (Q207360). Swedish Academy (Q207360) has 18 distinct "seats" numbered from 1 to 18, and I want to be able to make a query listing all persons that have been the "holder" of a specific "seat", e.g. seat number 7. The same could be used for various parliaments that have numbered individual seats. I guess a possibility is to create 18 separate objects, one per specific "seat", but such a scheme may not be feasible for parliaments with maybe hundreds of individual seats. So I would prefer to use a qualifier.

None of the properties in the following list seems to be a perfect choice, so I would appreciate a recommendation on what to use as "seat number" qualifier.

Property Comment
number of seats (P1342) No, that is not for a specific seat.
number of seats in assembly (P1410) No, that is not for a specific seat.
collection or exhibition size (P1436) No, that is not for a specific seat.
series ordinal (P1545) This is the "least bad" that I have found,
but it could be confused as a series ordinal (P1545) for the entire "academy/parliament" and not as the number for a specific "seat".
sport number (P1618) Possibly, but a hockey player and a member of an academy/parliament seem rather different.
road number (P1824) No, that would be confusing.
serial number (P2598) Possibly, but maybe confusing.
fleet or registration number (P2802) Possibly, but maybe confusing.
channel number (P3970) No, that would be confusing.

--Larske (talk) 09:23, 14 April 2018 (UTC)

It seems noone has a recommendation on what property to use as a qualifier for "seat number" to be used in a claim of member of (P463) a parliament like entity with individual and numbered seats. I will use series ordinal (P1545) then. --Larske (talk) 10:58, 21 April 2018 (UTC)

Catalog

Hoi, it is obvious that the use of "catalog" for projects like the "Black Lunch Table" is something that is to be phased out. These statements have qualifiers. I do not know how to do the conversion. So could someone help out and also show me how to do this? Thanks, GerardM (talk) 16:53, 19 April 2018 (UTC)

The conversion can most efficiently be done with bot code, so that no qualifier, reference, or any other detail like ranks get lost. You probably talk about a move from catalog (P972) to on focus list of Wikimedia project (P5008), right? Since this affects somewhat over 1000 items, it would take roughly 3 hours to move all of them at 6 edits per minute, a typical edit rate with the PAWS tool. This script would exactly do the move; if you want to move by yourself, log in to PAWS, start your server, create a new Python notebook, copy the code in the grey box from my linked example into the Python notebook and click “run”. It does not need any specific rights for your Wikidata account. These pages in enwiki should be updated accordingly as well, in order not to break the query functionality of the lists in BLT’s project space. —MisterSynergy (talk) 18:10, 19 April 2018 (UTC)
It is not efficient for me to learn to use a new tool. It is why I ask for it to be done. Thanks, GerardM (talk) 05:08, 20 April 2018 (UTC)
If nobody else volunteers, I can do this—hopefully this evening or otherwise tomorrow. —MisterSynergy (talk) 06:18, 20 April 2018 (UTC)
@GerardM: I just started the move. The lists at enwiki will be updated once all cases are moved. There are four lists which were broken already before the move was started, see the no items section in this Listeria status page. —MisterSynergy (talk) 10:45, 21 April 2018 (UTC)
@MisterSynergy: Thank you so much that you do this for me :). Let me know when you are done and, I will update the Listeria lists I know about. Please identify the catalogs you processed so that I can inform the people involved. Thanks, GerardM (talk) 10:58, 21 April 2018 (UTC)
As requested, this is only about the Black Lunch Table (Q28781198) “catalog”. Most or even all other cases were already moved a couple of days ago. I will ping you again once the move process has finished. —MisterSynergy (talk) 11:22, 21 April 2018 (UTC)

@GerardM: this is completed now. There are no items left to move, and I have also updated all lists in English Wikipedia (cf. en:Special:Contributions/MisterSynergy). To the best of my knowledge, none of the queries has a different result now, since I triggered updates for each one after the update. —MisterSynergy (talk) 14:23, 21 April 2018 (UTC)

More than 50% of my edits get reverted. And "Days of The Week" are separated into 2 items; undiscussed.

  1. Everything starts with one day I find that the Chinese, English, Deutsch version of the "Days of the Week" are not connected. 42 languages are in one group and 11 languages are in the other group, and I found no discussion about the split in Wikidata or the English Wiki, so I merged them. And then I got reverted by User:Andreasmperu. So I merged them with a message "No one single language separates this exact same topic. No separation discussion found in wikidata or english wiki of the topic 'Days of the week'", and leave a comment on the Talk:Q41825 "Separation will create unnecessary difficulties to compare difficult languages of the 'Days in the Week'.". And then I got reverted by the same admin again with just two words "wrong item" with no reply on the discussion page. [14] [15]
  2. As a result, more than 50% of my edits in Wikidata get reverted without a reason. If you do Google translate, you will know these two are the same topic "a ship destroyed and sank".
  3. G translate will tell you that these two(w:en:Chuanyue) are about the same topic of "Time Travel Fiction".
  4. Click in to look at the map and coordination and you know one is about the same town, and the other is not a town.
  5. You don't even need Google translate to know these two are the exact same group of Baptist Universities. Just spend 5 seconds to click in.

What an irony that an IP user is more willing to explain and communicate than an administrator. 118.143.147.130 22:01, 9 April 2018 (UTC)

In the first case, you seem to have conflated "Names of the days of the week" with "Day of the week". That's an error, since those are different concepts, and was rightly reverted. Similarly, not all time travel in fiction is a time travel novel. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:24, 9 April 2018 (UTC)
Per Andy, you in fact get 5/5 wrong.
  1. Two different concepts, names of days of the week versus day of the week.
  2. Disambiguation page versus article
  3. As noted, not all time travel in fiction is a time travel novel; different concepts.
  4. zh.wiki has articles attached to both items. You'd need to sort that out before they could be merged.
  5. Disambiguation page versus list article. You can argue that zh.wiki should change their article to a list, but until they do, it's a dab, and we need to reflect that.
So. Sorry about that. --Tagishsimon (talk) 22:54, 9 April 2018 (UTC)
  1. The English Wiki "Names of the days of the week" is actually the same article moved from "Days of the week". What topic of contents is different before the en-wiki article moved from "Days of the Week" and after moved to the current "w:Names of the days of the week"? They are still all about the exact same topic: how to count the seven days, their astronomy origins and their names in different languages. "No one single language Wiki separates this exact same topic into two articles" is the strongest evidence that none of the 53 languages thinks that they are about different topics.
  2. I don't think good idea to separate the exact same topic of different languages simply because one has less content. This will break the cross-reference and cross-improvement connections among different languages.
  3. Of course, fiction includes novels and films. It is wrong to say that "小說" in Chinese only includes written books. 小說 does include films and other forms of art, so 穿越小說 is same as "Time travel in fiction" and both of them include "Time travel novel". Google translate tells you the same 小說=Fiction Fiction=小說.
  4. I wonder whether you guys did click in for 5 seconds. w:zh:金字牌鎮 and w:sv:Jinzipai (köpinghuvudort i Kina, Anhui Sheng, lat 29,85, long 117,81) are the same location at (29.85 N, 117.81 E). w:zh:金字牌 is a "message certificate of royal orders", which is definitely not a location.
  5. Bad idea to separate the exact same topic of different languages; not to mention even their content are the same. So if I change the w:Baptist University to disambiguation, this one is done? 118.143.147.130 12:24, 10 April 2018 (UTC)
  1. (All IMO...) This one is something of a clusterfuck. The two wikidata items are distinct and for that reason should not be merged. However I concede that there are claims and wikilinks on day of the week (Q41825) that belong instead on names of the days of the week (Q42962347) or on another item (e.g. the GND ID seems to point to a concept of 'Working Days'). I concede, too, that not much may be left on day of the week (Q41825) by the time we finish.
  2. These remain two distinct items, one pointing to articles on concept of shipwreck/ing, the other a zh disambiguation page. zh's article on shipwrecks was pointed at the wikidata item maritime disaster (Q2620513) ... I've moved it to shipwrecking (Q906512). I think that's now sorted?
  3. So for this one ... chuanyue (Q10453828) is a distinct subclass of time travel in fiction (Q253732) and the two should not be merged. The zh. wikilink from chuanyue (Q10453828) looks appropriate; is about tt novels. The en. wikilink was to an article on Chuanyue which is a distinct subclass of chuanyue (Q10453828) ... I've created a new item, chuanyue (Q51733250) and moved the en. wikilink to it. I'm hoping we're good on this now.
  4. This one should be sorted now.
  5. Yes; I've done that. With luck, also sorted.
More generally, you'll not get much dispute that there is a lot wrong or that can be improved in many wikidata items. Exactly how to effect improvement can be a more thorny issue. --Tagishsimon (talk) 00:47, 11 April 2018 (UTC)
@Tagishsimon: I'm afraid that I have to bump WD:RFP/R#Andreasmperu again here. --Liuxinyu970226 (talk) 10:11, 11 April 2018 (UTC)
Actually it’s xiaoshuo (Q59126), which is more general than novel, but still a subclass of fiction.--Stevenliuyi (talk) 02:47, 22 April 2018 (UTC)

Taxonomic etymology - in search of inspiration (Q21090291)

This is similar to Synalpheus pinkfloydi sp. nov., a new pistol shrimp from the tropical eastern Pacific (Decapoda: Alpheidae) (Q29390847) (see above) and blocks were executed by User:Rschen7754 who did not help to resolve such matters. Multiple widely uses properties at statement level are removed here by Mr. Mabbett. We have more than 3,900 items for ZooKeys (Q219980), but only this one is missing some of this properties at statement level. Is there any reason why we should do that? There are more issues reintroduced with the latest revert, but I think we should do this step by step. Let's start... --Succu (talk) 19:57, 21 April 2018 (UTC)

If I had money for every time you decided to ping me and drag my name through the mud... --Rschen7754 20:08, 21 April 2018 (UTC)
Sorry, but this is not about you. This is about some unresolved modeling problems (including some longer lasting conficts). --Succu (talk) 20:19, 21 April 2018 (UTC)
Looks like you made it about you at WD:AN. --Succu (talk) 20:37, 21 April 2018 (UTC)

ISU figure skater ID (P2694) format

There are some format issues to sort out, please see Property talk:P2694.
--- Jura 06:13, 22 April 2018 (UTC)

ICTV taxonomy latest version (Q45362532)

Hi all,

I'm using the virus ICTV taxonomy of viruses to reference viruses on a website I'm working on and I was looking for a linked data format that would help me in keeping these taxonomic references up to date. I was first glad to find on Wikidata a contribution about the ICTV taxonomy (List of ICTV master species list on wikidata) since till now the ICTV website only provides an annual official release of the virus taxonomy in an Excel sheet format with about 1 thousand updates each year over the last releases (https://talk.ictvonline.org/files/master-species-lists/m/msl).

However, I'm interested in having fixed URIs that would have updated content when the taxonomy of these viral taxons are updated by the ICTV (basically updated labels and parent taxons, the previous label could be put in the "Also known as" section or in a more adapted statement ), but on Wikidata it seems that each release of the ICTV taxonomy exists with a kind of independance which is a kind of duplication of the information (cf: List of ICTV master species list on Wikidata): Even if common terms in the different version of this taxonomy are well linked to the corresponding ICTV versions there is an incoherence when the terms are updated since a new URI with the new term is created. For example according to the ICTV history tracking system on taxonomy we can find that the current taxon officialy named Avian avulavirus 7 was previously named Avian paramyxovirus 7 but there is 2 distinct URIs for these terms on Wikidata (Q29004365 and Q18907779). https://talk.ictvonline.org/taxonomy/p/taxonomy-history?taxnode_id=20161025


To me, and probably to most of the virologists, it is crutial to know what is the current ICTV taxonomy (ie:latest version) and keep track of the changes made so that we can know what is the current official name for a taxon (even if updated) and what is its current lineage (even if modified)... At the moment Wikidata is somehow providing a biased exhaustive view of all the existing releases of this ICTV taxonomy.

I'm new to Wikidata and I'm new to Linked data too... so I was wondering if we could ask for a kind of merging contribution named "Latest version of the ICTV taxonomy" that would have fixed URIs for each taxon including the next updates of the ICTV taxonomy without changing the URIs? ... a new URI would only be created for new taxons (not for updates in the taxon names or lineage) and removed taxon in the taxonomy would consequently be marqued as moved to their parent taxon.

@succu: ... Avian avulavirus 7 (Q29004365) and Avian paramyxovirus 7 (Q18907779) both your creations. --Tagishsimon (talk) 01:23, 11 April 2018 (UTC)
Paging @succu: ... would you mind explaining your creation of duplicates, please. --Tagishsimon (talk) 15:12, 13 April 2018 (UTC)
Virus nomenclature is different from that for plants or animals, virus names can be deleted or renamed. The ICTV Master Species List (Q45362532) (the latest is ICTV Master Species List 2017 v1 (Q51526638), not complete here) did not contain the former name. You have to connenct the former name manually via replaced synonym (for nom. nov.) (P694), but this connection is not very complete. --Succu (talk) 15:50, 13 April 2018 (UTC)
Thanks Succu. @D3fk dev: I think we aspire to the arrangement you suggest - a single stable item for each virus. Right now the way in which ICTV release information is not conducive to us being able to achieve that. Although the website will provide a change history for each virus, the spreadsheets do not provide specific information on the nature of a change made to virus attributes from year to year. Even the URL for the change history of a virus will change from year to year. ICTV seem to lack, or seem not to promulgate, a primary key value for each taxon; and reserve the right to change all attributes of the virus, including its name. Together, that's a data nightmare. It would take some masterful web-scraping to glean the information we need from the ICTV website, which is a shame, because ICTV clearly have the data that we (and others - you - require). I think someone needs to make enquiries with info@ictvonline.org to see if we can get a release of data that ties together the names of viruses in each of the yearly releases, so that we can start to merge items and provide, as properties and qualifiers, their attributes as they vary over time. --Tagishsimon (talk) 01:29, 14 April 2018 (UTC)
Thanks Succu and Thanks Tagishsimon @succu: @Tagishsimon: Happy to see that you aspire to fix this problem that way. I've been in contact recently with ICTV poeple through the info@ictvonline.org contact point about having access to possible other formats of the ICTV database (ideally Linked Data)... They are really open to create specific releases of their species spreadsheets (actually they are already providing their data in the form of a slightly modified spreadsheet to a number of taxonomic databases including Species 2000 "Catalogue of Life" and the World Registry of Marine Species). They've said that they were planning to find an alternate way to provide their data to the scientific community and the public... I've already suggested Wikidata as the best solution for this and they said they find number of advantages in using Wikidata (I didn't know at that time that there was already a contribution for the ICTV species list on Wikidata). Could you contact them about the ideal format (e.g.: including an ID for each taxon and the previous known names and parent taxon in case of update) for the succubot so that this consistency problem on Wikidata could be fixed in short terms? -- D3fk dev (talk) 9:51, 16 April 2018 (UTC)
@succu: @D3fk dev: I pinged info@ and Prof. Elliot J. Lefkowitz was kind enough to respond ... in short, providing alternate means of access to taxa history data is in their to-do list, but the list is long and the work is done in their free-time. (And they have data gaps in places - e.g. no info in the database as to which of the 1979 taxa were merged, in 1999, into either Human enterovirus C or Poliovirus, in https://talk.ictvonline.org/taxonomy/p/taxonomy-history?taxnode_id=20171985 ) I have not, at this time, asked them to do anything to alleviate our issue; I think we would need to have someone, or a team, ready to work with them before asking them to consider the effort required to provide a special cut of data for us. I'm now thinking that we might be able to make progress with a modicum of web scraping, but need some time to ponder on that. --Tagishsimon (talk) 20:51, 16 April 2018 (UTC)
@succu: @Tagishsimon: good news! However, don't you think that simply having access to a slightly modified version of their ICTV spreadsheet including 1 row per taxon and an additional column for the unique ICTV taxon ID + an additional column for the previous taxon name known + a latest column for the previous parent taxon known... should be sufficient for the succubot to make links within existing data on Wikidata and solve this issue? -- D3fk dev (talk) 7:43, 17 April 2018 (UTC)
There is certainly a tabular data representation to be had which conveys the information we need. Lefkowitz made the point that they had in the past sought to do exactly that, but ended up with a complex hard to understand spreadsheet which was difficult to maintain. I don't think your 2 -line spec covers all of the complexities of merge, split, rename, reparent; and it is because I don't have the headspace right now to think through the issues (because, IRL) that I chose not to persue a suggestion that the volunteers at ICTV redirect their effort to satisfy our needs. Equally, slow time, I'm still looking at their website with a view to scraping. But right now I have a sash-window to dismantle. Very exciting. --Tagishsimon (talk) 11:56, 17 April 2018 (UTC)
@Tagishsimon: Actually ICTV updates complexity could be resumed by :
  • New taxon(= a new ICTV unique ID in the spreadsheet = a new URI to create: solved by the ICTV unique ID column)
  • Renamed taxon (= same ICTV unique ID = same URI on wikidata with changed label and the previous label moved in a "previous/former name" property of the taxon type or in the "also known as": situation solved by the previous taxon name column, the current taxon name and the ICTV unique ID)
  • Moved taxon= reparent (= Same ICTV unique ID, same taxon name but different previous parent taxon name from the current parent taxon name = same URI but changed parent taxon property of the property of the taxon type statement)
  • Moved and renamed taxon (=Same ICTV unique ID, different taxon name from the previous taxon name and different parent taxon name from the current parent taxon name)
  • Deleted taxon (ICTV unique ID not anymore in the taxons list = URI unlinked from ICTV source and marked as deprecated)
"Spliting" might be resumed by the creation of new taxon or a new taxon+ a renamed taxon
"Merging" seems equivalent to a deletion or a deletion + a renamed/moved taxon
I agree that keeping all historical data in a single file would be a mess but if they continue to work with yearly distinct ICTV release files it should be ok.
The main interest of having a dedicated spreadsheets release for Wikidata would be the stability of the data format and the direct implication of ICTV in delivering their data to the scientific comunity and to the public through Wikidata; versus a web scraping only initiated by wikidata that could miss data if ICTV decides to change their web interface. That being said, I hope your sash-window was dismantled and fixed without problem -- D3fk dev (talk) 13:22, 17 April 2018 (UTC)
There is another difference: We accept different taxonomic viewpoints for plants, animals etc. and model them around taxon name (P225). But for virus taxonomy there is only one taxonomic viewpoint, that of ICTV. So your proposal change labels etc. is not working in the general context how WD handles taxonomic issues. But WD should be able to keep track that the species Bovine papilloma virus from the ICTV 1st Report (MSL #01) in 1971 evolved into Deltapapillomavirus 4 (Q18972422), Epsilonpapillomavirus 1 (Q18972592) and Xipapillomavirus 1 (Q18972997) today. A additional column for renamed taxa would help ofcourse. --Succu (talk) 15:34, 17 April 2018 (UTC)
@succu: Yes, you're right ICTV represents on the web only 'one' taxonomic viewpoint on the virus taxonomy information available through the web and ICTV with their yearly update system probably works somehow differently than others in the world of taxonomy ... but this ICTV point of view on virus taxonomy is recognized within the complete virologist community as the single Official taxonomic viewpoint on "virus taxonomy" and all the other taxonomic viewpoints existing on the web at the moment that are different from the ICTV point of view on the virus taxonomy are just late to catch up on the ICTV taxonomy(eg: Species 2000 "Catalogue of Life", the World Registry of Marine Species, or even the NCBI taxonomy that mentions "The NCBI taxonomy database is not an authoritative source for nomenclature or classification - please consult the relevant scientific literature for the most reliable information.").
So trying to retranscribe in Wikidata all existing virus taxonomies through the web at a moment is really confusing since it participates to maintain deprecated nomenclatures in the world of viruses nomenclature. It would be wonderfull if Linked data could make more sense in the world of taxonomy and not simply tending to be exhaustive.
Moreover, if URIs of the ICTV virus taxonomy could be fixed with updated labels on taxon updates by ICTV (as well as updated statements on parent branching and naming history), Wikidata could became a wonderfull tool for spreading an up to date virus taxonomy to the world.... Simply imagine how wonderful it would be to simply have to reference a WD taxon URI once and to get the latest known name and lineage for this taxon at each request! Wikidata would become for sure the main data provider for most of users of the virus taxonomy. To reach that goal the Wikidata bot used for virus taxonomy could be adapted to handle the virus taxonomy slightly differently from other kingdoms taxonomies... simply in order to make more sense within the data. You're right that WD should also be able to keep track of the evolution of the species/taxon names (history allows to catch up on the current taxonomy), but it could be done by recreating a new URI for the former species/taxon names at each ICTV taxon update. A subclass of taxon (Q16521) like "Former taxon" or "Historical taxon" could be imagined or even a subclass of taxon name (P225) like "Former taxon name" could be useful. At the moment on WD we can find several instances of a taxon for a current taxon and all the historical versions of this taxon but in reality it is question of the same taxon that was renamed but WD doesn't link this information for now -- D3fk dev (talk) 8:22, 19 April 2018 (UTC)
This will not work. E.g. what should we do with abolished (deleted) names. Delete them here too? Different taxonomic viewpoints are represented here by references (at least we try). A SPARQL query like this should give you all the accepted names according to ICTV Master Species List 2017 v1 (Q51526638) (no guarantee it does indeed). There is another problem: There are a lot of older Wikipedia articles never got updated to the newest "offical viewpoint". Even Wikispecies host track. --Succu (talk) 19:34, 20 April 2018 (UTC)
@succu: I suggest that abolished taxons (deleted names) should simply be considered as moved to their parent taxon (This is how it is handled in the ICTV spreadsheets)... therefore, a WD merging process should do the stuff (+ creation of a "former taxon" instance with the abolished taxon name for the label in order to maintain history).
I have no problem in having a "complete" list of the latest master species names entered in WD, tagged with ICTV latest version (a SPARQL request like this listing parent taxons would have the power to provide you with the lineage).... the main problem here is that if I'm using a WD URI to reference a taxon name on my website, this URI will be obsolete on the next ICTV update of the taxon name: so all entities in the world(websites, databases, URIs ... ) referencing that URI for the taxon name will display obsolete information since they are using an URI that return a deprecated taxon name that is even not anymore a taxon name (but a former taxon name)... this is not the way linked data should work ... URIs for the latest taxon name should always return the latest taxon name not a deprecated taxon name: Linked data have that power to create consistency with data.
So, the second problem you mentioned should be solved by this update labels approach: If the wikidata articles or even the wikispecies host track are using an URI to reference a taxon name, they will be updated in the same time that the label and the statements of that URI: e.g. for now, referencing Alethinophid 3 reptarenavirus (Q22106453) should return Rotterdam reptarenavirus since it was renamed to this taxon name in the latest ICTV release ICTV Master Species List 2017 v1 (Q51526638), but it is not returning this taxon name, it is not returning a currently existing taxon name and there is no way to have Q22106453 linked with the real data in WD at the moment. -- D3fk dev (talk) 16:00, 22 April 2018 (UTC)

Payment cards vs systems

Is there any valid way that payment card (Q1436963) can be linked to payment system (Q986008)? It's not a subclass, because a payment card isn't a type of payment system, just a thing used by such a system. I don't think part of (P361) or used by (P1535) can be used either, since the inverse statements don't apply (there are payment systems that don't use cards, like cash and barter). Ghouston (talk) 00:18, 20 April 2018 (UTC)

@Ghouston: A possible way: payment card (Q1436963)=> manifestation of (P1557) => payment system (Q986008).--Micru (talk) 18:20, 20 April 2018 (UTC)
I think it has the same problem as subclass, since a payment card isn't a payment system, just a part of one. facet of (P1269) doesn't have the inverse constraint like part of (P361), but would also imply to me that every payment system would have it. Ghouston (talk) 23:05, 20 April 2018 (UTC)

part of (P361) seems problematic in general, like on motor car (Q1420) which lists various parts, some of which an automobile doesn't actually need to have. E.g., an electric automobile doesn't have a fuel tank (Q1411232). Ghouston (talk) 23:18, 20 April 2018 (UTC)

I think that one can be fixed with motor car (Q1420) has part(s) of the class (P2670) automotive part (Q46765731). Ghouston (talk) 00:06, 21 April 2018 (UTC)
Except then things like wheel (Q446) aren't automotive part (Q46765731) as such...this is too hard. Ghouston (talk) 00:10, 21 April 2018 (UTC)

linking between corresponding template and category

How do I connect on wikidata bewteen a cateroy and template which deal with same thing? My example is Template:Roshei Yeshiva of Yeshivat Har Etzion (Q22741153) and Q32819726 which both deal with the rosh yeshiva (Q2308372) (head rabbi/dean) of Yeshivat Har Etzion (Q2910398). Thanks, DGtal (talk) 06:57, 22 April 2018 (UTC)

Property description error - link to distinct values not working, there is only unique value

Property_talk:P213 has a link to Wikidata:Database reports/Constraint violations/P213#distinct values, but that target section does not exist. There is a section Wikidata:Database_reports/Constraint_violations/P213#"Unique_value"_violations. 78.48.189.69 09:53, 22 April 2018 (UTC)

Phoenician language

I would like to add a statement that the name of Gades (Q3094251) was 'Gadir' in Phoenician (Q36734), but 'Phoenician' isn't available in the language drop-down. Phoenician has ISO 639-2 code 'phn'. What needs to happen to make Phoenician a valid language for <name> statements? - PKM (talk) 21:54, 15 April 2018 (UTC)

@PKM:
Tracked in Phabricator
Task T155425
You should be subscribed to the ticket now. Mahir256 (talk) 23:18, 15 April 2018 (UTC)
@PKM: but please don't forget that Phoenician (Q36734) had no such letters as "G", "a", "d", "i" or "r". Infovarius (talk) 16:51, 18 April 2018 (UTC)
It's hard to reconcile that with statements like "Phoenician gadir ‘masonry wall enclosing a town’." (Oxford Concise Dictionary of Place-Names, under 'Agadir (Morrocco)'). Perhaps different romanizations? I have a handful of sources that call the Phoebnician city 'Gadir'. If you have sources that indicate a different name, it would be great to add both to the item. - PKM (talk) 18:37, 18 April 2018 (UTC)
I agree that Gadir is a valid Phoenician word. It can be rendered in different scripts and above is rendered in Latin (romanization). If you consult the IANA lang tag list, you'll see the default script for this language is "Phnx" Phoenician, therefore the language should be called "Phoenician (Latin transcription)" and the lang tag should be as in "gadir"@phn-Latn --Vladimir Alexiev (talk) 13:15, 19 April 2018 (UTC)
Looks like langcom overlooked that. Wonder what they are actually checking/discussing.
--- Jura 13:31, 19 April 2018 (UTC)
@Vladimir Alexiev:, thanks for updating the Phabricator ticket. - PKM (talk) 21:52, 22 April 2018 (UTC)

Allowing serial number for award

Currently serial number (P2598) is not allowed qualifier for award received (P166). At least in Soviet Union orders had serial numbers, so will be good idea to allow such use case. --EugeneZelenko (talk) 02:54, 22 April 2018 (UTC)

That sounds more like catalog code (P528). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:02, 22 April 2018 (UTC)
It's not catalog code. Each instance of order had serial number. --EugeneZelenko (talk) 23:24, 22 April 2018 (UTC)

Colors and pigments

Just confirming - are we agreed that we need separate items for colors and pigments? (See rose madder (Q2609895) where the linked articles seem to be a mix of info about the pigment, the color, or both). It seems to me they should be separate, linked by color (P462). - PKM (talk) 21:25, 20 April 2018 (UTC)

+1 for separated items. And pigment should be named using the name of the substance/chemical compound. Snipre (talk) 22:16, 20 April 2018 (UTC)
Thanks for the +1. Are you saying your preferred label for Prussian blue (Q421894) would be 'Ferric ferrocyanide'? I would disagree with that. I prefer that standard artists' pigment names as used in Getty AAT, with the chemical compound names as aliases. But then that's the domain I am used to working with.- PKM (talk) 22:50, 20 April 2018 (UTC)
@PKM: If there is a risk of confusion between the color and the pigment, yes, the best way to avoid people doing a mess is to use the chemical name of the pigment instead of the rendering color. The color name can always be added as alias. Snipre (talk) 12:39, 23 April 2018 (UTC)

Software blocks fixing - Human item - Latvian name in label copied to several other language-specific labels

I tried to fix, but software blocks me: 

Could not save due to an error.
The save has failed.
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.

Fixed some on each of the following items, but not all:

Quick statements down

Clicking on the quick statements link only gives the very cryptic message "500 - Internal Server Error". I can't find any other info about this. Is there any indication of what the problem is or how long this will take? Normally I would just contact User:Magnus_Manske but I noticed that this link works and only this link is broken. Thx. Jane023 (talk) 05:57, 28 April 2018 (UTC)

It looks like anything under /wikidata-todo results in an HTTP 500 error code, so perhaps the webservice for that tool simply isn’t running at all? (That would also explain why ~tools.wikidata-todo/error.log was last written to on April 13.) In that case one of the tool maintainers would only have to run webservice start – but unfortunately Magnus Manske is the only maintainer according to toolsadmin. --Lucas Werkmeister (talk) 15:40, 28 April 2018 (UTC)
@Jane023: he fixed it :) --Lucas Werkmeister (talk) 19:17, 28 April 2018 (UTC)
Yay! Jane023 (talk) 21:55, 28 April 2018 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 08:26, 29 April 2018 (UTC)

Display of a thumbnail for image properties

Tracked in Phabricator
Task T46727
Resolved
Tracked in Phabricator
Task T192601
Resolved
Tracked in Phabricator
Task T192869

Hello all,

Since yesterday, the statements including a Commons media file are now displaying a thumbnail of the picture in Wikidata. Clicking on the name of the file opens Mediaviewer. (try it on your favorite item)

Note that you may need to purge the page to see this change appears.

It's possible that you encounter a bug, in that case, feel free to let a comment and a screenshot on that ticket.

Thanks, Lea Lacroix (WMDE) (talk) 05:27, 20 April 2018 (UTC)

@Lea Lacroix (WMDE): both logged in and logged out I don't see any change. I don't see a thumbnail and I don't get a link to the MediaViewer. Is the change live at the moment? Can you provide a screenshot here how it is supposed to look? Multichill (talk) 09:36, 20 April 2018 (UTC)
Hello Multichill and all,
Because of a problem during the deployment train, the feature is currently removed. Sorry for the inconvenience, I'll let you know as soon as it is back :) Lea Lacroix (WMDE) (talk) 10:44, 20 April 2018 (UTC)
@Lea Lacroix (WMDE): now I get the image, but the link is to the local File page instead of directly to Commons. That seems to be a regression. Multichill (talk) 19:20, 23 April 2018 (UTC)
Thanks for noticing. I tracked this on Phabricator. Lea Lacroix (WMDE) (talk) 07:51, 24 April 2018 (UTC)

Quick community survey: notability of family members due to “structural need”

There are occasionally items about family members of notable persons (parents, children, siblings, and so on) listed at Wikidata:Requests for deletions (there are some cases listed these days). When they are linked from the item of the notable person, it is often argued that these family members are also notable due to “structural need”. However, I am not sure whether such relations really create “structural need”. One could easily extend this approach and consider the relatives of the family members also as notable, and very soon each and every human being would be “notable” for Wikidata. We simply don’t have the capacities to manage this appropriately; I have also serious concerns about privacy issues, particularly (but not only) if items about minors are created.

The ongoing Wikidata:Requests for comment/Privacy and Living People as well as Wikidata:Living persons (draft) and Wikidata:Living people do not cover this problem appropriately, to my opinion. What does the community think about the problem? In which situations does the “structural need” criterion justify to have family member items? In which cases not? What’s the use case to equip family members with individual items when their only (notability-relevant) achievement is to be the relevant of someone else? Which minimum sourcing requirements should we have for those cases? What else do you have to add to this survey? I am open for any kind of input here, thus feel free to express your opinion. Thanks, MisterSynergy (talk) 10:56, 20 April 2018 (UTC)

The RFC and living persons policy proposal are not concerned with notability, that's a separate issue. I think if the family members would be indicated in any way on a language wikipedia article about the "notable" person, then we should have (limited) items for them in wikidata. Of course the information on them should come from reliable sources, as covered in the RFC etc. ArthurPSmith (talk) 12:22, 20 April 2018 (UTC)
Part of the problem is that we do not define "structural need". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:52, 20 April 2018 (UTC)
The structural need for family members is particularly to establish and show relations.. So fathers, mothers link their children to grand parents.. Genealogy is one structural need. Another is for people who won an award or who held a position. The need is in making such lists complete. Thanks, GerardM (talk) 13:54, 20 April 2018 (UTC)
The structural need is that if you want to name somebody's spouse or child you need to create an item for them. The questions are whether their spouse and child really need to be named, and if so, how many additional statements can be added to these spouse/child items. Potentially such additional statements could require the creation of further items for the same structural need, and so on until everybody is on Wikidata. Ghouston (talk) 02:44, 21 April 2018 (UTC)
  • There are a lot of historical dynasties in art, politics, local land ownership etc that it is useful to be able to retrieve precisely. If we have an item for the grandparent and an item for the grandchild, then I think in most cases it would be useful to have an item for the parent in between. In very many cases the relevant people will probably not be living; but even when they are living, the connection is very likely to be public and generally known, through standard reference sources such as (in the UK) Who's Who, Burke's Peerage, Debrett's Distinguished People of Today, newspaper profiles, etc. When we have such a source for the information, that is readily publicly accessible, then I think the item would be appropriate to create. For more distant relationships relative (P1038) with kinship to subject (P1039) is available. Jheald (talk) 14:33, 20 April 2018 (UTC)
The links also make possible amusing queries such as current members of the House of Commons descended from possibly mythical ancestors. Jheald (talk) 14:45, 20 April 2018 (UTC)
  • For deceased persons, I think it is helpful to create items for those between people and their grandparents.
    --- Jura 06:29, 21 April 2018 (UTC)
  • Keep it limited to notable persons. If parents is not notable then relative (P1038) should be used for grandparents and other relationships with qualificator kinship to subject (P1039). Breg Pmt (talk) 09:01, 21 April 2018 (UTC)
  • I appreciate the concerns where the newly-created items are themselves living people, and a decision on that might be useful, but I would strongly urge that we don't make a blanket policy against creating these linking items for non-living people. This sort of thing is massively useful for building relationship links between historical figures, as Jura & Jheald say, and IME the relative (P1038) approach for "skipping" people doesn't work well compared to a standard family relationship.
    One approach might be to firm up "structural need" to say that ideally you need the item to be linked from *two* independently notable items, or at least be part of a chain that connects two items. So we would allow, for example, linking someone to their great-grandmother by creating the father and grandmother items, but discourage creating items for all the cousins etc who aren't part of the direct chain. Andrew Gray (talk) 09:38, 21 April 2018 (UTC)
  • For me, as long as you have at least one source, there is no problem to create *all* known familly numbers and it's often much easier and proper to create these items than to use overrefined and convoluted statements. The only exception is maybe living people, a different question that need a different approach. Cdlt, VIGNERON (talk) 08:39, 24 April 2018 (UTC)
  • I don't think Wikidata need to restrain the rules for dead people. I'm more worried by the property nombre d'enfants number of children (P1971). There is not a lot of datas on that subject on the Internet except genealogical websites. But but babies and child dead very young are rarely mentioned. When someone change the value of number of children (P1971), I can't fact-check quickly if it's a vandalism or not. Pyb (talk) 09:45, 24 April 2018 (UTC)

Will "applies to part" work well with conversion factors?

year (Q577) has recently been edited to add several conversion to standard unit (P2442), each with a property applies to part (P518), for different kinds of years, such as common year and leap year. Will the various bits of software that automatically interpret conversion factors be able to take into account the applies to part (P518) or will the software become confused and only see several contradictory conversion factors? Jc3s5h (talk)

On further consideration I have removed the conversion factors because the item contains different from (P1889) for common year and leap year. Jc3s5h (talk) 13:56, 23 April 2018 (UTC)

Are "different from" and "subclass of" incompatible?

year (Q577) asserts it is different from (P1889) calendar year (Q3186692). But "Calendar year" says it is subclass of (P279) of year (Q577). Are these statements incompatible? If we decide a year is not different from a calendar year, they we will have to increase the tolerance on the year-to-seconds conversion factor so the tolerance is about a month, since many calendars have "leap months" to keep them roughly in sync with the year. Jc3s5h (talk) 16:27, 23 April 2018 (UTC)

I'm not sure that "different from" is very useful in this context, since subclasses are always different from their parent classes, otherwise they should be merged. Subclasses of year (Q577) seem to be done on the principle that year (Q577) is any time period roughly based on the period of the Earth's orbit around the Sun; if year (Q577) is given a precise definition, it doesn't work. Ghouston (talk) 23:36, 23 April 2018 (UTC)
When we've got year (Q577) duration (P2047) 31,558,000±1,000 second, it really doesn't make sense that calendar year (Q3186692) is a subclass, since calendar year (Q3186692) can be a quite different period of time depending on the calendar. Ghouston (talk) 23:40, 23 April 2018 (UTC)
Likewise, 1966 (Q2649) instance of (P31) year (Q577) doesn't make sense, because 1966 (Q2649) is an instance of a calendar year, specifically a year in the Gregorian calendar (Q12138), and wasn't a period of 31,558,000±1,000 seconds. As a non-leap year, before leap seconds were invented, it was 31,536,000 seconds. Ghouston (talk) 23:49, 23 April 2018 (UTC)
@Ghouston, Jc3s5h: I think 1966 (Q2649) and 2018 (Q25291) should not be an instance of year (Q577), but an instance of calendar year (Q3186692), because 2018 (Q25291) is not an instance of an instance of unit of time (Q1790144), but an instance of time interval (Q186081) or point in time (Q186408). Maybe we need to remove all "instance of (P31) year (Q577)" statements from each specific year items, and make them instances of leap year (Q19828) or common year (Q235729), which are subclasses of calendar year (Q3186692). Also, we have to remove "subclass of (P279) year (Q577)" from calendar year (Q3186692). --Okkn (talk) 10:17, 24 April 2018 (UTC)

Property needed - Member of a scientific society

  1. honorary member (Q10519151) - honorary member - Ehrenmitglied
  2. corresponding member (Q2625729) - corresponding member - Korrespondierendes Mitglied - en:Category:Corresponding members of academies of sciences

77.180.35.91 18:34, 23 April 2018 (UTC)

What's wrong with member of (P463) and subject has role (P2868) as qualifier? Sjoerd de Bruin (talk) 20:18, 23 April 2018 (UTC)
User:Sjoerddebruin - maybe ok. More complex in queries. 92.228.155.106 00:03, 24 April 2018 (UTC)
Complexity of queries is always case dependent. If you want to create a list of all honorary and corresponding members of an scientific society, the query is less complex if they share the same property. --Pasleim (talk) 07:03, 24 April 2018 (UTC)

Properties "other ontologies intance/subclass of"

{{Ping project|Informatics}} What do you think about adding properties for item superclasses as defined by other ontologies (e.g. philosophy of space and time (Q1754940) → Dewey subclass of: metaphysics (Q35277))? Reproduce ontologies inside Wikidata would allow to easily view how a particular item is classified in the various ontologies and compare it to how it's classified in Wikidata.--Malore (talk) 00:51, 24 April 2018 (UTC)

WikiProject Ontology has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. Pinging correct project. Dhx1 (talk) 01:00, 24 April 2018 (UTC)

@Malore: This already exists - use broader concept (P4900). ArthurPSmith (talk) 12:16, 24 April 2018 (UTC)

Wikidata weekly summary #309

Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • Track the number of constraints by type (phab:T191891)
    • Include thumbnails in image statements (phab:T46727)
    • Install Query Service on Factgrid (phab:T189863)
    • Page creation and edit rates (phab:T184948)
    • Implement unit constraint (phab:T164372)
    • Implement no bounds constraint (phab:T170610)
    • Started working on the constraint violation type to restrict to indicate that certain properties can only be used on items or lexemes (phab:T164744)
    • Creating a Docker container for QuickStatements (phab:T192079)
    • Removed hardcoded demo data in Wikibase Lexeme (phab:T188899)
    • New data type for linking to Forms (phab:T165579)

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.


Unmerge 2017 in Quebec (Q21027564)

Hi, can someone please unmerge 2017 in Quebec (Q21027564)? It's a mixed up item about two different things. I guess here is the problem --> [[16]]. I don't dare to do it myself, it could end up even worse. Thanks! Wikidelo (talk) 21:52, 30 April 2018 (UTC)

done --Tagishsimon (talk) 23:20, 30 April 2018 (UTC)
This section was archived on a request by: Tagishsimon (talk) 23:23, 30 April 2018 (UTC)

Changing the Wikipathways Identifier (P2410) from data type String to type External Identifier

Currently, the Wikipathways ID is of type String. It is, however, an external identifier. It apparently can be changed accordingly but requires community consensus. Can we reach consensus here? Who supports, who objects and why?

--Andrawaag (talk) 11:02, 24 April 2018 (UTC)

BEIC

I believe all uses of BEIC here are incorrectly done. Instead of using described by source (P1343) in this way, shouldn't there by a property for "BEIC ID"? --EncycloPetey (talk) 23:40, 24 April 2018 (UTC)

@EncycloPetey: You are welcome to propose such a property. Mahir256 (talk) 00:46, 25 April 2018 (UTC)
Thanks for responding, but you didn't answer my question. --EncycloPetey (talk) 01:42, 25 April 2018 (UTC)
If you want my opinion, @EncycloPetey:, I am uncertain whether "Aristophanes -444--385" or "Aristophanes" count as identifiers for use within BEIC, especially since the parameter for which either of these may be given within the URL is noted to be a search for 'free text'. If you want others' opinions, you are welcome to propose the property and see what others think of it. Mahir256 (talk) 04:23, 25 April 2018 (UTC)

P969 (P969) vs. located on street (P669) for japanese address

I'm confused about the using of P969 (P969) and located on street (P669). I would like to add the street address of the headquarters location headquarters location (P159) for japanese companys. For instance: 東京都杉並区阿佐谷北1－46－4 友和ビル5階

On the Property Talk Page of P969 you can read: "Use located on street (P669) and street number (P670) for multilinguage use" BUT on the Property Talk Page of P669 you can read: "In such cases [japanese addresses] you use located at street address (P969)" Now I don't know which is the right way. Any recommendation? Diggr (talk) 13:52, 18 April 2018 (UTC)

@okkn:^^--Liuxinyu970226 (talk) 14:18, 18 April 2018 (UTC)
@Diggr, Liuxinyu970226: With a very few exceptions, street names are seldom used in postal addresses in Japan. That is because most of the Japanese streets do not have their name. (see w:en:Japanese addressing system) So we can't use the pair of located on street (P669) and house number (P670) to express the address of locations in Japan. As you may know, the value type of P969 (P969) is not item but string, so the values depend on Latin script (Q8229), Japanese writing system (Q190502), etc. I can strongly feel the importance of new system or model to represent Japanese addresses that is independent of languages. --Okkn (talk) 15:46, 18 April 2018 (UTC)
@Liuxinyu970226, okkn: Okay. If I understand you right, I should use P969 (P969) for now, but it's not the elegant way. It would be better to have a new model. As an example, I edit the japanese company Madhouse (Q650867) and added P969 (P969) and postal code (P281) (the use of postal code seems unproblematic to me). Are there any objections to that? I think it is really necessary that it is possible to represent addresses (with local characteristics) correctly in wikidata. Diggr (talk) 08:36, 25 April 2018 (UTC)
If you are asking my consideration on both property, I'm afraid that I personally don't like P969 (P969), because that isn't a monolingual text property, and such contexts on zhwiki are mostly "separated" by languages, but the de facto way seems okay for me. --Liuxinyu970226 (talk) 09:07, 25 April 2018 (UTC)

So consider converting P969 (P969) to a new monolingual text property, please?!

As mentioned this noon, I would request to have text-only addresses in more than one languages. --Liuxinyu970226 (talk) 13:13, 25 April 2018 (UTC)

Trailing space problem

Tracked in Phabricator
Task T193214

Context of this question is values for P2580 (P2580). For whatever reason, BBL *require* a trailing space on some of their URL fragments, so:

How do we get a trailing space into the BBL value in Olga Beggrow-Hartmann (Q36538626)? The user interface will not allow a plain trailing space. And using %20 in the P2380 value results in an output of %2520. --Tagishsimon (talk) 21:13, 23 April 2018 (UTC)

Property_talk:P2580#Trailing_space_in_IDs 77.180.110.58 14:09, 25 April 2018 (UTC)

Mass corrections to format of property

The URLs of an external site changed their formatting and now all property values https://www.wikidata.org/wiki/Property:P3064 for Lepindex need to have the taxonid without a trailing ".0" - is there a way someone could automatically fix this change? Shyamal (talk) 02:51, 24 April 2018 (UTC)

Here is change is quite simple, it's easy to fix (someone with a bot or anyone with QuickStatements). On the talk page of the property (Property talk:P3064), Succu said he can do it with his bot. Cdlt, VIGNERON (talk) 06:42, 25 April 2018 (UTC)
See https://en.wikipedia.org/wiki/Wikipedia_talk:WikiProject_Lepidoptera#Update Shyamal (talk) 11:37, 25 April 2018 (UTC)

Q668 popular items

Greetings,
Is it possible to get a list of 20 or 50 popular items related to India (Q668)  View with Reasonator View with SQID , similar to Wikidata popular items, using SPARQL query or something else? It would be helpful for WD:IND, and the upcoming newsletter. --Titodutta (talk) 14:30, 25 April 2018 (UTC)

@Titodutta: Related how? country (P17)? country of citizenship (P27)? Some other property? In any case, you can adapt the SQL query present in this page. Mahir256 (talk) 15:22, 25 April 2018 (UTC)
Both of these are fine User:Mahir256. I am a looking for a list of 20 most popular items (anyhow directly) related to India (Q668). --Titodutta (talk) 15:33, 25 April 2018 (UTC)
@Titodutta: If count of sitelinks is used as a proxy for popular, then this report may help:
SELECT ?item ?itemLabel ?sitelinks WHERE {
  {?item wdt:P17 wd:Q668.}               #country is India
  UNION                                  #and/or
  {?item wdt:P27 wd:Q668.}               #country of citizenship is India
  ?item wikibase:sitelinks ?sitelinks .  #number of sitelinks
  filter(?sitelinks >20)                 #discard if less than 21
  SERVICE wikibase:label { bd:serviceParam wikibase:language 'en' }
} order by desc(?sitelinks)
Try it!
- hth --Tagishsimon (talk) 15:39, 25 April 2018 (UTC)
And then this one shows P31 values, but as a result gets lots of duplicate rows. --Tagishsimon (talk) 15:43, 25 April 2018 (UTC)
@Tagishsimon: Part of the appeal of the popular items page is that this is tracked based on the number of edits to the items, which neither of the queries above really addresses. Mahir256 (talk) 16:06, 25 April 2018 (UTC)
Good stuff. How dow e go about adapting that for Titodutta's purpose. --Tagishsimon (talk) 17:42, 25 April 2018 (UTC)

Wikidata:Requests for comment/Sort identifier statements on items that are instances of human

Wikidata:Requests for comment/Sort identifier statements on items that are instances of human 77.180.110.58 17:49, 25 April 2018 (UTC)

Inherited values

Some values (like "credit card length" or all has characteristic (P1552) values) are inerithed for all instances and subclasses of the item. Is there some way to:

  • indicate it?
  • prevent instances and subclasses of that item to be filled with different values (or at least point out when this occurs)?
  • view all values that are inherited from superclasses?

--Malore (talk) 17:00, 19 April 2018 (UTC)

Isn't it automatically the case that properties are inherited by instances or subclasses? If an instance or subclass doesn't match for a particular property, perhaps it shouldn't have been made an instance or subclass, or perhaps that property shouldn't have been on the class. Ghouston (talk) 23:40, 19 April 2018 (UTC)
But what you are actually saying, yeah it would be nice if you could see an item with all of the properties, including those inherited, and get a warning like a constraint violation if there were conflicting values. Ghouston (talk) 23:46, 19 April 2018 (UTC)
@Ghouston: It's not automatic. For example, all identifiers are not inherited, images are not inherited, "has part" is not inherited, "inception" is not inherited (also if children items shouldn't have an inception date previous to the parents one), and so on.--Malore (talk) 00:53, 20 April 2018 (UTC)
Hmm, yes changing properties makes sense in these cases, because the instance or subclass has more specific values that can be used. It would be nice if the superclass values were also valid, e.g., the image on human (Q5) was one that could be applicable to any human, but that won't necessarily be the case. Ghouston (talk) 01:02, 20 April 2018 (UTC)
@Ghouston:Maybe I used the wrong examples, but there are values that are inherited (all "has quality" values), values that are inherited but can be further specified ("inception", "images") and values that are not inherited (identifiers, "has part", "named after") because they refer only to that specific word or concept.--Malore (talk) 10:51, 20 April 2018 (UTC)
Formally claims in one item do not apply for anything outside of that item, thus a formal or even automatic “inheritance” is not possible—although it implicitly often makes sense to assume one. There are some exceptions: we have, for example, very few instances of transitive Wikidata property (Q18647515) (see here), where some kind of transitivity over property paths of those listed properties is implied. I am not sure whether there is anything else, but at this point I wouldn’t expect much more. —MisterSynergy (talk) 11:06, 20 April 2018 (UTC)
@MisterSynergy:Is there a reason for this? Because I think this feature would allow to save a lot of time and to easier spot wrong values and properties.--Malore (talk) 15:40, 20 April 2018 (UTC)
I was not involved when this was set up and I do not really know the reasons, thus I have to speculate. There are probably way too many exceptions within any set of items that instantiate a particular class item, so that automatic inheritances would go wrong way too often. With transitive subclass relations (as they are in place right now), things would even be more complicated. The world it quite complex and diverse, and so is the representation of it that we build here …
However, with the Wikidata Query Service it should be possible to investigate and compare all kinds of data sets, also particularly for wrong values/outliers/special cases and so on. Most users have to learn SPARQL first, unfortunately. —MisterSynergy (talk) 16:17, 20 April 2018 (UTC)
So it's just the opposite to what I guessed, but it makes it easier. mw:Wikibase/DataModel mentions inheritance but doesn't seem to say anything about this. Ghouston (talk) 11:06, 21 April 2018 (UTC)
@MisterSynergy: Value inheritance can be assigned to properties (everywhere that property is used, its value is inherited to all item children) or to specific items (the value of some given property of this single item is inherited to his children). Value inheritance for properties is definitely more prone to errors and exceptions, but for single items the exceptions should be a lot less and the fact that the value is visible in many items instead of only one should make easier to spot mistakes. However, it occurred to me that a reason such a feature is not available can be that it is too resource intensive because a single edit can cause thousand of edits.--Malore (talk) 10:56, 23 April 2018 (UTC)
I don’t think the resources matter here. If you change the label of a highly used item, e.g. about a country, there are thousands of items and connected pages which need to be touched as well.
Which advantage would an explicitly stated inheritance have? You can equally well write down the values everywhere, or read them quite easily from the class item. I’m not sure whether I really get the point why you need this at all. —MisterSynergy (talk) 14:06, 23 April 2018 (UTC)

@MisterSynergy:IMO, the advantages are two:

  • don't have to manually repeat the value for every instances/subclasses, especially if there are thousands of them. Probably it could be done with QuickStatements or using the API but it is way more complicate;
  • it is easier to spot inconsistencies if an instance/subclass has a different value than the parent item while they should have the same value: it can be determined on a case by case basis if the children represent an exception or the parent item has a wrong value.

--Malore (talk) 22:45, 23 April 2018 (UTC)

@Ghouston: If the value is not true for subclasses it can be easily deleted in the same way that it was created. However, maybe the best approach is to simply visualize a warning when these values are not respected instead of copying the values for all children. I think there are three ways to implement it:
  • through a "has value" property similar to "has quality" but for values instead of properties;
  • through an "has quality" qualifier like "value" or "preferred value" (if children items can have different values);
  • through something like the ranking icon.
Personally, I prefer the second option. What do you think?--Malore (talk) 11:24, 24 April 2018 (UTC)
I found out something similar already exists: properties for this type (P1963)--Malore (talk) 15:59, 26 April 2018 (UTC)

Single value constraint and deprecated statement

The item Yuncheng (Q73073) is flagged in the "single value" constraints violation of China administrative division code (P442) because the value changed in year 2000. However, the item seems correctly represented to me, with "start time" and "end time" properties, as well as using "deprecated" rank for the previous value.

Is there a way to tell the constraint violation checks to ignore past values? I feel like deprecated ranks should be ignored for these checks, or maybe a better management of start time (P580) and end time (P582) is needed, which would allow to check for overlapping periods as well?

Koxinga (talk) 05:56, 23 April 2018 (UTC)

  • If was a correct value in the past, it should not be using deprecated rank. Normal rank is just fine, and preferred rank for the currently valid value.
  • While it is using deprecated rank at the moment, the covi gadget does not report a violation. KrBot's list does, but the software generating these two reports is completely different.
  • There was a related discussion at Wikidata talk:WikiProject property constraints#Count best-rank statements in “single value” constraint recently. Very likely there will be a new “single best-rank value” constraint type soon that allows to have multiple values, but only one with best rank. The introduction of such a constraint is motivated exactly by the situation you are experiencing. I suggest to wait for it; maybe @Lucas Werkmeister (WMDE) can tell when it’s about to arrive here … —MisterSynergy (talk) 06:29, 23 April 2018 (UTC)
@Koxinga, MisterSynergy: This seems like yet another use case for separator (P4155) on China administrative division code (P442)’s single-value constraint. Mahir256 (talk) 13:08, 23 April 2018 (UTC)
I thought about it, but what would be the qualifier to use as separator? end time (P582)? Koxinga (talk) 15:50, 23 April 2018 (UTC)
@MisterSynergy: single-best-value constraint (Q52060874) should be deployed now, if I’m not mistaken, but I haven’t tested it on Wikidata yet. --Lucas Werkmeister (WMDE) (talk) 13:40, 26 April 2018 (UTC)

New badge available: good list

Tracked in Phabricator
Task T190976
Resolved

Hello all,

Some Wikipedias have a special content feature for good lists. This is now available as a badge in Wikidata: when editing a sitelink, you can select "good list" as a badge. You can see the list of all the badges here. Lea Lacroix (WMDE) (talk) 09:52, 25 April 2018 (UTC)

Variable results for unknown time values

For different date values with precision=9 (year), the value for month and day is variable. For instance, date of death (P570) for Yves Goldenberg (Q12743958) has the month and day as 0, while for Gheorghe Petrașcu (Q3595783) the month and day are 1. Why does this happen?--Strainu (talk) 21:34, 25 April 2018 (UTC)

You will find some information in archive: Wikidata:Project chat/Archive/2017/06#Merge tool doesn't merge dates as expected. It depends on the tool that added the statement. Matěj Suchánek (talk) 15:09, 26 April 2018 (UTC)

Item about a human that exists since 2013 and has external ID - up for deletion

"Abe Goldfarb" (Q4666390) - Wikidata is not English Wikipedia. In IMDb it says "Abe Goldfarb is an actor and writer, known for When in Rome (2010), The Horror at Gallery Kay (2018) and Neurotica (2017)." When one wants to write about these, then this item is useful. 78.51.216.7 22:28, 25 April 2018 (UTC)

This should be argued in the deletion discussion, not here. --Tagishsimon (talk) 22:39, 25 April 2018 (UTC)
User:Tagishsimon this is very obscure. A page with Q-numbers as section names, people nominate something, claim "not notable" and after it is done, noone knows what happened. 78.55.205.137 12:49, 26 April 2018 (UTC)
The lack of history after a deletion is a known annoyance. The individual's en.wiki article was deleted as non-notable ... wikidata often follows, where the wikidata item has created as a result of the en.wiki article. I think you'll just have to let this one go. --Tagishsimon (talk) 13:39, 26 April 2018 (UTC)

Bot edit war

https://www.wikidata.org/w/index.php?title=Q6678817&offset=20180423100050&action=history 78.55.205.137 12:47, 26 April 2018 (UTC)

Hopefully it will be fixed by this, but, @Ivan A. Krestinin, Magnus Manske: the owners of User:KrBot and User:QuickStatementsBot will consider if they need to rein in their bots based on that history. --Tagishsimon (talk) 15:00, 26 April 2018 (UTC)

Family tree

Hey, would it be possible to generate a family tree of a person using Wikidata. That would be an interesting visualisation. Capankajsmilyo (talk) 10:33, 24 April 2018 (UTC)

Something like the Ancestors Tool? Or the other way round? --Beat Estermann (talk) 10:46, 24 April 2018 (UTC)
There is also GeneaWikid (unfortunately very cluttered). But already linked Ancestors Tools is the best we have (although old and unmaintained), you can enable it in Gadget panel in preferences.--Jklamo (talk) 13:05, 24 April 2018 (UTC)
And Relator - Salgo60 (talk) 05:14, 27 April 2018 (UTC)

subclass of event

Is it OK for Powderham Castle (Q2106892) or The second High School of Mikołaj Kopernik in Cieszyn (Q9296000) to be an indirect subclass of occurrence (Q1190554)? Mateusz Konieczny (talk) 08:53, 26 April 2018 (UTC)

@Mateusz Konieczny: Rather doubtful. I'm not seeing a recent edit on the first of these ... can you explain more about the indirect subclass business ... where does it show up? --15:02, 26 April 2018 (UTC)
Not seeing this.
I've added {{Item documentation}} to the talk pages of historic house museum (Q2087181), English country house (Q1343246), and castle (Q23413), which are the direct classes that Powderham Castle (Q2106892) is in. The templates should show the full parent class tree for each of those classes, but I'm not seeing occurrence (Q1190554) included anywhere. Perhaps somebody has fixed this? Jheald (talk) 07:19, 27 April 2018 (UTC)

Planes and flights

Hi there! I'd like us to have Flightradar24 (Q1164485) properties. The question is, do we have items on individual planes? Can we? And what about flights? Shouldn't we have them too? Thierry Caro (talk) 17:18, 26 April 2018 (UTC)

By chance I come across the newly created OK-MEK (Q52195573) just now. Doubtless we have others; Enola Gay is the only one that comes to mind. On individual planes, my vote is, knock yourself out; they appear to meet Wikidata:Notability point 2. Somewhat less convinced by flights, though routes / flight numbers has some appeal. --Tagishsimon (talk) 19:17, 26 April 2018 (UTC)
See query. Most of them are false positives, but there are 100+ correct mostly at the end of the list. There are plenty of single plane categories on Commons:Category:Aircraft_by_registration, but unfortunately two registrations of same aircraft are two categories.--Jklamo (talk) 19:35, 26 April 2018 (UTC)
Regarding flights between Paris (Q90) and Houston (Q16555), for instance, we may have items named 'air route between Paris and Houston' to deal with everything happening between those two cities or – and that's what I was thinking about – items for each of the scheduled routes connecting the two by a specific airline, for example 'AF 0636' for the Air France (Q131005) flight. I was not thinking either about having items per individual flight such as 'AF 0636 on March 26, 2018'. So, would anyone be uncomfortable with us having 'AF 0636' and the likes? Thierry Caro (talk) 08:29, 27 April 2018 (UTC)
Comfortable. --Tagishsimon (talk) 11:17, 27 April 2018 (UTC)
@Thierry Caro: My only concern is how stable the flight numbers are. I know that they are taken out of service after a major accident (hence why accidents can use them), but apart from that do routes keep the same flight numbers indefinitely, or can they change every so often? Items on individual planes would be very useful, though, as they can be matched up with the commons categories and we can then use commons:Template:Wikidata Infobox in those categories. Thanks. Mike Peel (talk) 11:32, 27 April 2018 (UTC)
 Comment The idea to add flights has also arisen on the German project chat a fortnight ago: Wikidata:Forum#Hilfe gesucht. --Marsupium (talk) 10:22, 27 April 2018 (UTC)

ISNI unique value violation created by mix'n'match

User:ArthurPSmith, on 2018-02-13 you added an ISNI to several items where it already existed, some examples: [17], [18], [19]. I don't know the tool, but maybe this can be prevented in the future? Or a bot could remove duplicates? I found the duplicates when checking Wikidata:Database_reports/Constraint_violations/P213#"Unique_value"_violations. 78.48.189.69 10:05, 22 April 2018 (UTC)

User:Sic19, at 2018-03-20 07:06 (same minute) you added the same ISNI twice, once with spaces, once without [20]. Maybe mix'n'match has a problem with handling two representations of ISNI? 78.48.189.69 10:13, 22 April 2018 (UTC)

I usually change the formatting manually after adding an ISNI with mix'n'match and haven't seen it add one formatted with spaces before. Not sure what happened here and I hope it is an isolated instance. Simon Cobb (Sic19 ; talk page) 17:06, 22 April 2018 (UTC)
Yeah, I noticed later that the mix'n'match ISNI upload contained id's in a different format than we store them in wikidata, so it was failing to recognize already existing entries etc. It's been a while since I tried it; this was definitely broken, so if it hasn't already been fixed, the mix'n'match for the various ISNI catalogs should probably be reloaded with ID's in wikidata format at least. ArthurPSmith (talk) 12:41, 23 April 2018 (UTC)
@Magnus Manske: Mahir256 (talk) 13:09, 23 April 2018 (UTC)
ISNIs were imported in the wrong format into Mix'n'match. Garbage in, garbage out. Fixing Mix'n'match data now. --Magnus Manske (talk) 15:36, 23 April 2018 (UTC)

@Magnus Manske, Sic19, ArthurPSmith: - seems the Wikidata-specific way of storing ISNI caused trouble. Why not fix the root cause? 77.180.35.91 18:20, 23 April 2018 (UTC)

It's not Wikidata-specific, it's the standard ISNI display format, but obviously there are (at least) 2 different ways of representing ISNI as a string and you have to pick just one of them. This one at least makes clear it's a string and not a number (where leading zeroes might be truncated, giving yet a third representation). ArthurPSmith (talk) 19:30, 23 April 2018 (UTC)
ArthurPSmith, "It's not Wikidata-specific" - then which other DB, similar in size to Wikidata or VIAF stores it like Wikidata? Regarding the number of presentations you talk nonsense. They are endless. But what matters is how the ISNI format is defined by the ISO standard:
4 Structure and syntax of the ISNI
4.1 An ISNI shall consist of 16 digits. It shall consist of two components:
a) 15 decimal digits, and
b) a check character.
4.2 [...]
4.3 When an ISNI is written, printed or otherwise presented in a human-readable format:
a) it shall be preceded by the letters ISNI, separated from the identifier by a space, and
b) the 16 digits shall be displayed as four blocks of four digits, with each block separated from the next
by a space.
EXAMPLE ISNI 1422 4586 3573 0476
The characters “ISNI” and the space characters shall not be considered to form part of the ISNI.
What is not clear about "The characters “ISNI” and the space characters shall not be considered to form part of the ISNI"? 77.179.186.215 00:36, 27 April 2018 (UTC)
Well Wikidata are certainly not the only ones confused about this then. The GRID institution database, for instance, in its JSON data file or external_ids.csv file, pretty uniformly has ISNI in the space-separated format. ArthurPSmith (talk) 00:42, 27 April 2018 (UTC)
Thanks for reporting on GRID! 85.179.115.121 18:42, 27 April 2018 (UTC)
It would be ideal, 77.179, if we could have a lot less heat ('you talk nonsense', 'What is not clear about'), and a lot more light. In all good faith, it's not clear which of at least three possible representations you're so fiercely supporting.
  • A contiguous 16 digit string (where in, to echo APSs point, we should have concerns about downstream truncation of leading zeros)
  • A quad of space separated numeric quads
  • A quad of space separated numeric quads preceded by ISNI <space>
On the one hand I'm guessing you're advocating the first of these, but on the other, in so far as we present a human-readable ISNI, the third of these looks more attractive. So, I'm a little confused (and not actually that interested, one way or the other, so long as the duplication problem goes away). What does bother me is unnecessary attacks on other volunteers who are working and advocating their positions in good faith, and who should be dealt with respectfully. --Tagishsimon (talk)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Tagishsimon, why do you repeat the -censored- about "at least three possible representations" - there are endless representations. But only one in the spec is defined to be an ISNI. All others are derived. And only one derived form is in the spec. I am not advocating any format, I am advocating compliance with the spec. Wikidata is a database. The value in the database should be the ISNI if the property is called ISNI. Otherwise the property would need a new label "Wikidata representation of ISNI."

ISO 27729 4.3 : The characters “ISNI” and the space characters shall not be considered to form part of the ISNI
ISNI Example Length Comment
ISO 27729 4.1 ISNI 0000000000000095 16 char some applications may by default remove leading zeros, easy to select by double click, used in isni.org and viaf.org-URLs
ISO 27729 4.3 ISNI ISNI 0000 0000 0000 0095 24 char may be on different lines, e.g. tables, may flow different in rtl-scripts, no selection by double click
Wikidata ISNI 0000 0000 0000 0095 19 char may be on different lines, e.g. tables, may flow different in rtl-scripts, no selection by double click

Can you create a SPARQL for listing all external ID properties and show the regex? 85.179.115.121 18:42, 27 April 2018 (UTC) 

SELECT distinct ?property ?propertyLabel ?regex WHERE {
   ?property wdt:P31/wdt:P279* wd:Q19595382 .
  ?property wdt:P1793 ?regex . 
  SERVICE wikibase:label { bd:serviceParam wikibase:language 'en' }
} order by ?propertyLabel
Try it!
- it seems so. --Tagishsimon (talk) 18:55, 27 April 2018 (UTC)

Tagishsimon, helpful, already makes more inconsistencies in Wikidata visible, e.g. [0-9] and \d are used, so grouping will fail. But the query does not match the request which was for "all external ID properties" (Data type= ...), not for direct(!) instances of Wikidata property for authority control for people (Q19595382). A workaround could be to select all properties that have regex. But how to select all properties, i.e. select not based on instance of, but on the fact they are properties? Now at Wikidata:Request a query#Show Wikidata properties that are external identifiers and have a regex. 85.179.115.121 21:08, 27 April 2018 (UTC)

Motorways and controlled-access roads

M1 motorway (Q19875) and other British motorways seem to have instance of (P31)=controlled-access highway (Q46622). "Motorway" seems to be an alias for that, and the description is "highway designed exclusively for high-speed vehicular traffic, with all traffic flow and ingress/egress regulated". Except that these motorways are not controlled-access, and don't have regulated ingress/egress. Maybe this is the result of a past-merge-gone-wrong, although I can't easily spot it. Any thoughts? Thanks. Mike Peel (talk) 00:39, 23 April 2018 (UTC)

controlled-access highway (Q46622) looks like a euphemism to avoid having to choose between motorway, freeway, expressway ... to stop the English speakers falling out amongst themselves. There are regulations on access to UK motorways, and so 'ingress/egress regulated' is more or less appropriate. The only fault I see with M1 motorway (Q19875) is instance of (P31)=road (Q34442) since controlled-access highway (Q46622) is already a subclass of road (Q34442). --Tagishsimon (talk) 00:58, 23 April 2018 (UTC)
@Mike Peel, Tagishsimon: In jawiki they have ja:自動車専用道路 (from limited-access road (Q14875255)) which could probably be motorways that (with the exception of automobile-only) do not control accessibility. --Liuxinyu970226 (talk) 04:16, 23 April 2018 (UTC)
Ah, looking into this some more, it's likely to have come from enwp - see en:Controlled-access highway (Although as an aside, something like "restricted-access highway" would be more correct). So let me revise my question: should we continue to use a single Qid for the different terms, or split them into separate ones ('motorway' for UK items vs 'freeway' for US items and so on)? Thanks. Mike Peel (talk) 12:14, 23 April 2018 (UTC)
Agree 'restricted' probably better than 'controlled'. Individual QIds become useful where there are values attaching only to subsets of roads: one can make the case that right now, country is such a value (at least for UK roads) ... certainly not harmful & probably beneficial, so long as they're subclasses of 'restricted-access road'. GFI. --Tagishsimon (talk) 13:25, 23 April 2018 (UTC)

Very easy, UK's M1 motorway is an instance of UK motorway. 85.179.115.121 18:51, 27 April 2018 (UTC)

Alexis900 (talk • contribs • logs) Asqueladd (talk • contribs • logs) BeneBot* (talk • contribs • logs) Detcin (talk • contribs • logs) Dough4872 (talk • contribs • logs) Gz260 (talk • contribs • logs) Happy5214 (talk • contribs • logs) Imzadi1979 (talk • contribs • logs) Jakec (talk • contribs • logs) Labant (talk • contribs • logs) Liuxinyu970226 (talk • contribs • logs) Ljthefro (talk • contribs • logs) mxn (talk • contribs • logs) naveenpf (talk • contribs • logs) Puclik1 (talk • contribs • logs) Rschen7754 (talk • contribs • logs) Scott5114 (talk • contribs • logs) SounderBruce (talk • contribs • logs) TCN7JM (talk • contribs • logs) TimChen (talk • contribs • logs) Bodhisattwa (talk • contribs • logs) Daniel Mietchen (talk • contribs • logs) Tris T7 TT me Izolight (talk • contribs • logs) Gnoeee (talk • contribs • logs)

Notified participants of WikiProject Roads--Jklamo (talk) 11:51, 23 April 2018 (UTC)

Abriss der Psychoanalyse (Freud)

The articles in German, Spanish and French about Freud's book Abriss der Psychoanalyse are used by item Q19606247, while the article in English is used by item Q16002914. --Pablo.ad (talk) 02:38, 25 April 2018 (UTC)

Seems like quite a mess, but the end solution will probably not bring those sitelinks together. Sjoerd de Bruin (talk) 08:34, 25 April 2018 (UTC)
Hello Pablo.ad, It's fixed, thank you : here --> Q16002914, best regards, --Pierrette13 (talk) 20:37, 27 April 2018 (UTC)

Is there a problem renaming a page that has translations, Wikidata:Data donation?

Hi all

I'm thinking about suggesting renaming Wikidata:Data donation to something like Wikidata:Publishing data on Wikidata, the current title is not a correct description of the page and also suggests people publishing data on Wikidata are giving something up which is not very helpful. The page has 20+ translations, is renaming the page as simple as moving it, or will this break something? Also what do people think about the name change, its an important page to get correct as it is a well used page for partners interested in sharing their content on Wikidata.

Thanks

--John Cummings (talk) 07:48, 25 April 2018 (UTC)

Translated pages offer a special moving mechanism in place of the normal Special:Move. If there are no conflicts there will be no problem. Sjoerd de Bruin (talk) 08:32, 25 April 2018 (UTC)
If it also has subpages, then there can be problems, see phab:T114592. If not, then it should be straightforward... Thanks. Mike Peel (talk) 08:48, 25 April 2018 (UTC)
Thanks @Sjoerddebruin: and @Mike Peel:, should I just put a question on the talk page to see if anyone objects to the move? Could I ask one of you to move the page once agreed? There are no subpages but I don't want to make my first practice move on something with so many other languages. --John Cummings (talk) 13:58, 25 April 2018 (UTC)
Yes. ;-) Thanks. Mike Peel (talk) 23:31, 27 April 2018 (UTC)

named after (P138)

I'm trying to add named after (P138) to Fisher's equation (Q1763840), however the equation is also known as Kolmogorov–Petrovsky–Piskunov (KPP) equation. I could simply add all four persons to named after (P138), but then we won't be able to tell that one of them applies to one name of equation, and the other three applies to another name. Is there a way to keep this information?--Stevenliuyi (talk) 19:57, 24 April 2018 (UTC)

Which image is better?

Is there any way (other than manual labour) to determine which image (Wikipedia infobox or Wikidata P18) is better for articles listed here? Capankajsmilyo (talk) 06:52, 28 April 2018 (UTC)

Changing NCI Thesaurus ID (P1748) to data type "External identifier"

Currently, NCI Thesaurus ID (P1748) is of type String. It is, however, an External identifier. It apparently can be changed accordingly but requires community consensus. Can we reach consensus here? Who supports, who objects and why?

--Andrawaag (talk) 11:00, 24 April 2018 (UTC)

--I completely support this change: its does not make sense to have NCI Thesaurus ID as a string will all other medical ontologies are "External Identifiers"--Amb sib (talk) 11:18, 24 April 2018 (UTC)

So to be clear, external identifiers are most useful when they uniquely identify the item they are attached to. The problem with the above cases seems to be that these, despite being called "identifiers", are not used in that way within wikidata: they do identify something external, but a single external entity is attached to multiple items within wikidata. We can change the way these are being used within wikidata (fix what would be constraint violations) and then request they have their datatype changed, or leave things as they are. ArthurPSmith (talk) 12:23, 24 April 2018 (UTC)
 Support I often use this property and I know this is obviously an external identifier. Most of the single value violations were imported from Disease Ontology. --Okkn (talk) 14:30, 24 April 2018 (UTC)
 Support Andrew Su (talk) 23:32, 24 April 2018 (UTC)
 Support --Daniel Mietchen (talk) 22:16, 28 April 2018 (UTC)

Small businesses and notability

The Wikidata:Notability guidelines are significantly looser than those of the English Wikipedia. The requirement is that an entry refers to an instance of a clearly identifiable conceptual or material entity. The entity must be notable, in the sense that it can be described using serious and publicly available references. For "small businesses", would an entry in a chamber of commerce directory (for example [21]) which includes a business website and hours operation be sufficient to meet this threshold?

Slightly differently, is there a rule regarding individual Walmart/McDonalds franchise locations? Right now these appear to be excluded, is this by policy or simply because they haven't been added yet? Power~enwiki (talk) 18:37, 27 April 2018 (UTC)

MD Imtiaz Ahammad Kopiersperre Jklamo ArthurPSmith S.K. Givegivetake fnielsen rjlabs ChristianKl Vladimir Alexiev Parikan User:Cardinha00 MB-one User:Simonmarch User:Jneubert Mathieudu68 User:Kippelboy User:Datawiki30 User:PKM User:RollTide882071 Andber08 Sidpark SilentSpike Susanna Ånäs (Susannaanas) User:Johanricher User:Celead User:Finnusertop cdo256 Mathieu Kappler RShigapov User:So9q User:1-Byte pmt Rtnf econterms Dollarsign8 User:Izolight maiki c960657 User:Automotom applsdev Bubalina Fordaemdur DaxServer

Notified participants of WikiProject Companies

I've posted on the WikiProject talk page. Power~enwiki (talk) 20:46, 28 April 2018 (UTC)

Number of deaths in a battle

Hoi, in a battle there are participants, in my opinion, the people who commanded, the number of deaths could be qualifiers to the country / power / unit that participated. It is not obvious what units belong to what participant, particularly in translation and in combination with historic countries.. How to model this? Thanks, GerardM (talk) 20:34, 28 April 2018 (UTC)

Jarbot

User:JarBot is currently mass creating new objects, with links to article in the Arabian language-Wikipedia. However, many of these object already exist. Merging takes long time, because if you enter Arabian language-Wikipedia and write with Latin letters, text will appear in Arabian letters. J 1982 (talk) 17:47, 27 April 2018 (UTC)

@جار الله: @J 1982: Why complain about a bot without pinging its operator? Mahir256 (talk) 20:03, 28 April 2018 (UTC)
@جار الله: I am already seeing duplicates even on arwiki: w:ar:قالب:رؤساء الكونغو الديمقراطية ↔ w:ar:قالب:رؤساء جمهورية الكونغو الديمقراطية, for example. Mahir256 (talk) 20:21, 28 April 2018 (UTC)
@Mahir256, J 1982: as you can see now the bot just add link to item but when the bot creating new objects is only because these articles were not based on en articles. anywhy now the bot checks if the en Label is in enwiki the bot will add ar link and don't create new objects.--جار الله (talk) 20:44, 28 April 2018 (UTC)
About templates @Jura1: said to me that the bot should not created items for templates with only 1 sitelink, the bot will not created items for templates with only 1 sitelink anymore. and in arwiki I always handle the duplicates.--جار الله (talk) 20:44, 28 April 2018 (UTC)
Everything is getting better now (and I've fixed Latin letters settin at Arwiki), but I just want to say that not all objects alredy existing link to an article in English. For example, Belgium and Switzerland linked to French, just for examples. J 1982 (talk) 15:58, 29 April 2018 (UTC)

A WordPress plugin with Wikidata

Hi,

I created a WordPress plugin which works with Wikidata for my final project in the University, and I need some feedback. It only takes 10 minutes: https://contexto.thetwentyseven.co.uk/research/

PS: Basically, I developed Page Previews for WordPress. However, I did not know that this feature exist in Wikidata when I start developing the project but I will reference it now

Regards  – The preceding unsigned comment was added by Urirnal (talk • contribs) at 15:03, 29 April 2018 (UTC).

Why numbers are in English format only?

Hello, I was trying to add wikidata data in sawiki infoboxes, but all the output (numerical) is in english numbers. For example see 476 and 13 June 1992 on sawiki. All the text output is in local language, so should the numerical one be. I raised this issue at enwiki and was informed that this is due to wikidata in this comment. Can anyone please help me on this? Capankajsmilyo (talk) 13:21, 16 April 2018 (UTC)

The Wikidata module should handle the localization itself by calling frame:callParserFunction to substitute {{#time: }}, which works well on your wiki. Matěj Suchánek (talk) 14:52, 18 April 2018 (UTC)
It's not generally good coding practice to use parserfunctions for things that can be implemented in Lua. Pppery (talk) 02:54, 19 April 2018 (UTC)
Not doing so in case of localization, which is already well covered in MediaWiki, is just reinventing the wheel. Matěj Suchánek (talk) 07:19, 19 April 2018 (UTC)
We don't require extra code for text values, who do we need it for numerical ones? Capankajsmilyo (talk) 11:43, 19 April 2018 (UTC)
You need extra code for transforming identifier "Q2013" to something useful for the reader, like "Wikidata". Accordingly, you need extra code for transforming +2018-04-23T00:00:00Z. Matěj Suchánek (talk) 20:29, 23 April 2018 (UTC)
If every number stored in Wikidata - and that means numeric values, not just dates - has to be localised by the re-user, then each re-use in a language that doesn't use our Western Arabic numerals requires similar code to be written. Check out en:Modern Arabic mathematical notation #Numeral systems for an example if you don't understand what is being asked for. There are many more re-users of Wikidata than Wikipedias. So now who is reinventing the wheel? It's about time that the community realised that output is just as important as input, and when the same functionality is required by every re-user with a particular need, it is far more efficient to provide that functionality within Wikidata than to expect each end-user to have to provide it themselves. --RexxS (talk) 19:39, 23 April 2018 (UTC)
I do understand what is being asked for and my suggestion did reflect that. But note that Wikidata itself only acts as a data storage. Anyway, if there is a Lua call you miss, you can still give feedback on Wikidata:New convenience functions for Lua or generally on WD:DEV. Matěj Suchánek (talk) 20:29, 23 April 2018 (UTC)
Telling another editor how to localise a time function when they are making a general enquiry about numerals, which includes quantities, coordinates and many other numeric values, not just dates, doesn't support your assertion that you understand what is being asked for. Wikidata certainly does far more than act as data storage because it is structured data, and part of that is the rendering of labels, descriptions, sitelinks and other monolingual text in multiple different languages. The failure to render numerals in the same way is not an insurmountable problem, but it is inexplicable to force every re-user to make the localisation themselves when it could be done once, within the Wikidata interface. --RexxS (talk) 23:01, 29 April 2018 (UTC)
  • This is a known issue that frustrates lots of people. Wikidata has yet to fix the localization of date formats and currently the only way to deal with date formatting when invoking Wikidata in other Wikimedia projects is by local Lua modules or parser functions. See phab:T63958. Deryck Chan (talk) 09:42, 25 April 2018 (UTC)

What duplicate rates should we tolerate?

QZanden talk nl, en, de Capankajsmilyo talk en, hi, sa Epìdosis talk it, en, fr Liuxinyu970226 talk zh, en, ja, ko Marsupium talk Iamthecheese44 talk en Matlin (talk) 14:48, 25 August 2020 (UTC) 1234qwer1234qwer4 ru, de, en Hwem talk ru, en, ja 2le2im-bdc talk fr, de, en FakirNL (talk) nl, en, de Samoasambia (talk) fi, en, sv

Notified participants of WikiProject Duplicates

Creation of duplicates is a general problem, especially for mass creating items from external databases. Discussing the most recent and very general case I hope we can find consensus on what to tolerate for future cases!

Since Tuesday, Reinheitsgebot has created more than 3000 human items with the mixnmatch_people_creator by Magnus Manske (source code). I've reviewed the last 20 items of the first run as a random sample. Of those 5 were duplicates. This is 25%.

I see duplicates as a serious problem and think this is much too high. I propose to complement the first paragraph of Wikidata:Bots#Approval process after “before the test run is started.” with

“A bot task creating items must show not to exceed a duplicate rate of THRESHOLD% demonstrated with a REASONABLE sample.”

and Wikidata:Blocking policy after “Unauthorised and/or malfunctioning bots” with

“and users running automated tasks except for those that can be stopped separately (QuickStatements batches). Malfunctioning includes creating items with a duplicate rate higher than THRESHOLD% demonstrated with a REASONABLE sample.”

I'd personally favor a threshold of perhaps 2%. I expect many will think this is too low, but let's see! The qualities of the sample would have to be figured out.

Thanks for your thoughts! --Marsupium (talk) 15:04, 27 April 2018 (UTC)

It might be better if we were to look at what a reasonable sized sample was and worked backwards from there and decide what level of certainty we need (eg if there are x duplicates in a sample of n records, then we can state with 95% certainty that no more than y% are duplicate). Martinvl (talk) 16:33, 27 April 2018 (UTC)
  • Interesting question. I think it depends on:
    1. - the type of item (e.g. person or place),
    2. - the information available on creation (identical name in Latin script, sitelink, lifespan, coordinates, etc.),
    3. - the tools available to merge after creation or de-duplicate them prior to creation
    4. - and finally the tools available to add more data to created items (to figure out that they are actually duplicates).
    The recent tendency seems to be to merge items (thanks to GZWflood).
    --- Jura 16:42, 27 April 2018 (UTC)
@Jura1: To 1.) I'm more engaged in persons, how do you think places should be treated differently? To 2.-4.) I should have been more specific about that, all is meant for the comparison human vs. bot editor, both will make mistakes, but I think we should settle which error rate of a bot that a careful human editor (with all available tools and means of 2.-4.) wouldn't have made is still tolerable and which not. Does that make sense? (BTW: In a way User:GZWDer (flood)’s items are in a way less harmful by not having any statements.) For this case I've pointed out answers to 2.-4. at User talk:Magnus Manske #Reduce duplicates by mixnmatch_peoplecreator. --Marsupium (talk) 12:52, 28 April 2018 (UTC), 13:07, 28 April 2018 (UTC)
I guess people are pretty much always a 1:1 relationship (it is a duplicate, or it is someone else) whereas places are often a bit fuzzier (is this village the same as that parish? does that protected area mean quite the same thing as the island it covers? is this building the same as the business in it?) - so you might want different approaches for them. Andrew Gray (talk) 14:46, 28 April 2018 (UTC)
  • I blew the dust off some of my stats books and looked at the problem from a numerical point of view. The process would be as follows:
  • Choose the value of p (typically 0.02 for 2%)
  • Randomly select n items from the list to be merged and check for duplicates.
  • Assume that the number of duplicates is given by d
  • If d < (n*p - 1.65*sqrt(n*p)), then accept the batch as the proportion of duplicates is less than p with a 95% certainty.
  • If d > (n*p + 1.65*sqrt(n*p)), then reject the batch as the proportion of duplicates is greater than p with a 95% certainty.
  • For any other value of d, we have insufficient information, so do some more tests (ie increase n)
  • Repeat until we get can either accept or reject the batch.
Readers who have a stats background will recognise that I am doing a 95% confidence test with the poison distribution approximated by a normal distribution. Martinvl (talk) 17:51, 27 April 2018 (UTC)
@Martinvl: To define a p-level is of course a very clean way. I support to use it in dubious cases, for the case above I don't have to compute anything to know that it is <0,02 :-). I'm afraid there many preliminary more controversial questions than how to find the duplicate rate of a batch. :-) --Marsupium (talk) 12:52, 28 April 2018 (UTC)
  • As far as the value of the threshold (p) goes, 25% is definitely too high; 2% seems a little low. I'd advocate for 5% to 10% as allowable - that cuts the number that might have to be handled manually by a factor of 10 or 20 from the original dataset, so hopefully manageable. A 25% duplicate rate leaves remaining manual work comparable to what you would have to do with no automation at all, so I'd argue the automation is not helpful in such a case. ArthurPSmith (talk) 21:58, 27 April 2018 (UTC)
  • I'm not sure a fixed threshold is helpful here, as it depends on the scale of the import - an import that creates 10 new entries with 5 duplicates is quite different from one that creates a million entries with 50,000 duplicates, particularly if the first is about complex topics but the latter is about very simple ones. It's better to ask people to minimise duplicates as much as possible, but not be afraid to create them in complex cases where a human needs to be involved. Thanks. Mike Peel (talk) 23:30, 27 April 2018 (UTC)
I agree with this. Also, for larger datasets, if the newly made items have over 20 referenced statements and the older items less than 5 unreferenced statements, then it is worth the effort to (manually) merge later in order to enrich the data available in the dataset. Jane023 (talk) 06:10, 28 April 2018 (UTC)
@Mike Peel, Jane023: Do you think we could address this with stating that it is okay for a single bot to create perhaps 100 items at a time, but that it has to stop for the next 100 before the number of duplicates of the last 100 got sufficiently (yet to define) reduced (manually)? Do you have another suggestion for at least a rough standard? Should we decide case by case beforehand, but then what case size needs previous approval? Or do you think the current situation doesn't need reform? I'd appreciate to prevent mass creation of duplicates, but also give bot owners security on what is admissible to do. --Marsupium (talk) 12:52, 28 April 2018 (UTC)
There are two parts to the problem: 1) reducing duplicates on Wikidata and 2) massaging datasets before data upload to reduce the chance of duplicates. Your subject heading implies we can identify duplicates on creation, which is not the case. At best you need an overnight bot run to reveal the dupes. I think guidelines are in order for uploaders so they can both understand the problem and also how to test their data to get a feel for the duplicates. A max or min rate is just nonsense. Jane023 (talk) 13:10, 28 April 2018 (UTC)
@Jane023: We can: It would even be better to publish (a part of) a dataset outside the main namespace before uploading it to make public testing possible, but actually I had a test run in mind. (WD:Bots asks for a “test run of between 50 and 250 edits” after writing a RFP at Wikidata:Requests for permissions/Bot.)
To step back a bit: Do you think the mixnmatch_people_creator upload is desirable? If not, how can we prevent uploads like this one in the future? --Marsupium (talk) 13:58, 28 April 2018 (UTC)
To answer your first answer, I disagree. You can only see that there is a constraint violation which is not necessarily the same thing as a dupe. To answer your second question, I think it depends. You cannot just generate them, but you can use that output for a good start. It's like VIAF - we don't base biographies on it, but we do want to link the item to it once the item is done. Jane023 (talk) 14:34, 28 April 2018 (UTC)

Same old, old arguments ignored

  • The first argument forgotten, ignored; NOT including data means that we can not share the data that we omit because of fear. Data not shared is us being in error because we ignore our objective; sharing the sum of all knowledge.
  • Another fallacy is that "their" data is to be be better than our data. We talk about acceptable percentages but as it is, our error rate and Wikipedia's error rate is unknown, ignored. When we accept in new data no more than 2%, it means that with a 3% error rate OUR failure is 97%.
  • Now consider 100% quality new data but with 6% duplicates, it means that we can compare existing data with the new data. Even with our data having something like 6% errors, we can find what is similar
  • Now consider our 280+ text databases linked together. With no automated consolidation to show that these links are indeed about the same subject. There is a reasonable opportunity to show the likelihood where the errors are.. Now we consider the acceptable rate of duplication in new data. When are we going to consider an acceptable rate of duplication in our own data?
Thanks, GerardM (talk) 17:06, 28 April 2018 (UTC)
I didn't talk about duplicates in new data but actually duplicates in our own data and how to keep it low. --Marsupium (talk) 20:57, 28 April 2018 (UTC)

Hi, I wrote a long reply to this, but it got lost in an edit conflict :-( Long story short, ~4200 items created, didn't anticipate a lot of duplicates, lots of careful checks in place, just checked ~20 items manually, didn't find a single duplicate. People creation bot is deactivated for now. --Magnus Manske (talk) 21:15, 29 April 2018 (UTC)

United States President

If a President servers eight years, can he run again in four years after his term was up?

No. See the text of the ammendment that limits the number of terms a president may serve. Jc3s5h (talk) 19:41, 29 April 2018 (UTC)

create item constraints

Item constraint would allow to check if all instances/subclasses of a certain item satisfy a particular constraint, such as:

Actually, something similar already exists - properties for this type (P1963) or even has characteristic (P1552) - but it doesn't check for constraint violations. What do you think about it?--Malore (talk) 17:26, 26 April 2018 (UTC)

You can use SPARQL queries to extract lists of items not respecting some configurations. No need of specific system to do that. Snipre (talk) 19:18, 26 April 2018 (UTC)
@Snipre:Yes, but I can do that only if I know SPARQL and it's much more effective have exclamation points warnings whenever there is a constraint violation than have to perform a manual query.--Malore (talk) 19:51, 26 April 2018 (UTC)
But the absence of a property is not really a constraint violation, is it; at least in the form we now know it - as something that's broken. How does your proposal materially differ from the excellent Recoin? --Tagishsimon (talk) 20:16, 26 April 2018 (UTC)
@Tagishsimon: item-requires-statement constraint (Q21503247) treats the absense of a property as a constraint violation. Furthermore, "item constraints" can indicate also other inconsistencies like wrong values, a wrong number of values, etc (just like property constraints). Thank you for pointing out Recoin - I didn't know it and it's a great tool. However, Recoin is based on common usage of properties and it offers only suggestions, while with "item constraint" we are sure that the item lacks a property or has a wrong value or something else.--Malore (talk) 11:10, 27 April 2018 (UTC)
Again, when we see how the constraint violation reports are used, I prefer to avoid creation of a system which will generate reports without any care of them. And one remark about constraint definition is that some contraints have so many exceptions that the constraints don't have any sense. So better manage SPARQL queries in the framework of a project or in an user page. And finally constraint violation reports are currently managed by bots and not by wikibase so this will represent some additional work for the rare bot operators doing that kind of job. Again an individual solution is better because it can be tailor-made with SPARQL or directly wit a bot than a general system which won't be used. Snipre (talk) 11:30, 27 April 2018 (UTC)
@Snipre: If I understand correctly, there are two problems:
  • not enough bots;
  • too many exceptions;
As regards exceptions, IMO they are useful because they allow to understand if the constraint is true or should be deleted. Instead, currently it's much more difficult to understand if a properties for this type (P1963) or a has characteristic (P1552) is effectively valid for all instances/subclasses.--Malore (talk) 13:51, 27 April 2018 (UTC)
@Malore: Currently constraint violation reports are supported by bots and not by Wikibase, so yes, we relying on some bot operators and from what I know only 2 bots maintain the current reports. And since several months, we are reducing the number of constraints because we have too many exceptions. Before creating a constraint, you really have to be sure that that constraint is valid and I am afraid that people will just create constraints without looking at the consequences. Snipre (talk) 07:26, 30 April 2018 (UTC)
There is also Wikidata:WikiProject ShEx, which is exploring the use of Shape Expressions (Q29377880) — a method to describe data models and constraints in terms of shapes of RDF graphs — in Wikidata contexts. --Daniel Mietchen (talk) 22:12, 28 April 2018 (UTC)

best practice for elections in which some candidated withdrew

My example in 2018 Meretz leadership election (Q48519106) where 9 registered as candidates, 5 of which resigned before the ballots opened. Do I add the candidates that quit and deprecate their names? If so, what do I add in reason for deprecated rank (P2241)? Thanks, DGtal (talk) 07:15, 30 April 2018 (UTC)

Six months left before Wikidata's sixth birthday

Hello all,

Every year since Wikidata went live, on October 29th, 2012, the Wikidata community celebrates its anniversary all around the world. Local groups organize meetups, workshops, share cake and greetings, to celebrate the work of the editors building up the knowledge base and its ecosystem of tools.

Last year, for the 5th birthday, the WikidataCon, first international conference dedicated to Wikidata was organized in Berlin. Waiting for its return in 2019, the local events continue! We encourage the community to meet up, the chapters and user groups to organize events in order to build a Wikidata group in their area. Even if you're just starting editing, or you've never been organizing an event before, you can be part of the Wikidata's sixth birthday!

The events will take place around October 29th, 2018 (it doesn't have to be on this exact day). You can find a lot of information on this page: Wikidata:Sixth Birthday. You will find the list of events that will take place around the world, ideas of formats, ways to get funded, and documentation about organizing events.

We're excited to see many small or big Wikidata events happening everywhere!

If you have any question or need help, feel free to contact me. Lea Lacroix (WMDE) (talk) 13:09, 30 April 2018 (UTC)

Wikidata weekly summary #310

Here's your quick overview of what has been happening around Wikidata over the last week.

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.

Goal as a criterion

I have added to goal (Q18530): subclass of (P279) --> criterion (Q1789452) because goal (Q18530) can be used with criterion used (P1013) (value type constraint appear). But I am not sure if it is correct... Xaris333 (talk) 20:11, 30 April 2018 (UTC)

Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2018/04&oldid=1102553132"