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

Wikidata:Project chat

From Wikidata
(Redirected from Wikidata:Project Chat)
Jump to: navigation, search
Wikidata project chat
Place used to discuss any and all aspects of Wikidata: the project itself, policy and proposals, individual data items, technical issues, etc.
Please take a look at the frequently asked questions to see if your question has already been answered.
Please use {{Q}} or {{P}}, the first time you mention an item, or property, respectively.
Requests for deletions can be made here. Merging instructions can be found here.
IRC channel: #wikidata connect
On this page, old discussions are archived. An overview of all archives can be found at this page's archive index. The current archive is located at 2017/12.

Project
chat

Administrators'
noticeboard

Development
team

Bureaucrats'
noticeboard

Translators'
noticeboard

Requests
for permissions

Requests
for deletions

Property
proposal

Properties
for deletion

Requests
for comment

Partnerships
and imports

Interwiki
conflicts

Request
a query

Bot
requests

Contents

Google Knowledge Graph API[edit]

I have the Google Knowledge Graph API and a valid key, but can't find the identifier for Kevin Borland (Q40786519). Can someone look it up for me?  – The preceding unsigned comment was added by Richard Arthur Norton (1958- ) (talk • contribs) at 00:47, 2 December 2017‎ (UTC).

Recoin2: New version of relative completeness status indicator for all Wikidata entities[edit]

Dear community,

We are happy to announce version 2 of Recoin, the relative completeness status indicator. Recoin adds a color-coded progress bar to entity pages showing how extensive information about the entity is in comparison with similar entities (see example picture for the album Abbey Road).

Recoin on Abbey Road

In version 2, Recoin is rolled out to virtually all entities in Wikidata (see e.g. Q565400 or Q15074414 ). The core version uses only Wikibase properties, but based on earlier feedback there is also a version specifically for ID properties only.

We think Recoin can be of great help for quickly getting impressions of data quality about entities. We look very much forward to your feedback.

Also, if you support making a gadget out of it please get in contact with us.

More information can be found at Wikidata:Recoin.

Best regards,

Ls1g (talk) 08:20, 6 December 2017 (UTC)

Good news Ls1g! I'm happy to see you moved to the Wikimedia Toolforge, that takes care of the privacy concerns. The project seems to be maturing. Time to turn it into a gadget?
Maybe also time to move it to Wikidata:Recoin? Multichill (talk) 17:32, 6 December 2017 (UTC)
Thanks for the hints! (2) Just moved it to Wikidata:Recoin. (1) Gadget: That would be great - can you advice us what would be the process? Ls1g (talk) 11:08, 7 December 2017 (UTC)
Sadly had to deactivate it again, yesterday it worked fine, but today it slowed down the webpage so much it became barely usable. Does it calculate the completeness after every edit, or only at initial page load, or does some caching? I noticed that properties I added already were still listed among those missing. Ahoerstemeier (talk) 09:49, 7 December 2017 (UTC)
Can you describe the issue further? For me it does not affect the webpage loading, the page loads fast as normal, then a second or two later the indicator is added. Regarding computation, it is done only once, at page load, and uses the SPARQL endpoint. So to see changes the page has to be reloaded (and the SPARQL endpoint has new changes usually after 1-2 minutes). Ls1g (talk) 10:30, 7 December 2017 (UTC)
Update: My colleague just told me it seems something can be optimized regarding the speed of the underlying DB queries, I'll report later. Ls1g (talk) 11:08, 7 December 2017 (UTC)
It was not the page loading which was slow, it was the time needed when adding a statement until I could select the property or type anything. As it went back to normal directly after deactivating Recoin I guess that was the culprit - though a few sections below others report a similar slowness which now recorded as Bug T182322. Ahoerstemeier (talk) 14:27, 7 December 2017 (UTC)
@Ahoerstemeier, Ls1g: There is comment below about page loading slowness as a universal issue. It may be coincident of issues.  — billinghurst sDrewth 22:35, 8 December 2017 (UTC)
Thanks for pointing this out! We also did some optimizations on our side, now Recoin should load considerably faster. Ls1g (talk) 13:07, 11 December 2017 (UTC)
Thanks, I already activated it again and it works like a charm now. And nice to see items I care for to have a green indicator. Ahoerstemeier (talk) 09:09, 13 December 2017 (UTC)

@Ls1g: Hi, I work mainly on "people" items (Q5), and for newly created items, which only have P31... it displays a list if IDs (VIAF, LCCN, etc. -- see screenshot, but NOT sex or gender (P21), country of citizenship (P27), date of birth (P569), date of death (P570), image (P18), occupation (P106), + name in native language (P1559), family name (P734)/given name (P735) that are the most essential properties for people... why is that ? --Hsarrazin (talk) 09:45, 13 December 2017 (UTC)

@Hsarrazin: Hi, please use the other script (recoin.js) - there are two variants of Recoin, recoin.js shows "normal" properties, which is what you and presumably most people want, whereas recoin_id.js, the one you have currently enabled, is a derivative that only shows IDs like VIAF etc. Thanks for pointing this out, I'll fix the description to make the distinction more clear. Ls1g (talk) 13:00, 13 December 2017 (UTC)
@Ls1g: ahhhh... thanks for this piece of info :D --Hsarrazin (talk) 13:15, 13 December 2017 (UTC)

Original research in the Wikidata?[edit]

In Wikipedia, you can make a statement and indicate its source. For example, w: Musician: "A musician (or instrumentalist) is a person who plays a musical instrument or is musically talented. [1]". In Wikidata, this possibility is absent for some cases, which allows you to create original research here: musician (Q639669) (activity of a person) is both profession (Q28640) (person) and performer (Q16010345). Any references to such statements can not be given now. --Fractaler (talk) 13:19, 7 December 2017 (UTC)

Fractaler, fixed by removing the false claims from "musician" [1] 92.226.213.195 14:11, 7 December 2017 (UTC)

I undid the vandalism from the IP user. Please do not mess around with wikidata ontology like that. As to Fractaler's question, there's no reason why references cannot be given for instance and subclass statements; generally they are considered to be obvious and not require references, but if references are needed here I think for example this one which states (1) a person who makes music a profession, especially as a performer of music as a definition works. ArthurPSmith (talk) 15:20, 7 December 2017 (UTC)
The one who is messing is ArthurPSmith. Edit by idiot rreverted, not every musician is a performer. Read enWP "Anyone who composes, conducts, or performs music may also be referred to as a musician" 213.39.186.16 15:29, 7 December 2017 (UTC)
So, here we have: musician (Q639669) is a person. --Fractaler (talk) 18:53, 7 December 2017 (UTC)
Actually, the wikidata item in question refers to a *class* of people. An instance of musician (Q639669) would be a person. musician (Q639669) itself is one of many conceptual classifications of people (like "scientist", "entrepreneur", etc) that are generally described as professions, occupations, hobbies, skills, etc. ArthurPSmith (talk) 19:42, 7 December 2017 (UTC)
As profession? Link, please. Or: original research.--Fractaler (talk) 19:49, 7 December 2017 (UTC)
Link above. ArthurPSmith (talk) 20:21, 7 December 2017 (UTC)
Link above said: "musician: a person who ...". --Fractaler (talk) 06:34, 8 December 2017 (UTC)
No different from ȘOFER (Q40348) which is also an instance of profession (via a subclass), and defined in same source as a person whose profession is to represent clients in a court of law or to advise or act for clients in other legal matters. All professions, occupations, etc. in wikidata are like this: classes whose instances would be individual persons, but the class itself is an instance of "profession" (or one of the other similar classes). ArthurPSmith (talk) 14:36, 8 December 2017 (UTC)
Yes! Source said: person, but Wikidata said: profession. What this original research are doing in the Wikidata?--Fractaler (talk) 18:06, 8 December 2017 (UTC)
Source also said "profession". ArthurPSmith (talk) 18:15, 8 December 2017 (UTC)
Note the enwiki quote above says "A musician is a person..." i.e. an instance of the class musician is also an instance of the class person. But it doesn't say anything about what the class musician itself is (other than a way of filtering persons). That's what professions, occupations, etc. are - classes of people. ArthurPSmith (talk) 18:24, 8 December 2017 (UTC)
Source don't said: a musician is a profession. Ok, if professions are classes of people, then what is their activity, their professional occupation?--Fractaler (talk) 18:51, 8 December 2017 (UTC)
I can't find many wikidata items for the activity of a professional/occupation, but here's one: teaching (Q28545). For a musician I suppose it would be "making music". art (Q735) more generally encompasses what musicians do. ArthurPSmith (talk) 19:28, 8 December 2017 (UTC)
Wikidata's original research said: 1) musician (Q639669) -> profession (Q28640) -> occupation (Q12737077) -> activity (Q1914636) 2) ȘOFER (Q40348) -> legal profession (Q15987302) -> profession (Q28640) -> occupation (Q12737077) -> activity (Q1914636), etc. musician (Q639669) is activity (Q1914636)? ȘOFER (Q40348) is activity (Q1914636)? Or musician (Q639669) (ȘOFER (Q40348), etc.) is human (Q5)? --Fractaler (talk) 14:07, 9 December 2017 (UTC)
Ah, so perhaps occupation shouldn't be a subclass of "activity"? I think you have a case there. ArthurPSmith (talk) 15:55, 9 December 2017 (UTC)
But now WD said: occupation (Q12737077): "any activity of a person". --Fractaler (talk) 16:13, 9 December 2017 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I agree having occupation as a subclass of activity is a problem. “Singing” is an activity. “Singer” is not. The question then becomes what is an occupation? Common sense says a singer is a person who sings. Getty AAT says painters are “People who practice the art of painting”. Since we seem to have agreed that human (Q5) shouldn’t have subclasses (although it has a few), that simple logic doesn’t work here. Is occupation some sort of role? Is it a “description of a person based on activities that person engages in”? An “aspect of human”? - PKM (talk) 19:52, 9 December 2017 (UTC)

@PKM, ArthurPSmith, and Fractaler: I found "Wikidata:What Wikidata is not" guideline saying that original research is disallowed... or that Wikidata is "not a place to put original research". George Ho (talk) 04:03, 11 December 2017 (UTC)
This isn’t really about original research. No one is disputing the “fact” that a musician is a person. The question is how that fact should be represented in our particular ontology. - PKM (talk) 04:40, 11 December 2017 (UTC)
WD now has: "human+human's activity" (what is it?). Moreover, what kind of activity? catalytic activity (Q14353133) (catalytic activity), gastric activity, volcanic activity, etc. also is a activity? As far as I understand, organism (Q7239) is the carrier of either innate (inherited characteristics) or acquired properties (like Acquired characteristic (Q4674771)). The professional/occupation/etc. is similar to the acquired property, but certainly not innate. What is a profession, occupation? It is some skill (Q205961) (acquired characteristic) +/- document on this +/- place of activity? So, human (Q5) can be native speaker (Q924130) (language's carrier), profession/occupation's carrier, etc. And if musician (Q639669) is human (Q5) as "carrier of acquired properties" (and only), then we need "musician (Q639669) (occupation)". And then musician (Q639669) has acquired characteristic (occupation (P106)) "musician (Q639669) (occupation)". Fractaler (talk) 07:53, 11 December 2017 (UTC)
profession (Q28640) has job function (Q828721)? human (Q5) has job function (Q828721)? --Fractaler (talk) 14:09, 13 December 2017 (UTC)

merging entries[edit]

Sorry, I am really new to Wikidata...

I wanted to update this entry : no label (Q21094001)

But then found it was not possible because of this one : La Nouvelle Liberté (Q3211132)

The system suggests merging entries. Good move. But... is it something we simply do by hand ? Or is there a way to merge them with a tool ? How would that work ? Please do not do it for me. Explain me if I can do it myself and if so how.

Thanks

Anthere (talk)

In this case merge isn't the correct gadget, better use move and propose it for deletion, they are different type of item. --ValterVB (talk) 16:28, 7 December 2017 (UTC)
ok thanks Anthere (talk) 19:58, 7 December 2017 (UTC)
@Anthere: there are also two tools to move bits, one for interlinks and another for statements on an item.  — billinghurst sDrewth 22:23, 8 December 2017 (UTC)
@billinghurst: hi. Where are those tools ? Anthere (talk) (nota bene : someone took care of merging the two)
@Anthere: One is a gadget, and the other is User:Matěj Suchánek/moveClaim.js.  — billinghurst sDrewth 20:12, 11 December 2017 (UTC)

Wikidata Liaison for WikiJournals[edit]

Hello all,

I'm involved in the WikiJournal user group (Wikipedia-integrated academic journals; see refs below for more details). WikiJMed is the most established if you want to look at examples. I feel that there are likely great things that can be done in collaboration with Wikidata, but I'm sufficiently novice to not know the possibilities. Some half-formed ideas from a thread on FB's Wikipedia Weekly:

  1. add metadata for WikiJournal articles automatically to WikiData
  2. integrate with data presentation tools like Scholia
  3. broaden with some ResearchGate-like manuscript archiving system using wikimedia commons
  4. add AltMetrics or similar badges (currently available only for whole journal, not individual articles)

I'm therefore looking for a Wikidata Liaison to advise / mediate / work with the current WikiJournals to help with ideas and implementation. What do people reckon?

More info:

Evolution and evolvability (talk) 12:18, 8 December 2017 (UTC)

@Evolution and evolvability: I'm happy to help, but I don't think you need a single=-person liaison. If you do (1) in your list above then (2) will follow automatically. Also, we have Wikidata:WikiProject Periodicals, Wikidata:WikiProject Source MetaData and m:WikiCite which may be useful to you. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:38, 8 December 2017 (UTC)
@Pigsonthewing: You're right of course that a single person isn't really a necessity and that the same can be achieved in a distributed manner across multiple people. I'll have a look into the links you suggested and get back once I've got my thoughts in order (still so much to learn about wikidata!) Evolution and evolvability (talk) 00:20, 11 December 2017 (UTC)
One initial example idea, would it be possible to make a tool that allows people to generate a unified Wikimedia login based on their ORCID ID (analogous to how eLife allows people to log in using their ORCID account). It could even have an option to auto-generate a userpage with basic data from their ORCID page? Evolution and evolvability (talk) 02:44, 11 December 2017 (UTC)
Technically possible yes, but at a cost. See meta:ORCID for some thoughts on possible future integration. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:56, 11 December 2017 (UTC)

Help:Sources does not mention Wikisource[edit]

How do I reference a Wikisource page from an item? Also, shouldn't this basic need be explained in Help:Sources ? Jane023 (talk) 16:45, 9 December 2017 (UTC)

Jane, what do you mean? In an item, you can add a sitelink to Wikisource. Do you mean smth else?--Ymblanter (talk) 17:21, 9 December 2017 (UTC)
For example, this painting, A Rocky Landscape with Great Oaks (Q21015022), should reference de:s:Verzeichniss der von Speck’schen Gemälde-Sammlung, Gemälde Nr. 1 bis 89, since that is where the image comes from. Jane023 (talk) 17:45, 9 December 2017 (UTC)
Hmm and now I see that Woman in the window (Q30302574) is also in the same catalog. We probably have more. Jane023 (talk) 17:50, 9 December 2017 (UTC)
I know very little about Wikisource but I guess you can add it both as sitelink or as a reference.--Ymblanter (talk) 18:11, 9 December 2017 (UTC)
I did this for now. Jane023 (talk) 18:29, 9 December 2017 (UTC)
I thought that any Wikimedia project is not a good source by itself... May be better to use imported from (P143)? --Infovarius (talk) 19:03, 13 December 2017 (UTC)
ah, but wikisource is not the source of the reference : the source is the edition of the work that can be found on wikisource :) --Hsarrazin (talk) 08:19, 14 December 2017 (UTC)
In references, I use “stated in” with the title of the edition and then a reference URL to Wikisource. - PKM (talk) 20:05, 9 December 2017 (UTC)
Ah that would mean making an item. Apparently these don't get automatically created for Wikisource editions? I would expect the landing page to have an item (or the index page). There is no item in this specific case. Jane023 (talk) 20:53, 9 December 2017 (UTC)
<sigh> I spend at least half of my referencing time making works and editions. - PKM (talk) 00:34, 10 December 2017 (UTC)
I am not against making an item but I just can't wrap my head around the distinction between works and editions. Maybe for Hamlet, but art catalogs? (throws up hands) Jane023 (talk) 13:09, 10 December 2017 (UTC)
Sounds good. We need to make sure these don't get replaced with Wikidata:Property proposal/Wikimedia reference link.
--- Jura 20:18, 9 December 2017 (UTC)

Political prisoner[edit]

Good morning! What property can I use for political prisoner (Q217105)? significant event (P793) with qualifiers or occupation (P106) wich I find interpret as occupation of a person; see also "field of work" (Property:P101), "position held" (Property:P39) Breg Pmt (talk) 06:10, 10 December 2017 (UTC)

Political prisoner is a very POV title. In my country someone I consider a terrorist and mass-murderer is often considered a Political prisoner by my enemies. I recommend using prisoner (Q1862087) under significant event (P793) to steer clear of the POV issue. DGtal (talk) 06:34, 10 December 2017 (UTC)
We have place of detention (P2632) to mark where someone was imprisoned. subject has role (P2868) could be used to express that the person is a political prisoner but that should only be done with sources that clearly say so. ChristianKl () 11:52, 11 December 2017 (UTC)
With place of detention (P2632) together with location (P276) and (from - to dates) as qualifiers for subject has role (P2868) as "main" property ? Breg Pmt (talk) 23:08, 11 December 2017 (UTC)
I don't see why you would need location (P276) but otherwise, I agree. ChristianKl () 22:15, 13 December 2017 (UTC)
@ChristianKl: Like for Otto Frank (Q7336) how can I indicate that he was prisoned in Auschwitz concentration camp (Q7341) located in Germany, away from where he lived?. Pmt (talk) 06:05, 16 December 2017 (UTC)
You can use place of detention (P2632) Auschwitz concentration camp (Q7341) and additionally you can use residence (P551) to state where he lived previously. Of course adding qualifiers helps with getting the data clearer. ChristianKl () 12:55, 16 December 2017 (UTC)

Beatrice von Dovsky (Q88821)[edit]

Beatrice von Dovsky: birth and death month got mixed up; aka Beatrice Dovsky, Beatrice Drewikowsky (birth name, also spelled Drevikowsky), Beatrice von Vay (married name), see German Wikipedia for sources or look here pdf p. 955. -- 84.41.34.154 09:28, 10 December 2017 (UTC)

  • I added the dates found on the gravestone. Depending on what is thought to be preferred or incorrect, ranks can be set.
    --- Jura 09:51, 10 December 2017 (UTC)
    • GND doesn't list any months. It was a mix-up, pure and simple. Please delete the wrong dates and please also add the akas. -- 84.41.34.154 10:16, 10 December 2017 (UTC)
      • GND generally has the full dates in the rdf or some other format. As such the other dates shouldn't be deleted. If they are incorrect, they can be set to "deprecated rank". As the item isn't protected, you can change the ranks and add more references directly.
        --- Jura 10:42, 10 December 2017 (UTC)
        • I see. The rdf has gndo:dateOfBirth "1866-07-14" and gndo:dateOfDeath "1923-11-18". Don't understand though why this shouldn't be deleted. -- 84.41.34.154 11:32, 10 December 2017 (UTC)
          • Mainly the first reason at Help:Ranking#Deprecated_rank. In some complicated way, I think it's also possible to find where GND got its value from (was it look up an abbreviation in a PDF?).
            --- Jura 11:40, 10 December 2017 (UTC)
            • Source is "BSB-Musik". BSB stands for Bavarian State Library. -- 84.41.34.154 12:09, 10 December 2017 (UTC)
          • It's valuable information for the GND people when they know that we consider a datum in their databbase to be wrong. On our side, it's valuable to have the data deprecated because it means that a bot who copies data from GND to Wikidata can see that we consider this value to be wrong and thus it won't add the value with the normal rank. ChristianKl () 19:06, 10 December 2017 (UTC)

Help:Sources doesn't include gravestones[edit]

What would be a good way to include them? At Q88821#P569, there is now type of reference (P3865)=headstone (Q203443) and the image (P18) with the image at Commons. A bit basic, but it could be a start.
--- Jura 22:07, 15 December 2017 (UTC)

Retracted paper[edit]

Hoi, I created a qualifier for "retracted paper". I added to one of the more infamous papers. Please comment and we can start add this to for instance what "Retraction watch" has to say and we can include it where a template for retracted papers is used on en.wp. Thanks, GerardM (talk) 13:29, 10 December 2017 (UTC)

Try constructing something with significant event (P793) instead of this constraint violating construction. Sjoerd de Bruin (talk) 13:48, 10 December 2017 (UTC)
That works for me.. Should this not be higher up in the ranking for the item? Thanks, GerardM (talk) 13:57, 10 December 2017 (UTC)
If you use significant event than the item should be named something like "retraction of the paper" and subclass event. ChristianKl () 14:19, 10 December 2017 (UTC)
Wikidata is setup for different data users being able to display data in any order they like. The order in which properties are listed shouldn't be construed as reflecting a rank. ChristianKl () 19:44, 10 December 2017 (UTC)
Why should it be a subclass of a scientific paper. It still is; it is only retracted. I do not understand your notion about ranks here. Thanks, GerardM (talk) 20:45, 10 December 2017 (UTC)
I thought that you meant with "rank" that you have a problem that it's lower down the page. If that's not the case, then that's great. If your argument is that about it being a "scientific paper" I don't think something stops being a scientific paper when it's redacted. We have external ID property that have type constraint that they can only be used with "scientific paper". Those don't suddenly become invalid when the paper gets redacted. ChristianKl () 00:15, 12 December 2017 (UTC)

Mattsenate (talk) 13:11, 8 August 2014 (UTC) KHammerstein (WMF) (talk) 13:15, 8 August 2014 (UTC) Mitar (talk) 13:17, 8 August 2014 (UTC) Mvolz (talk) 18:07, 8 August 2014 (UTC) Daniel Mietchen (talk) 18:09, 8 August 2014 (UTC) Merrilee (talk) 13:37, 9 August 2014 (UTC) Pharos (talk) 14:09, 9 August 2014 (UTC) DarTar (talk) 15:46, 9 August 2014 (UTC) HLHJ (talk) 09:11, 11 August 2014 (UTC) Lawsonstu (talk) 15:15, 11 August 2014 (UTC) Blue Rasberry (talk) 18:02, 11 August 2014 (UTC) Micru (talk) 20:11, 12 August 2014 (UTC) JakobVoss (talk) 12:23, 20 August 2014 (UTC) Finn Årup Nielsen (fnielsen) (talk) 02:06, 23 August 2014 (UTC) Jodi.a.schneider (talk) 09:24, 25 August 2014 (UTC) Abecker (talk) 23:35, 5 September 2014 (UTC) Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:21, 24 October 2014 (UTC) Mike Linksvayer (talk) 23:26, 18 October 2014 (UTC) Kopiersperre (talk) 20:33, 20 October 2014 (UTC) Jonathan Dugan (talk) 21:03, 20 October 2014 (UTC) Hfordsa (talk) 19:26, 5 November 2014 (UTC) Vladimir Alexiev (talk) 15:09, 23 January 2015 (UTC) Runner1928 (talk) 03:25, 6 May 2015 (UTC) Pete F (talk) econterms (talk) 13:51, 19 August 2015 (UTC) Sj (talk) author  TomT0m / talk page guillom (talk) 21:57, 4 January 2016 (UTC) ·addshore· talk to me! 17:43, 18 January 2016 (UTC) Bodhisattwa (talk) 16:08, 29 January 2016 (UTC) Ainali (talk) 16:51, 29 January 2016 (UTC) LeadSongDog (talk) 21:42, 23 March 2016 (UTC) RobLa-WMF (talk) 01:24, 25 March 2016 (UTC) BrillLyle (talk) 04:27, 30 March 2016 (UTC) Kosboot (talk) 20:45, 30 March 2016 (UTC) Sydney Poore/FloNight♥♥♥♥ 15:10, 14 April 2016 (UTC) Peaceray (talk) 18:40, 28 April 2016 (UTC) PKM (talk) 16:29, 1 May 2016 (UTC) Aubrey (talk) 12:42, 25 August 2016 (UTC) Chiara (talk) 12:47, 25 August 2016 (UTC) Marchitelli (talk) 19:02, 1 September 2016 (UTC) YULdigitalpreservation (talk) 17:44, 9 December 2016 (UTC) Satdeep Gill (talk) 14:59, 2 February 2017 (UTC) Pintoch (talk) 09:44, 28 February 2017 (UTC) Raymond Ellis (talk) 16:06, 1 April 2017 (UTC) Crazy1880 (talk) 18:21, 16 June 2017 (UTC) T Arrow (talk) 07:55, 22 June 2017 (UTC) GerardM (talk) 08:25, 30 July 2017 (UTC) With a particular interest of opening up sources about Botany and opening up any freely licensed publications. Clifford Anderson (talk) 18:26, 11 August 2017 (UTC) Jsamwrites (talk) 07:52, 27 August 2017 (UTC) Krishna Chaitanya Velaga (talk) 09:52, 19 September 2017 (UTC) Capankajsmilyo (talk) 18:32, 19 September 2017 (UTC) Hsarrazin (talk) 20:41, 15 October 2017 (UTC) Mlemusrojas (talk) 10:15, 6 December 2017 (UTC) Pictogram voting comment.svg Notified participants of WikiProject Source MetaData --Succu (talk) 23:01, 10 December 2017 (UTC)

I was pointed out to a recent paper that would help us by providing a differentiating in three classes. Have a read. Thanks, GerardM (talk) 15:09, 11 December 2017 (UTC)
The article seems to argue that it would be good if there are amendments and corrections to published articles. Wikidata won't host either and our role is to describe what the journal in question did with their article. If they say they retracted it, we should be calling the process retraction. If they say they amended it, we should call it amendments. Furthermore it would be interesting to have a look at the metadata that the journals use to say that they retract or amend a paper. ChristianKl () 12:52, 13 December 2017 (UTC)

IP edits[edit]

Do we have any research showing which percentage of IP edits is vandalism? We definitely have a lot of IP vandalism in highly visible articles, which stays there for hours and sometimes days, which basically compromises on the idea of using Wikidata directly on the projects. Do we know whether we at least have sufficient benefits from IP edits?--Ymblanter (User talk:Ymblanter) 13:58, 10 December 2017 (UTC)

No research, but some experience: The amount/percentage of vandalism strongly depends on topics and language. I had an eye on German labels and descriptions a while ago using Pasleim’s reCh-Tool and stopped doing so after barely any vandalism by IP and new users showed up. Tha vast majority of edits was good, or at least clearly no vandalism. Spanish and English labels and descriptions, however, seem to attract much more bad-faith edits. —MisterSynergy (talk) 06:31, 11 December 2017 (UTC)
I see on a regular basis vandalism in English and Russian on my watchlist.--Ymblanter (talk) 09:42, 11 December 2017 (UTC)
I've been patrolling recent changes an hour or so each week and there's definitely both good and bad from IP edits. Some IP edits do a lot and seem very conscientious, I think we would lose quite a bit in banning them all. There are also named users who do damage. I think the main thing we need is better organization of our patrolling. How do other wiki's handle it? ArthurPSmith (talk) 15:28, 11 December 2017 (UTC)
As you probably know, in the English Wikipedia there is no organized patrolling except for (i) new page patrol, which goes to the queue, and (ii) pending changes on the pages where those are installed, (iii) ClueBot. Pending changes, installed on the most visible pages, could actually help, I proposed this here some time ago but received no support. Some projects, for example, the Russian Wikipedia, use flagged revisions. All pages in the main space are subject to these flagged revisions, which are by default shown to the readers. Anybody with the flag can patrol some of all of the flagged revisions, and there is a queue organized. When I last edited there more than six years ago, the project was clearly understaffed, and the queue was steadily growing. Sometimes a user would decide to flag thousands revisions per day, which resulted in bad quality (borderline vandalism or incompetent) edits getting through; nobody cared. I do not think we have enough manpower for flagged revisions.--Ymblanter (talk) 15:44, 11 December 2017 (UTC)
Last time I checked the recent changes for properties for ip edits, it turned out to be 80%+ vandalism. We need a better way to deal with this vandalism otherwise we end up with blunt solution like semi-protecting all properties. Multichill (talk) 16:57, 11 December 2017 (UTC)
For properties, to be honest, I do not see why they are not yet all semiprotected. There are not so many of them, we can easily patrol all of them and answer requests (and the only type of requests I can imagine for properties would be adding translations), and vandalism in properties can do too much damage. However, I am wondering whether we should be better off if Ips could not edit items (semi-protection of all items is not an option since project editors who are not autoconfirmed on Wikidata must be able to move and delete pages.--Ymblanter (talk) 09:55, 12 December 2017 (UTC)
I would support semiprotecting all properties. Even when it comes to the translations it's preferable when they are set by people who understand Wikidata and have already made a bunch of other edits. ChristianKl () 13:28, 12 December 2017 (UTC)
Should we start an RfC on this issue?--Ymblanter (talk) 14:55, 12 December 2017 (UTC)
We must have already had one, at least here at the project chat... Matěj Suchánek (talk) 15:05, 12 December 2017 (UTC)
We have too much vandalism on Wikidata, and it's a lot of work to check and find bad edits. It's also one reason why many Wikipedias don't want to use Wikidata in their wikis because then the vandalism made on Wikidata appears on Wikipedia. We had at least two previous discussions about using the Pending changes on Wikidata: Wikidata:Project_chat/Archive/2013/05#Pending_changes and Wikidata:Project_chat/Archive/2017/09#Pending_changes?. I think the extension would be interesting to see on Wikidata, but I think it would not work on Wikibase system as it's now. Users interested of Pending changes can also read m:Requests for comment/Flagged revisions deployment. I fear we don't have enough interested users to check the pending changes and the queue would get too long. It works very well on some wikis, for example on the Finnish Wikipedia the waiting time before edits get approves is usually very short. Stryn (talk) 17:15, 11 December 2017 (UTC)
One solution to getting better at fighting vandalism is to increase the amount of people who follow commonly used items, by letting people automatically follow the item towards which a statement they point as I laid out in my wishlist item.
The other is to get better integration into the Wikipedia watchlists so that vandalism that affects Wikipedia shows up on their watchlists.
I don't think pending changes would be a good step for Wikidata. ChristianKl () 18:09, 11 December 2017 (UTC)
When it comes to Wikis that do use pending changes, there's no reason why they couldn't also use that mechanism for Wikidata related changes. ChristianKl () 20:26, 11 December 2017 (UTC)
Great ideas by ChristianKl here! If the right tools existed, Wikidata would make patrolling more efficient for the Wikipedias, thanks to more people (from various Wikipedias) being available in many timezones to mark edits as patrolled. Syced (talk) 06:26, 12 December 2017 (UTC)
My experience is that the majority of Wikipedia editors will never consciously edit Wikidata (they still get Wikidata edits by moving pages for example, but they would never go and make edits on purpose). That watchlists currently suck does not exactly help.--Ymblanter (talk) 14:54, 12 December 2017 (UTC)
Hopefully that will change in the future when we will have a way for them to edit Wikidata directly from their infoboxes. ChristianKl () 22:06, 12 December 2017 (UTC)
That will multiply our burden to fight this kind of "improvements". --Succu (talk) 22:13, 12 December 2017 (UTC)
The current trend in the English Wikipedia is such that most likely in a year, after a series of RfCs, it will be prohibited to display Wikidata directly in infoboxes in the main namespace. Wikidata vandalism (and our inability to deal with it efficiently) is the most significant reason for that.--Ymblanter (talk) 11:40, 13 December 2017 (UTC)
This is IMO a very pessimistic view on the enwiki activities these days. (Since I’ve read some of the discussions I understand how you came to that conclusion, but I don’t share it.) Yes, there are a couple of very loud editors who are aggressively lobbying against Wikidata, but I can’t see that there is a clear majority of enwiki editors in support of a full Wikidata ban in main namespace. Let’s see how it develops. —MisterSynergy (talk) 12:08, 13 December 2017 (UTC)
Indeed we need to see what happens when actual RFCs are up and who comes to argue what opinions. I however share the concern that Wikidata vandalism is a serious problems, and, in particular, when vandalism in highly visible items does not get reverted for hours and sometimes for days. We must have some solutions at least for the most visible items, even if banning IP edits is not a solution.--Ymblanter (talk) 15:17, 13 December 2017 (UTC)
My last idea on this problem is to have two dispatch queues (ie. mechanism that is used to update Wikidata content in infoboxes). The main one would be for trusted users, like bots, admins or confirmed users, which would work in real time, the second one would only dispatch edits from IPs and newbies when they get patrolled. (Or to only enqueue edits if they are patrolled but that's just implementation detail.) Matěj Suchánek (talk) 15:05, 12 December 2017 (UTC)
any idea of what fraction of edits that need patrolling actually get patrolled here right now? ArthurPSmith (talk) 19:33, 12 December 2017 (UTC)
@ArthurPSmith: You can compare some filters on the new, confusing Special:RecentChanges with the patrol log to get a feel for it. —Justin (koavf)TCM 20:38, 12 December 2017 (UTC)
I don't understand users who call this new feature confusing. It brings new important features for patrolling we didn't have before. It's something Wikidata should take advantage of as soon as possible to combat vandalism more effectively. Matěj Suchánek (talk) 20:52, 12 December 2017 (UTC)
@Matěj Suchánek: Its interface is entirely unlike anything else in MediaWiki and how to add filters is counterintuitive to me. —Justin (koavf)TCM 21:01, 12 December 2017 (UTC)
Last month, there were 33,000 non-confirmed edits, of which almost 6,000 have been patrolled. Matěj Suchánek (talk) 20:52, 12 December 2017 (UTC)
I don't suppose there's any well to tell how many of those have been reverted? Can we tell how many are no longer the latest revision? But that doesn't necessarily help (if one IP makes 10 edits only the last one would be latest, but all would still be active). ArthurPSmith (talk) 21:00, 12 December 2017 (UTC)
I updated the query, the results should be more accurate now. Matěj Suchánek (talk) 21:20, 12 December 2017 (UTC)
In that query what's the difference between "non current" and "non current patrolled"?ChristianKl () 12:54, 13 December 2017 (UTC)
"non current patrolled" ⊆ "non current" (patrolled ∪ unpatrolled). Matěj Suchánek (talk) 17:29, 13 December 2017 (UTC)
  • .. so there are ca. 9000 current edits that are not patrolled.
    --- Jura 17:57, 13 December 2017 (UTC)
  • Ok, thanks for those numbers. It looks like we're doing some of the work needed, but we need to add about 4x our current effort to cover it all. Looking at the manual patrol log it seems we have maybe half a dozen regular users doing it all, plus DeltaBot's major work in this area. @YMS: seems to do by far the most among regular users - perhaps YMS has some insights on how we can improve here? It seems like we at least need a couple dozen more people working on this to tackle it properly. ArthurPSmith (talk) 20:27, 13 December 2017 (UTC)
I don't think I have any particular insights to share, but I indeed use a specialised tool I wrote for myself. It sure won't be ideal for everybody; it won't even be the best choice RC tool for everybody willing to use an RC tool, but I believe it to be a very powerful tool and I hope that I will be able to release a beta version of it to the public by Christmas. This will happen at Wikidata talk:WikiProject Counter-Vandalism first, so if anyone wants to give a try to yet another RC tool, watch this. --YMS (talk) 22:01, 13 December 2017 (UTC)
Could we in the meanwhile adapt this tool to Wikidata? I was trying yesterday, but without much success, and it is very helpful to patrol multiple good edits in the same item.--Ymblanter (talk) 06:11, 15 December 2017 (UTC)

"mayor of X" or "Mayor of X"[edit]

"mayor of X" or "Mayor of X" for the position? I am looking at all presidents, governors, and mayors and there is a mix of both. --RAN (talk) 20:28, 10 December 2017 (UTC)

In my view "mayor of X" should be the position (i.e. it should be an instance of position, or an instance of subclass of position, etc.) It should not be the person, and should not be a subclass of eg "politician". It would be good to clean this up systematically across the site. Jheald (talk) 15:16, 11 December 2017 (UTC)
The capitalized version IMHO is the formal title and the uncapitalized as the generic position. The Mayor of New York City, John Lind, has been serving as the mayor of New York City for 6 years. Standardization would be nice. --RAN (talk) 01:43, 12 December 2017 (UTC)

Professor[edit]

Does anyone have any ideas as to how to better disambiguate professor (Q121594) and professor (Q25339110)? I suspect professor (Q121594) should be the profession and professor (Q25339110) should be the academic rank, and therefore not a subclass of university teacher (Q1622272) (i.e. not a profession). I think the ENwiki links perhaps need to be swapped as well. Any academics want to weigh in? - PKM (talk) 20:31, 10 December 2017 (UTC)

This is completely mixed. For example, the Russian article combines both meanings.--Ymblanter (talk) 20:33, 10 December 2017 (UTC)
Please explain how professor is a profession imho it is only a rank. Thanks, GerardM (talk) 07:01, 11 December 2017 (UTC)
In some countries exists position, that sounds "professor". Maybe that led to misunderstanding? - Kareyac (talk) 12:59, 11 December 2017 (UTC)
The United States Bureau of Labor Statistics has Standard Occupational Classification System (Q7598269) codes for a number of professors by area of practice: 25-1011 Professors, Accounting; 25-1022 Professors, Actuarial Science; 25-1072 Professors, Advanced Nursing; 25-1032 Professors, Aeronautical Engineering, and so on [2]. I think that makes a pretty solid case for "Professor" being an occupation/profession in the US. - PKM (talk) 21:07, 11 December 2017 (UTC)

I think my problem with university teacher (Q1622272) vs. academic lecturer (Q1569495) related to this question. In my understanding first one is a person (and profession) who teaches on a university, second one is (the highest educational) position in a university. Every academic lecturer (Q1569495) is a university teacher (Q1622272), but only very few of university teacher (Q1622272) are academic lecturer (Q1569495). Samat (talk) 21:37, 11 December 2017 (UTC)

Looking at the English Wikipedia articles and English descriptions, I don't see why there are separate entries for professor (Q121594) and professor (Q25339110). They both seem to be about the rank. We already have a separate entry and enwiki article for members of the profession: faculty (Q5428874). Looking again at Samat's comment, I don't see why we have separate entries for faculty (Q5428874) and university teacher (Q1622272). Again, they seem to be the same topic in English, though there may be a distinction in other languages. And we shouldn't have faculty (Q5428874) has part (P527) assistant professor (Q5669847). That's now how has part (P527) should work. If anything, there should be a subclass of (P279) relation in the other direction. MartinPoulter (talk) 17:05, 13 December 2017 (UTC)
Professor means different things in different parts of the world. In the US and Canada a professor is anyone in a regular teaching position at a college or university. There are various ranks such as assistant professor and full professor. In some other countries professor refers only to those at the highest rank, the full professor. The English Wikipedia article en:Professor links to professor (Q25339110) professor (Q121594) (whose description does not match the article, I will fix it) and en:Professor (highest academic rank) links to professor (Q25339110) (whose description does match the article). StarryGrandma (talk) 20:15, 13 December 2017 (UTC)
Fixed links in the above. StarryGrandma (talk) 00:12, 14 December 2017 (UTC)

Contributions do not update[edit]

Hoi, it has been since November 27 that my last update happened. Do we have a problem ? Thanks, GerardM (talk) 15:11, 11 December 2017 (UTC)

Can you be more precise? Sjoerd de Bruin (talk) 16:07, 11 December 2017 (UTC)

Wikidata weekly summary #290[edit]

Football kit[edit]

Delusion23
WFC
happy5214
Fawkesfr
Xaris333
A.Bernhard
Cekli829
Japan Football
HakanIST
Jmmuguerza
H4stings
Unnited meta
محمد آدم
Pictogram voting comment.svg Notified participants of WikiProject Association football

Hello. I am using a lot the data of Wikidata to Wikipedia templates, especially about football. In templates about football teams we can have all the data from Wikidata except the football kit. I had proposed a property about that Wikidata:Property proposal/Home kit (football). I have understood why the proposal what not successful. So, I am writing here to listen to your ideas how we can have the football kit, as a data, to Wikidata. The idea is to be able to fetch the data from Wikidata so we can have the uniform it Wikipedia template. Please read the proposal first and then see how we are using the kid parameters in w:en:Real Madrid C.F.. I know that is not the real uniform, it is a representation of the uniform. But, its a very common way to show the kit of a football team in Wikipedia and we have all the pictures in commons. It would be easier if all Wikipedia could fetch the kit from Wikidata. Xaris333 (talk) 21:31, 11 December 2017 (UTC)

When creating a Wikidata property it's useful to think about how it generalizes. What other sports have things like a football kit? What's exactly the difference to an uniform? Is the kind of uniform that a Marching band carries something similar to a "football kit"?
Do you have good sources that describe how a football gets defined elsewhere? ChristianKl () 18:59, 12 December 2017 (UTC)
@ChristianKl: Of course we can have a general representation uniform property. It can apply for many sports (I haven't see anywhere in Wikipedia representation uniform about anything else except sports). Well, this article explain (with sources) the football kit. Is that what you are asking? Xaris333 (talk) 20:44, 12 December 2017 (UTC)
When creating a property, it's good when it can not only represent one use-case but it can also represent multiple other use-cases. Currently, Navy Working Uniform (Q19297117) is for example linked to no other item but it would likely make sense to link it to another item about the Navy. If the relationship between this uniform and some other Navy item is qualitatively the same as the one between the Football kit and the football club, both use-cases should be handeled by the same property. To do that effectively it would be good to have a list of use-cases that the property should be able to handle well. ChristianKl () 20:56, 12 December 2017 (UTC)
@ChristianKl: What about Runner1928 approach? Xaris333 (talk) 21:01, 12 December 2017 (UTC)
Perhaps you could use wears (P3828). Runner1928 (talk) 20:47, 12 December 2017 (UTC)
@Runner1928: You mean something like Real Madrid FC (Q8682) --> wears (P3828) --> Real Madrid C.F. home kit (Q28199035)? Xaris333 (talk) 20:52, 12 December 2017 (UTC)
Exactly. It looks like start time (P580) and end time (P582) would be appropriate qualifiers too. Runner1928 (talk) 20:55, 12 December 2017 (UTC)
@Runner1928: It may works. The only problem I can find is how to specify that Real Madrid C.F. home kit (Q28199035) is the home kit, Real Madrid C.F. away kit (Q45321396) the away kit and Real Madrid C.F. third kit (Q45321398) the third kit? Xaris333 (talk) 21:01, 12 December 2017 (UTC)
wears (P3828) uses astronaut (Q11631) space suit (Q223571) as one example, so I think using it for sport uniforms would also be okay. criterion used (P1013) could be used to specify home kit/away kit etc. ChristianKl () 21:03, 12 December 2017 (UTC)
@ChristianKl: Ηοw? Xaris333 (talk) 21:14, 12 December 2017 (UTC)
https://test.wikidata.org/wiki/Q110290 ChristianKl () 21:23, 12 December 2017 (UTC)


@Swpb, Dhx1, Pasleim: Do you agree with this approach? Xaris333 (talk) 21:28, 12 December 2017 (UTC)

  • Looks interesting. The actual items being used might need some tweaking, but I like the approach.
    --- Jura 21:34, 12 December 2017 (UTC)
  • Sure, using wears (P3828) makes more sense than adding a new property. Swpb (talk) 21:44, 12 December 2017 (UTC)
  • Symbol support vote.svg Support using wears (P3828). Dhx1 (talk) 11:52, 13 December 2017 (UTC)


I have completed Real Madrid C.F. home kit (Q28199035), Real Madrid C.F. away kit (Q45321396) and Real Madrid C.F. third kit (Q45321398) and added them to Real Madrid FC (Q8682). Please check that 4 items and make suggestion how we can improve them. Xaris333 (talk) 23:21, 12 December 2017 (UTC)

Instead of the statement instance of (P31)=kit (Q1381731) with qualifier applies to part (P518)=home kit (Q45321977), I would directly use instance of (P31)=home kit (Q45321977) --Pasleim (talk) 10:20, 13 December 2017 (UTC)
@Pasleim: home kit (Q45321977) is not only for football. So, we can't have subclass of (P279)=kit (Q1381731). Of course we can change the label to Home kit (association football). Xaris333 (talk) 11:49, 13 December 2017 (UTC)
I dislike having so much information in qualifiers for instance of (P31). I would add a new item "association football home kit" that subclasses both "home kit" (general) and kit (Q1381731). ChristianKl () 12:46, 13 December 2017 (UTC)
We have a more specific property for use. use="home kit" should work.
--- Jura 14:06, 13 December 2017 (UTC)
On what item would you add uses (P2283) "home kit"? ChristianKl () 15:20, 13 December 2017 (UTC)
It's use (P366) ;)
--- Jura 17:13, 13 December 2017 (UTC)
@Jura1: Can you change the item to understand what you mean? Xaris333 (talk) 12:06, 14 December 2017 (UTC)
@Xaris333: ✓ Done at Real Madrid C.F. home kit (Q28199035)
--- Jura 12:08, 14 December 2017 (UTC)
Thanks. Xaris333 (talk) 12:58, 14 December 2017 (UTC)

And one more question: do you prefer to have an item per period (like Real Madrid C.F. home kit (Q28199035)) or a general item "Real Madrid C.F. home kit" and make the changes every year? I am asking because an item "Real Madrid C.F. home kit 2016-17" maybe useless now. And would be easier for the user to change an exist item than create a new one. On the other hand, by creating each period a new item, we can have a "library" for all home kit of Real Madrid during the years. Xaris333 (talk) 16:05, 13 December 2017 (UTC)

It's my intuition to prefer are more general item that qualifies all the relevant statements with "start time"/"end time" and sets the most uptodate state as referred status would be the best but I don't have much subject matter experience, so I don't have a strong opinion. ChristianKl () 16:42, 13 December 2017 (UTC)
So you mean, for example, to have many body (Q170494) values for has part (P527) and have start time"/"end time with it? Xaris333 (talk) 12:06, 14 December 2017 (UTC)
Yes. ChristianKl () 12:50, 14 December 2017 (UTC)
Works for me. Thanks. Xaris333 (talk) 12:58, 14 December 2017 (UTC)

Thank you all for your help. I will start apply it. I don't know any tool that can help me. I think I have to do it by hand. Again, thanks for you help. Xaris333 (talk) 12:58, 14 December 2017 (UTC)

Editing labels and descriptions[edit]

Sometimes when I click the edit button to change the labels and descriptions of an item, a box appears with editable fields and sometimes I am redirected to the Special:SetLabelDescriptionAliases page and I can edit only one language. I don't understand the rule for this, and I don't know how can I reach the first interface in the second case (I would like to edit an other language, but there is no option for that on this surface). Samat (talk) 21:47, 11 December 2017 (UTC)

I think you end up on the "Special page" if you click "edit" before the page has been completely loaded. To get back just click on the back button on your web browser and try again. --Larske (talk) 22:02, 11 December 2017 (UTC)
Thank you for that explanation! - PKM (talk) 02:50, 12 December 2017 (UTC)
Thank you! It looks for me, that you are right. Samat (talk) 08:15, 12 December 2017 (UTC)

Changing the property[edit]

Sometimes there is a big statement group with statements, qualifiers and references, and I should change the property (wrong or deprecated).

  • Is there a way to change the property instead of copying by hand (filling up the separate fields one by one) the whole statement group?
  • Is it possible to copy a whole statement group to a new place (e.g. into a new item, based on a good example)?
  • Does Wikidata have a "wikicode" mode / editing surface maybe? :)

Samat (talk) 21:55, 11 December 2017 (UTC)

Some bots can do that. @Pasleim:. Xaris333 (talk) 22:16, 11 December 2017 (UTC)

  • "move whole statement group to different property" is only possible with bot (or pywikibot code in PAWS), as Xaris333 correctly stated
  • whole statement groups can be copied (or moved) to different items using moveClaim.js by User:Matěj Suchánek (put mw.loader.load( '//www.wikidata.org/w/index.php?title=User:Matěj_Suchánek/moveClaim.js&action=raw&ctype=text/javascript', 'text/javascript' ); // [[User:Matěj Suchánek/moveClaim.js]] to Special:MyPage/common.js, then you’ll see an extra button in each statement)
  • there is no "wikicode" or "wikisource" more for source-based manual editing available

MisterSynergy (talk) 14:39, 12 December 2017 (UTC)

Thank you! Samat (talk) 17:47, 12 December 2017 (UTC)

Consider merging Q19849270 and Q6867899?[edit]

The only conflict is in Cantonese Wikipedia, which IMHO we can encourage those users to discuss if both templates are mergeable or not. --117.14.250.110 02:23, 12 December 2017 (UTC)

no label (Q19849270) and Template:Centralized discussion (Q6867899) --Tagishsimon (talk) 02:42, 12 December 2017 (UTC)
  • There isn't much we can do at Wikidata about it.
    --- Jura 05:45, 12 December 2017 (UTC)
Interesting, merging two templates are having much more possible than two single articles, hence pinging @Deryck Chan: in your opinion should we do that? Affected templates: yue:Template:Bulletin and yue:Template:Cent. --Liuxinyu970226 (talk) 10:52, 12 December 2017 (UTC)
Because Cantonese Wikipedia has a great tradition of following both English Wikipedia and Chinese Wikipedia practices unless they conflict or we create our bespoke system, we've adopted both templates for different purposes.
Normally I'd say Symbol delete vote.svg Disagree because these just aren't the same template, but since yue:Template:Cent is only used by a small set of pages, we have two Symbol split vote.svg Split support and oppose options:
  1. Deprecate yue:Template:Cent and keep Template:Centralized discussion (Q6867899) only, knowing that this item will encompass a bunch of noticeboards that have different design intents.
  2. Keep the two items separate, move all templates based on zh:Template:Bulletin to no label (Q19849270), and said to be the same as (P460) or see also (P1659) the two items. Deryck Chan (talk) 15:15, 12 December 2017 (UTC)
  • This is not the place to discuss what yuewiki ought to do or not. We don't want to import debates from there here.
    Whether Wikidata items can be merged is determined by Help:Merge#Check_to_be_sure. I don't see how the two template items could satisfy that. I don't think people need to read Chinese to understand that.
    Possibly Template:Centralized discussion (Q6867899) needs to be split, but this isn't the question that was asked.
    --- Jura 15:31, 12 December 2017 (UTC)

Freedom in the World (Q1417780) status for countries?[edit]

This status (free, partly free, not free) is often used by NGOs or activist groups. For a project I'm working on it would be very useful to be able to filter countries according to it, but I can't figure out the best way to add this information. Things like review score (P444) and ranking (P1352) aren't quite right. Is there an existing property that would be more suitable? --Oravrattas (talk) 08:09, 12 December 2017 (UTC)

Maybe "has quality" "free according to Freedom House"? ChristianKl () 12:02, 12 December 2017 (UTC)
Oh, nice idea. has quality (P1552) could definitely work here. Though perhaps using Free country (Q3174312) with a statement supported by (P3680) qualifier of Freedom in the World (Q1417780) (or Freedom in the World 2017 (Q30539476)?) might be better than creating a specific "free according to Freedom House" item? --Oravrattas (talk) 16:20, 12 December 2017 (UTC)
I think in this case Freedom House not only supports the claim of it being free but also defines it. ChristianKl () 18:47, 12 December 2017 (UTC)

Where should problems with infoboxes that use Wikidata content by directed?[edit]

I just get access to the OTRS queue and one of the requests is from a person who had trouble with an EnWiki infobox that uses Wikidata content and throws Lua errors. Do we have a place towards which we can direct a person with such a request? ChristianKl () 13:30, 12 December 2017 (UTC)

The person must write the problem to a page to enwiki (template's talk page or w:en:Wikipedia:Village pump (technical)). It is not something that can be solved in wikidata. The problem, I guess, has to do with the modules of enwiki. Xaris333 (talk) 14:00, 12 December 2017 (UTC)

Mineralogy[edit]

Hello. Just a small check. Is my view ok or controversial? Any ideas/comments?
Hierarchy: science > exact sciences > earth sciences > geology > rocks > mineralogy (minerals & nonminerals; mineral classification) > mineral classes & subclasses (Strunz) > mineral supergroups > mineral groups > mineral subgroups (mineral series), minerals, mineral polytypes & varieties.
Orientation (Classification)
Thank you very much --Chris.urs-o (talk) 18:04, 12 December 2017 (UTC)
Are you trying to build a subclass tree? Because rocks, the structures, are quite certainly not a subclass of geology, the scientific discipline. --Yair rand (talk) 19:10, 12 December 2017 (UTC)
I was just thinking. Well, geology studies the Earth crust, I have to think about that. Earth's crust > geological unit > geological formation > rock. But I am working on the tree below mineralogy. Thx --Chris.urs-o (talk) 19:38, 12 December 2017 (UTC)
@Chris.urs-o: As Yair rand suggests, your > means different things here - sometimes it would be a subclass of (P279) relation (from right to left), sometimes it might be facet of (P1269), for some of these things instance of (P31) might also be appropriate. Also what you propose is not entirely consistent with what we have now in wikidata - for example at the moment mineral subclass (Q3965272) subclass of (P279) mineral group (Q1936581) while it appears from your list this relation would be in the other direction. I'm not familiar enough with this field to say what's right here. I suspect that most (all?) of your proposed specific mineral classes also already exist in Wikidata - for example sulfide mineral class (Q927407) - so I would caution you to first look at what already exists here in these areas. If you want to make changes to what's been done already, you can come here to seek consensus, but of course also include Wikidata:WikiProject Mineralogy of which you seem to be a participant already. ArthurPSmith (talk) 19:46, 12 December 2017 (UTC)
Okidoki. Mineral subclass subclass of mineral group is wrong. No panik, I am not changing anything major. Thx
Note: I (124k), Tobias1984 and Sbisolo (in the beginning) did many edits on Wikidata:WikiProject Mineralogy. --Chris.urs-o (talk) 21:08, 12 December 2017 (UTC)

Gathering examples of the search engine not ranking items well[edit]

I dislike that when I type "male" into sex or gender (P21) the item I search is on the third place. When I talked with Lydia, she recommended gathering more examples of where the search engine gives bad sorting. I created https://www.wikidata.org/wiki/User:ChristianKl/List_of_bad_sorting as a list to gather examples. If you have any interesting ones, I invite you to add them to the list. ChristianKl () 23:20, 12 December 2017 (UTC)

  • As P21 has a "one-of" constraint, selection could easily suggested its values first. I think I once read a proposal for that. Didn't you write that?
    --- Jura 17:09, 13 December 2017 (UTC)
  • My first proposal was that constraints in general result in items that violate the constraints being listed after the ones that do. That seems to be computationally too costly.
Limiting this to the one-of constraint would be a possibility that's computationally tractable but it won't fix the issue that "male organsim" gets ranked before "male" because "male organism" has >10 sitelinks and "male" has none. It would get "male" only to the second place by letting it rank better than Malé (Q9347). ChristianKl () 13:24, 14 December 2017 (UTC)

Feeding Wikidata into Wikipedia Infobox[edit]

This is in reference to an earlier query on contributing to Wikidata https://www.wikidata.org/wiki/Wikidata:Project_chat/Archive/2017/03#Contribute_entity_URI_to_Wikidata_2

We notice there are data which exists in Wikidata but are not exposed in Wikipedia infobox. How can we feed the data into Wikipedia infobox?

E.g. National Library Board (Q6974124) – The data “member of Biodiversity Heritage Library Consortium” exists in Wikidata, but not found in the infobox of Wikipedia article National Library Board (https://en.wikipedia.org/wiki/National_Library_Board). Is there a way to expose the data in Wikipedia infobox?  – The preceding unsigned comment was added by Nlbkos (talk • contribs) at 13. 12. 2017, 06:04‎ (UTC).

The infobox must support this data stream. You need to ask on the English Wikipedia whether they want to enable it. Matěj Suchánek (talk) 08:23, 13 December 2017 (UTC)
To be specific, if you want that something gets included into the authority control template of the English Wikipedia https://en.wikipedia.org/wiki/Template_talk:Authority_control is the place to raise the request. ChristianKl () 19:09, 13 December 2017 (UTC)

Editing name, description, "also known as"[edit]

Wikidata seems recently to have unilaterally decided that when I go to edit name, description, & "also known as" that I'm only editing the English. This is annoying, given that I read/write several languages. I did not deliberately change any settings. Is this some sort of global change? Is there anything I can do to be able to get back to editing these fields in several languages? - Jmabel (talk) 07:01, 13 December 2017 (UTC)

You mean you can see no other language field than English? Beneath the table, there should be "All entered languages". Can you see that? Few weeks ago, several users reported a similar problem, which was then solved by null-editing their userpages. I have just done this for you. Matěj Suchánek (talk) 08:26, 13 December 2017 (UTC)
Didn't work at first, but seems to be working now. For what it's worth, I still had to play with it a little, but now it works as it should. I'm afraid I can't fully describe what "playing with it" consisted of. - Jmabel (talk) 16:09, 13 December 2017 (UTC)

{{Section resolved|Jmabel (talk) 16:09, 13 December 2017 (UTC)}}


  • Couldn't we add a note about this on top of the special page?
    --- Jura 17:11, 13 December 2017 (UTC)

Improving the data import process[edit]

Community Noun project 26481.svg

Background:

The Request for Comment page and related discussion page linked to below follow on from the session held at WikidataCon 2017, which aimed to gather feedback from the community about problems/difficulties with the current data import flow. The main objective is to use this area to build a picture of what already exists and what's missing in the data import process. This can then be used to create actionable tasks (which we're proposing be managed on Phabricator).

What we want:

We want to get as many community members as possible involved in these early stages of planning. The end goal is to have a well structured task list showing the improvements we need to make to the data import process. We have created a draft page showing the general steps for the data import process, indicating the resources we have and what is still needed. We've also proposed a possible structure for a project on Phabricator, which also needs to be reviewed by the community.

If you have anything to add please go ahead and edit the page and/or join the discussion.

Wikidata:Requests for comment/Mapping and improving the data import process

Any help greatly appreciated! Thanks, NavinoEvans (talk) 10:22, 13 December 2017 (UTC)


The inclusion of the pages makes the TOC here rather big. Sjoerd de Bruin (talk) 10:51, 13 December 2017 (UTC)
@NavinoEvans: I would recommend that you create a Request for Comments for this discussion as that's a better way to deal with long documents than having them in the project chat. ChristianKl () 11:51, 13 December 2017 (UTC)
I'll sort this out to be an RFC today. Thanks, --John Cummings (talk) 13:23, 13 December 2017 (UTC)
  • Somehow this transclusion messes up the TOC of Project Chat.
    --- Jura 19:50, 13 December 2017 (UTC)
Changed to a linked RFC page, thanks. --John Cummings (talk) 20:08, 13 December 2017 (UTC)

Google often serves mobile pages for Wikidata even though I'm on a desktop[edit]

When I Google for Wikidata pages I often get m. mobile pages even though I'm on a desktop PC. Do other people find that the issue also exist for them? ChristianKl () 15:32, 13 December 2017 (UTC)

oh yess, very often !! :(( - also on wikipedia and wikisource, but it's even more problematic on wikidata, due to the format being data and not text... --Hsarrazin (talk) 15:49, 13 December 2017 (UTC)
Do you use a mobile device while logged in to Google? AFAIK it remembers many things about your behavior and this could be a side effect. Matěj Suchánek (talk) 17:57, 13 December 2017 (UTC)
I never had this effect (I am sometimes using IPad but then choose the desktop version).--Ymblanter (talk) 18:12, 13 December 2017 (UTC)
I do use Google sometimes on my smartphone with uses the same Google account as my desktop computer but it would seem strange if Google isn't smart enough to separate the two uses. ChristianKl () 19:07, 13 December 2017 (UTC)
There are several tasks for this in Phabricator, like phab:T60425 and phab:T72045. I find the fact that the mobile site doesn't redirect to the desktop site when you're on desktop extremely annoying, but the Foundation folks don't seem to agree. Maybe a few more voices on those tasks could have an impact? Jon Harald Søby (talk) 20:40, 13 December 2017 (UTC)

UK places with multiple heritage listings[edit]

I'm having another issue with UK listed buildings (so paging @Anvilaquarius: who came in on the last such thread).

Background: I have been preparing an upload to Commons of a substantial number of old maps and plans from 19th century books. (See batch preparation pages -- UK initially, but much more from the rest of the world to follow).

Through a process of georeferencing, I have the coordinates of the corners of the map or plan. Given these co-ordinates I have been using Wikidata to help identify what the map or plan may actually be of, for features up to about 250m from side to side. (For bigger objects than this, there are generally lots of Wikidata location hits, and it's difficult to filter which ones might be most likely for somebody to have made a map of, of that size -- but if people have some good ideas I would be very interested. Identification is already fairly unreliable at the 200m scale).

The identification process I have been using has been quite simple minded:

  • Find the nearest Wikidata item to the centre of the map, that has a Commons category (P373); or, failing that
  • Find the nearest Wikidata item.

Using the Wikidata look-up I have also been able to separate out plans for cathedrals, castles and churches / abbeys / monastic ruins.

But I have hit a problem with some places for which we have multiple items -- for example, looking at the Commons category page for c:Category:Gloucester Cathedral, it has four entries in the register of listed buildings, each of which we have a different item for, all of which link to the Commons cat, but only one of which (and not the nearest to the centre of the map) is instance of (P31) cathedral (Q2977). Similarly, here's the Commons category for c:Category:Guildford_Castle, with 5 different listings.

What should be done in such cases?

My initial reaction was to think the items should be merged here. But now I think (with a few exceptions) that that is probably not the right solution. The exceptions I would make, that IMO should be merged would be eg two ends of the same bridge, that happened to be different local authority areas; or places that have both a 'listed building' and a 'scheduled ancient monument' listing. Where these refer to the same place, I think it does make sense to combine them. But for eg Gloucester Cathedral, perhaps it does make sense to have separate items for eg the Cloisters and the Chapter House, since these are distinct identifiable structures within the whole. Perhaps then Cathedral Chapter House (Q17525927) and Cathedral Cloister And Lavatorium (Q17525928) should indeed remain separate from Gloucester Cathedral (Q262500), perhaps linked by part of (P361), with Gloucester Cathedral (Q262500) having the Gloucester Cathedral commonscat link. I'm less sure about no label (Q17525925), which maybe should be merged. Similarly, for Guildford Castle, does it make sense to keep all five separate listed items, such as Remains Of Shell Keep At Guildford Castle (Q17527536), marking them all as part of (P361) Guildford Castle (Q139629), with only the latter getting the Commons category (P373) ?

On the other hand, that would mean that we were divorcing the item with the Wikipedia link and the Commons cat from its historic buildings listings. Is that a good idea? And does this make for too much of a challenge for people to know when to merge items and when not to merge them?

What thoughts do people have? Jheald (talk) 18:18, 13 December 2017 (UTC)

Discussion also noted at Wikidata_talk:WikiProject_Built_heritage#Places_with_multiple_heritage_listings. Feel free to move the thread there if people think it would be more appropriate. Jheald (talk) 18:30, 13 December 2017 (UTC)

I do not quite understand what the problem is. Using the example of the Gloucester sathedral, why can not one have one item about the cathedral (with sitelinks to Wikipedia articles) and five items which list parts of the cathedral which are listed buildings (referenced properly) and possibly having separate Commons categories?--Ymblanter (talk) 18:32, 13 December 2017 (UTC)
I see it exactly like Ymblanter above. Reality is not as hierarchical as we would like it to be. Hence, listed buildings and building parts will alawys be different from the scope of Wikipedia articles and from the scope of Commons Categories, or from the scope of authority data entries. There's not much to be done about it but to try to describe the items as understandable as possible and link them via "part of" and similar properties. --Anvilaquarius (talk) 18:39, 13 December 2017 (UTC)
There are a couple of lighthouses with multiple listings as well. I linked them together with "has part" and that seems to work fine. Looking for images for the items, I found that Commons could use separate category for the foghorn of one of them ..
--- Jura 18:47, 13 December 2017 (UTC)

Thanks, everybody. Here's a query for multiple article-items pointing to a single Commons category, with at least one item having a National Heritage List for England number (P1216): tinyurl.com/ybdakrac. 1275 rows, so there may be some cleaning up to do. Jheald (talk) 22:51, 13 December 2017 (UTC)

Same building, different name / different use?[edit]

Next question: what about when we have the same building complex, re-purposed to have a slightly different use, under a slightly different name: eg

Gloucester Abbey (Q5572012) -> Gloucester Cathedral (Q262500) in 1540.

Is the first a facet of (P1269) of the second? Was it followed by (P156) the second? Property structure replaced by (P167) is one that is not applicable, I think, because the structure was not knocked down and replaced.

Also, what is the appropriate inception (P571) for Gloucester Cathedral (Q262500) ? The date of the first religious structure on the site? The date of the oldest section of the current building? The date on which it was re-purposed to its current use? All of the above, with suitable qualifiers, eg determination method (P459) -> a value representing one of the above, identity of subject in context (P4649) -> (eg) Gloucester Abbey (Q5572012) ? Jheald (talk) 00:32, 14 December 2017 (UTC)

@Jheald: It is always a definition problem: if I take the definition of Gloucester Abbey (Q5572012), this is an abbey (Q160742) which is a special case of monastery (Q44613). What is a monastery (Q44613) ? This an instance of institutional complex, place of worship and organization.
So if we take the definition of institutional complex for monastery, then Gloucester Cathedral (Q262500) is part of Gloucester Abbey (Q5572012), as the church is only a building among others in a monastery. Then for the inception date, the inception date for Gloucester Abbey (Q5572012) (church, monks settlement,...) is the date when the first building was created and for Gloucester Cathedral (Q262500) this is the date when the current building was erected. Snipre (talk) 15:57, 14 December 2017 (UTC)
The WP:en article is not clear as it defines the abbey as being the church, which not true. Snipre (talk) 16:09, 14 December 2017 (UTC)
From my understanding, Gloucester Cathedral (Q262500) was a abbey church (Q334383) between 1058 and 1540, and since 1540 it is a cathedral (Q2977). Snipre (talk) 16:16, 14 December 2017 (UTC)
Thanks @Snipre:. So you would recommend Q262500 be considered an item for the building(s) rather than the institution, and therefore perhaps have:
In general terms, in English I think both the words "cathedral" and "abbey" probably both encompass the entire complex including cloisters, chapter-house, etc; rather than just the church building specifically. So perhaps P31 "abbey" rather than "abbey church" for the earlier period.
What would you recommend for Gloucester Abbey (Q5572012), to indicate its relation to Gloucester Cathedral (Q262500) and its subsequent re-use as a cathedral? Jheald (talk) 17:50, 14 December 2017 (UTC)
@Jheald: I agree with your 2 first points but not with the reuse of inception (P571): the building was erected once, not twice. Use of inception (P571) to describe status change is wrong for me: better use significanr event property with a appropriate event to define the status change. Snipre (talk) 23:41, 14 December 2017 (UTC)
The question is, though, do we use Q262500 to represent "the" building, or the whole sequence of building complexes and associated institutions.
Similarly, should St Paul's Cathedral (Q173882) just be the building designed by Wren, or the institution of a cathedral dedicated to St Paul from AD 604 onwards?
In some ways it makes sense to have a primary item that is as general and encompassing as possible, in both space and time, and then further items that may concretely represent particular parts of the building or specifical constructions in its history, such as Old St Paul's Cathedral (Q2576524).
Or should we keep Q173882 for just the present-day structure? Jheald (talk) 00:34, 15 December 2017 (UTC)
In some ways there are parallels with the question of "my grandfather's axe" (or alternatively the Ship of Theseus (Q1050837)). Can my grandfather's axe still get a single item, if the head has been replaced twice and the handle three times? Jheald (talk) 00:43, 15 December 2017 (UTC)
@Jheald: Again this is the definition of the item which should answer your question, not me: a building can't be an association so you can't mix both concepts in the same item. Then you have to keep a logical structure so if you say that Gloucester Cathedral (Q262500) was once the abbey church of Gloucester Abbey (Q5572012) then you can't consider now the other buildings of the abbey as being elements of the Gloucester Cathedral (Q262500) defined as cathedral. But this las point is more a personal point of view. And for your last question about your "my grandfather's axe", I proposed you to look at any items about a person: does the item refer to the person as a baby, teenager or adult ? Was Albert Einstein (Q937) a physician since he was born or only after he finished his studies ? Snipre (talk) 23:05, 16 December 2017 (UTC)
  • It's hard to say. It seems that initially the cathedral was the church of the abbey. Now it seems to be the opposite: abbey buildings are an annex of the cathedral.
    --- Jura 16:10, 14 December 2017 (UTC)

Two commons links[edit]

We have Albert Memorial (Q281465) and Category:Albert Memorial (Q24704063), apparently just so that we can link to commons:Albert Memorial and commons:Category:Albert Memorial (respectively). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:50, 14 December 2017 (UTC)

What a mess. The two items need to be merged, and the merged item should have both a Commons category (P373) property and a Commons gallery (P935) property. As to whether the category or the gallery gets the sitelink, that's a coin-toss, but I would sitelink the category. Jheald (talk) 00:46, 15 December 2017 (UTC)
The whole system needs to be corrected. Galleries on the commons are almost always just irrelevant. The most important thing about the Commons sitelink is to get from the main Commons entry about an item to Wikidata quickly (and to quickly see if such an entry exists at all), so links from a crappy gallery page that no one ever visits are pretty useless. Jheald's proposal seems sensible, but only if Commons finally starts to display links to Wikidata derived from Commons Category and Commons Gallery property entries, not only from the flawed sitelinks entry. --Anvilaquarius (talk) 09:18, 15 December 2017 (UTC)
For the sitelinks, an alternative would be to create an item "Category:Albert Memorial", connected to Albert Memorial (Q281465) by category's main topic (P301)/topic's main category (P910) in each direction, sitelink that new item to the category, and put c:Template:Interwiki from Wikidata on the top of the Commons category. Both the category-item and the article-item here would also have Commons category (P373) statements. That method is probably how to maximise the interwiki linking from Commons to Wikipedias.
@Anvilaquarius: It doesn't help the multitude of casual browsers of Commons, but if you're doing a lot of work on Commons and Wikidata, you may find the wdcat.js script helpful, which shows when a Commons category is the object of an incoming Commons category (P373) statement on Wikidata. The script gets its information from a SPARQL lookup, so probably can't be scaled for mass usage, which for the moment has to rely on sitelinks and templates, but as an individual user I find it quite useful. Jheald (talk) 10:43, 15 December 2017 (UTC)
User:Matěj Suchánek has gone for a different solution, which is to mark Q24704063 as a Wikimedia permanent duplicated page (Q21286738). But since the Commons interwiki template can't interpret that, and there is no language that has a Wikipedia alternate article, I'm going to change it to instance of (P31) Wikimedia category (Q4167836), as discussed above. Jheald (talk) 12:09, 15 December 2017 (UTC)

Query returning same item several times[edit]

As I am still new to writing queries I am probably missing something, but can someone explain me why this query returns several items more than once. Whether I write just SELECT or SELECT DISTINCT makes no difference. Ahoerstemeier (talk) 12:01, 14 December 2017 (UTC)

Here is a version that opens as list. It seems to be mainly because some parks have multiple coordinates and the query retrieves the same coordinates in two different ways (wdt:P625 and p:P625). Not sure why Q1322875 appears 4 times as well. It disappears if you retrieve just as wdt:P625 [3]
--- Jura 14:14, 14 December 2017 (UTC)

Linking to Draft namespace[edit]

Hello. The topic was raised on the talk page of the Everipedia (Q44960346) item, which is linking to the Draft namespace of the English Wikipedia. It seems difficult to realize that you land in a draft when you come from another Wikipedia following the interwiki link. Have there been any previous discussion on this topic? — Envlh (talk) 19:32, 14 December 2017 (UTC)

I believe we used to not link at all to the Draft namespace, but then ran into problems when people moved pages from Main to Draft and back again (wikidata links, and therefore interlanguage links, were lost). Not sure what the policy is on this here at the moment. ArthurPSmith (talk) 19:38, 14 December 2017 (UTC)
Links to the draft namespace shouldn't exist, and in previous cases the English Wikipedia community was unhappy with them being set. I don't think we have to give EnWiki another reason to dislike Wikidata. Feel free to remove them when you see them. ChristianKl () 19:39, 14 December 2017 (UTC)
Last discussion on this topic was in 2014, Wikidata:Project chat/Archive/2014/05#What do do with Wikipedia:Draft articles. Thoasp (talk) 19:42, 14 December 2017 (UTC)
I'm pretty certain that this wasn't the last discussion. ChristianKl () 20:11, 14 December 2017 (UTC)
So Wikidata:Project chat/Archive/2017/02#Wikidata links to English Wikipedia draftspace? Matěj Suchánek (talk) 20:20, 14 December 2017 (UTC)
That seems more like the most recent one. ChristianKl () 22:59, 14 December 2017 (UTC)
  • Pictogram voting comment.svg Comment In short … Wikipedia does not consider them articles when they are in draft namespace. Draft namespace should simply be considered a more formal sandbox.  — billinghurst sDrewth 23:57, 14 December 2017 (UTC)
  • PetScan list to clean up: petscan:1988016. Matěj Suchánek (talk) 09:45, 15 December 2017 (UTC)
    I cleaned up 33 items from that list and now it's empty. ChristianKl () 15:08, 15 December 2017 (UTC)
Empty for a few minutes, then this. --Larske (talk) 16:07, 15 December 2017 (UTC)

I suggest that we create an abuse filter to restrict adding the draft and other non-notable namespace links on Wikidata items. Stryn (talk) 16:31, 15 December 2017 (UTC)

I Symbol support vote.svg Support such an abuse filter for draft. Which other non-notable namespaces do you have in mind? ChristianKl () 19:28, 15 December 2017 (UTC)
Wikidata:Notability: "To be valid, a link must not be a talk page, MediaWiki page, special page, file, translations page, pages in User or Draft namespace, pages used by LiquidThreads (i.e. pages in Thread and Summary namespace), pages used by Structured Discussions." Stryn (talk) 20:39, 15 December 2017 (UTC)
When it comes to the user/draft namespaces I can easily see how you filter for them and that the filter would be good. In general an item that's notable might have valid links to pages that alone wouldn't grant it notability. Before introducing filters I would like to see for each filter a petscan search that shows what items currently link to that namespace. ChristianKl () 01:01, 16 December 2017 (UTC)
We have had filters for user pages, template subpages and files for four years. Given how common draft articles are becoming, I can quickly make up a new filter for drafts. But I haven't seen much sitelinks to other forbidden namespaces, if any. Matěj Suchánek (talk) 09:44, 16 December 2017 (UTC)
I think such a draft filter would be good in the eternal project of removing reasons for EnWiki people to get annoyed at Wikidata. ChristianKl () 13:20, 16 December 2017 (UTC)
I believe we can restrict links to specific namespaces in the config. I think that would be better than an abuse filter for performance reasons. Anyone up for creating a ticket? --Lydia Pintscher (WMDE) (talk) 18:02, 17 December 2017 (UTC)
✓ phab:T183100  — billinghurst sDrewth 20:53, 17 December 2017 (UTC)

George Cornelius Gorham[edit]

I am attempting to link Q5538125 to wikisource:Author:George Cornelius Gorham, but the save button is greyed out. Would someone be able to give me a hand? Thanks, 142.161.81.20 21:30, 14 December 2017 (UTC)

Seems to be done special:diff/609122838.  — billinghurst sDrewth 00:00, 15 December 2017 (UTC)

British Museum[edit]

Hello. I have the feeling that we might need to change our 3 properties related to the British Museum (Q6373)British Museum person-institution (P1711), British Museum thesaurus ID (P3632) and British Museum place ID (P3633) – because they have apparently changed their formatter URL (P1630) but also possibly the IDs that we store. Are they still working for you? Thierry Caro (talk) 03:11, 15 December 2017 (UTC)

What's up with organizations destorying their links lately? Sjoerd de Bruin (talk) 14:10, 15 December 2017 (UTC)
Please create new properties if they actually changed the identifiers.
--- Jura 20:52, 16 December 2017 (UTC)

Wikimedia Commons links being implemented[edit]

WikidataLovesCommons

There was a previous discussion here in the project chat last month where someone asked if WD:N be expanded to include Wikimedia Commons categories (see this) and there was an overwhelming support, as Wikimedia Commons itself is busy with structuring its data wouldn't it now be time to add a Wikimedia Commons link 🔗 option instead of directing users to “Other sites”? I agree that which Wikimedia Commons categories get linked should be heavily scrutinised, but it should be time to technically bridge 🌉 Wikidata with Wikimedia Commons.

Sent from my Microsoft Lumia 950 XL with Microsoft Windows 10 Mobile 📱. -- 徵國單  (討論 🀄) (方孔錢 💴) 09:41, 15 December 2017 (UTC)


  • Congrats to your new Microsoft Lumia 950 XL. I think we already have links to Commons.
    --- Jura 09:49, 15 December 2017 (UTC)
  • Hooray! There will be a commons_category area to add a link and one for a commons_page so we get back links to Wikidata? --RAN (talk) 17:23, 15 December 2017 (UTC)
I have Q45747062 a case where a Commons category (Category:Journal de Bruxelles nr 76) needs to be added. See also next subject.Smiley.toerist (talk) 14:26, 15 December 2017 (UTC)
✓ Done if you check how I did it, you can do it yourself for the other ones. Mbch331 (talk) 17:32, 15 December 2017 (UTC)

Joining wikidata, the Commons and French Wikimedia[edit]

'Journal de Bruxelles' was revolutionairy Belgian periodical. Two issues from the year 1799 have been digitalized in French Wikisource. It also exists as a Commons category. There is also a article in French Wikimedia about 'Journal de Bruxelles', only the wrong one, as there where several periodicals with the same name. Can this be added to Wikidata? Is Wikisource excluded from Wikidata? (As I see no Wikidata link in Wikisource) As the content contains news, these could be linked to certain subjects such as ships, generals, etc. The links are the following:

Journal de Bruxelles (1790-1800)/76-1799 https://fr.wikisource.org/wiki/Journal_de_Bruxelles_(1790-1800)/76-1799 https://fr.wikisource.org/wiki/Livre:Journal_de_Bruxelles_nr_76_1799 https://commons.wikimedia.org/wiki/Category:Journal_de_Bruxelles_nr_76

Journal de Bruxelles (1790-1800)/83-1799 https://fr.wikisource.org/wiki/Journal_de_Bruxelles_(1790-1800)/83-1799 https://fr.wikisource.org/wiki/Livre:Journal_de_Bruxelles_nr_83_1799 https://commons.wikimedia.org/wiki/Category:Journal_de_Bruxelles_nr_83

General article https://fr.wikipedia.org/wiki/Journal_de_Bruxelles, but as there are several periodicals named 'Journal de Bruxelles' this is the wrong one.Smiley.toerist (talk) 11:18, 15 December 2017 (UTC)


Searching in phabricator[edit]

I was wondering how is the experience of other non-developers when searching in Phabricator? Can you usually find stuff? Have you tried searching for specific bugs? Personally, I haven't done so recently, but it happened that I couldn't even find requests I wrote there myself. I'm asking as some seem to want to use it for more Wikidata editing related stuff.
--- Jura 16:00, 15 December 2017 (UTC)

I can quite easily find the tasks what I want to find. https://phabricator.wikimedia.org/search/query/X10Ih0v0dkQo/#R tells you haven't started any tasks on Phabricator with account "Jura1". Stryn (talk) 16:37, 15 December 2017 (UTC)
@Stryn: This is in part because he goes by Esc3300 on Phabricator. Mahir256 (talk) 23:54, 15 December 2017 (UTC)
So it's doesn't work ;)
Can you find tasks about property suggestions? For some reason keyword search didn't always work for me.
--- Jura 16:49, 15 December 2017 (UTC)
Does this help? https://phabricator.wikimedia.org/search/query/.jWJkAE34cCe/ - PKM (talk) 19:23, 15 December 2017 (UTC)
  • Well, it's a general question about the experience people may have with phabricator for finding things (beyond clicking on a ticket number displayed in a thread).
    --- Jura 20:09, 15 December 2017 (UTC)
You're worried about using it more for editing related things. I think this can be helped by having specific boards for these specific tasks that are easier to scan than the monster Wikidata board. --Lydia Pintscher (WMDE) (talk) 18:05, 17 December 2017 (UTC)

Linkedin[edit]

Property:P2035 was formatted as a full url, is there anyway we can change it to just the ID portion of the url and make it a true identifier? For Jared Robert Banta (Q45753535) we add in https://www.linkedin.com/in/jared-banta-7a369a48. It would be better if we had it as "jared-banta-7a369a48" and made it a true identifier? Any thoughts? --RAN (talk) 17:27, 15 December 2017 (UTC)

  • They used to have a different format, incompatible with that datatype. As we attempt to provide stable properties, I don't think it's good to shuffle it around. On the mere GUI side, I think it could be displayed with others further below.
    --- Jura 17:44, 15 December 2017 (UTC)

Query wikidata start getting messy because of "bad imports" of dates with different precisions and same rank[edit]

When more and more data is added I can see that we get a problem with dates added with different precisions and same ranks ==> you get a non "expected" result when asking queries

Problem creating queries gets very complex if you just would like to have back "one item"

  1. Suggestions:
    1. change the wdt prefix to just return the value with best precision if more has the same rank
    2. train people doing database imports of the consequences of what they do
    3. Histropedia added a new option "Merge duplicated results" link ?????


Example Lars Johan Werle (Q4107873) has more death dates P570 added with different precisions but same rank ==> we get 2 results back

The following query uses these:

  • Properties: family name (P734) View with Reasonator View with SQID, date of death (P570) View with Reasonator View with SQID
    1 SELECT ?item ?itemLabel ?death WHERE {
    2   ?item wdt:P734 wd:Q37507741.
    3   ?item wdt:P570 ?death.
    4   SERVICE wikibase:label {
    5     bd:serviceParam wikibase:language "en","sv" . 
    6   }
    7 }
    
- Salgo60 (talk) 09:15, 16 December 2017 (UTC)
  • If it's a problem at Hist..pedia, you might want to report it there.
    I don't see anything "bad" about the two dates of different precision and referencing, but with equal rank at Q4107873#P570. The next step could be that you add additional references for the more precise dates. Once done, you could set that date to preferred rank.
    --- Jura 09:27, 16 December 2017 (UTC)
No the Histropedia example was just one way of solving it. I guess they have seen the same problem
  1. Mass importing means +10 000 items added then people just jump to the next import it's feels like "touch and go"
    1. they never use the data ==> they dont see the problems....
    2. Example
      1. in Sweden for e.g. we have +7000 items already added with day as precision and rank normal using P3217 as reference ==>
        1. its research done by the Swedish National Archives
        2. its 4 full time very skilled historians quality assuring the result done by 50 professional authors everyone in the field knows about the quality work they do and this organization has done this since 1917 ==>
          1. uploading data with a date with less precision and same rank is wrong 99 times out of 100...
          2. one problem with Wikidata's designed today is that is not easy to understand the quality of P3217
        3. then someone upload another source with less good precision
        4. then all queries with wdt will return 2 results back. Yes you could better queries with grouping but querying Wikidata is already complex dont make it a "science...."
I suggested also a preview mode that could be used before mass import to see what constraints problems you will have see Talk Property constraints portal
- Salgo60 (talk) 10:00, 16 December 2017 (UTC)
  • I still don't get what's should be wrong with the two statements with equal ranks, but different precision/referencing at Q4107873#P570. It's part of Wikidata's design that multiple statements about the same fact can be added. It might even be the key improvement over some Wikipedias.
    --- Jura 10:08, 16 December 2017 (UTC)
You need to differentiate between
  1. how things are stored - its wrong to store two values with the same rank for P570 we have single value property constraints on P570
  2. how to query data - you use wdt to take the truth.... ==> for a date type we get a problem as we can have more values with same rank and different precision --> we have a problem
  3. expected result from the user querying the data - I guess normal is one record per person and his/her death date is expected when using wdt
The simple solution is to ask people to take more responsibility and check before uploading e.g. using Constraint Checker you violate Single_value and get the following "error"

Constraint checker

The reality is that this check is not done by people batch uploading plus maybe also has less good understanding of the quality of data sources used.... houston houston we have a problem
With my query example above using wdt returns all values with same rank I expected one result row which was what I got before mass uploads ==> I can't trust my queries....
Solution: ??? How to handle uploads violating constraints rules??? The last months I have spent maybe 30 hours cleaning dats to avoid this error when people query...
-Salgo60 (talk) 12:09, 16 December 2017 (UTC)
  • The single value constraint on P570 isn't mandatory. Even if it was, it wouldn't override the general principle that multiple different statements about the same fact should be recorded. The solution is that either you need to evaluate the different sources and determine which one should get preferred rank or refine your query.
    --- Jura 12:39, 16 December 2017 (UTC)
    And one of my suggestions is that wdt should just return the result with best rank and best precision not more values....
    evaluate the different sources and do cleaning after other people before every query returning 10000 records ?!?!? Serious? - Salgo60 (talk) 12:54, 16 December 2017 (UTC)
    (edit conflict) Best precision does not imply best rank. There may be a date with precision to days but may be totally wrong. This is where ranks come to the scene. Matěj Suchánek (talk) 13:00, 16 December 2017 (UTC)
    • If you are indifferent to the absence of a reference at Q4107873#P570 (value with day precision currently has a Wikipedia import source only), you could query the most precise value only.
      --- Jura 13:11, 16 December 2017 (UTC)
  • I have no problem understanding that I also can change all queries for me this is an "architectural problem" having prefixes like wdt you try to implement a concept that is easier for the person writing the query ==> returned is just the best ranked value. With the datatype "datetime" combined with "mass import" of sources with different precision we can see that we dont get just one value instead we get one value for every precision added..... In my book of design this is an user case not taken care of that will have a ”high" cost for people to write correct queries. My example with Histropedia was to point out that others have maybe already seen this problem and try to solve it in the query of Histropedia interface. For me that is the wrong place to correct it... User:ChristianKl I think points in the right direction to make it easier to write correct queries. Then the people doing mass import also need better feedback before starting a import that violates a lot of rules... Wikidata is growing and fixing everything by asking the users to write "better" more advanced queries will end in a "product" less usable - Salgo60 (talk) 16:24, 16 December 2017 (UTC)
  • There's no rule that it's mandatory to add a rank. Nonmandatory constraints suggests that it would be better if a person would judge which item should get preferred status but it doesn't say that a person who enters data violates our rules when they don't tackle the task of ranking. ChristianKl () 17:00, 16 December 2017 (UTC)
@ChristianKl: I think this is another flavour of the problem
  • that we dont "rank"/"define quality" of sources.
  • that we dont communicate to the Wikipedia reader if the fact s based on an authoritative source or just a community created source
Lesson learned is most people in Wikidata dont have the skills/knowledge to understand every source quality e.g. I requested a year ago Property P3217 that is from the Swedish National archive(about SBL). They have 4 fulltime people just quality assure the 100 biographies they produce every year with the help of about 50 authors. When speaking to University teachers in history they tell me if students havnt checked with SBL they have to go back and do the homework....
In Wikidata/Wikipedia most people have no understanding of this quality ==> you start upload birth dates with precision year and "degenerate" Wikidata and they get no warning that we already have a value from a high quality source.....
I am convinced that both for people importing data and people reading Wikipedia with data "transferred" from Wikidata should get some indication of the quality and correctness of the information. On the session State of the project I asked that question (see 42 min video) and the answer was Trust is important and that is nothing yu get for free and it mus also be communicated to the readers of Wikipedia that this is data you can trust and not that if you are lucky no one has destroyed the quality by importing some lower quality data... - Salgo60 (talk) 20:17, 16 December 2017 (UTC)
You confuse multiple different issues:
(A) We show multiple values with wdt when we have multiple values stored. As far as trust goes, offering multiple values when we don't know which value is the best one is good. A reader who cares about which value is correct can look at the values we have with the corresponding sources and make his own judgement about which source is best. It's also great when he shares his research by setting appropriate ranks. Above you recommend that only one values is shown even when we have no reason to believe that it's better than the other values we store.
(B) The extend to which data from the Swedish National Archive should get preferred status. If you believe that it generally should get preferred status, run a bot that gives all the entries that are soruces from it preferred status. Of course, at best you would also check for cases where that means that there are claims with preferred to see what's up with those cases.
(C) When reading Wikipedia a reader isn't supposed to trust everything but check sources. Normally, there's no requirement in Wikipedia to source every claim. When Wikipedia imports Wikidata every imported claim has a source that a reader can judge for themselves and that's an improvement over the status quo of Wikipedia.
(D) It would be great to have additional ways to judge source quality. Signed statements will be one way. We might also have other ways in the future to automatically judge source quality but that isn't an easy task. ChristianKl () 21:08, 16 December 2017 (UTC)
  • Like that. Wikidata will never block imports just because the data are different from what we have had. (This reflects the diversity of knowledge available...) Matěj Suchánek (talk) 12:42, 16 December 2017 (UTC)
    I have never spoken about blocking we speak about a problem with more date values with same rank and different precision and that wdt returns both values plus people mass uploading dont rank - Salgo60 (talk) 12:54, 16 December 2017 (UTC)
    Have not? Am I misreading the topic's title? Matěj Suchánek (talk) 13:00, 16 December 2017 (UTC)
Lars Johan Werle (Q4107873) is a nice example. If you browse it you see that 23 June 1926 currently only has a week reference from Russian Wikipedia and three good references for 1926. The Primary source tool provides a link to a stronger reference. Encouraging the browsing of the item to actually look at the reference allows the data users to see the Primary source reference and add it (I did that in this case). Giving to the well developed Primary Sources tool it now has three entries ;)
When doing a mass upload, adding data as "normal rank" is correct behavior. ChristianKl () 13:02, 16 December 2017 (UTC)
@ChristianKl: maybe time to redesign what correct behavior is.....Someone has to pay the price and right now its user querying who cant trust the result which is a very high price = in the end can we trust Wikidata?
Larske (talkcontribslogs) helped me with a query ==> result its more messy than I thought.... and nearly impossible for me to check every record... feels better to have the data outside Wikidata...
  • The SBL dataset P3217 has as I said very very few errors P3217 (most errors I guess is me uploading it to Wikidata.... ;-)
    • The total dataset is 7756 records and with this query we just check birth dates and can see 16% has more dates with different precisions mostly because new uploads the last months....
  • Lesson learned
      1.  ????
      2. Adding data to Wikidata and you wish to maintain the quality then you have a maintenance H#!?!?!?!ll ;-)
        1. I guess the upload rate will increase 10 times just the next year.....
  • Houston houston Wikidata has a problem.....
- Salgo60 (talk) 20:17, 16 December 2017 (UTC)
If you believe that data is that great, why don't you put it into the preferred rank? ChristianKl () 21:10, 16 December 2017 (UTC)
  • The best solution I see would be to add a new clause to the SPARQL interface that actually gives you only one answer. ChristianKl () 13:04, 16 December 2017 (UTC)
  • In this case we don't want to discourage a person from adding the data. It's not the responsibilty of the person who adds the data to go to every example and decide what should have the preferred rank. ChristianKl () 16:46, 16 December 2017 (UTC)
  • OK do you oppose having constraint checks for mass upload also? - Salgo60 (talk) 12:41, 17 December 2017 (UTC)


New clause? We already have that, it's called SAMPLE(). See Wikidata:WikiProject sum of all paintings/Creator missing collection authority control on how to use it. Multichill (talk) 09:22, 17 December 2017 (UTC)
How do we motivate that wdt just returns one value for other data types but for dates with different precisions in the database we get more...
SAMPLE is one way and then you need start grouping etc... --> queries get more complex. For me a good design is return the date with best precision for best rank and not more rows - Salgo60 (talk) 12:29, 17 December 2017 (UTC)
I would like to have the values returned with the most reliable source. This is not necessarily the value with the best precision. What is the adavantage of having precise but probably wrong values compared to having less precise but reliable values? --Pasleim (talk) 12:59, 17 December 2017 (UTC)
But ranking with the above logic is the responsibility of the person doing the query ==> you will often have entries with
  • same rank
  • different precision
today you get every entry with the same rank and all precisions back in theory you could get for Napoleon (Q517) birth the following returned
11 - day: 15 augusti 1769
10 - month: augusti 1769
9 - year: 1769
8 - decade: 1760
7 - century: 1700
6 - millennium: 1000
....
and then what do we do with different calendars ;-)
my suggestion is if everyone has the same rank pick best precision
- Salgo60 (talk) 13:14, 17 December 2017 (UTC)
wdt will always give multiple values when there are multiple values in the database for a given statement and there's no ranking information that gives information about which value happens to be best.
If cases where the user cares about data quality that's the correct behavior because it allows the user to decide for themselves which value they trust the most. In your case you are saying that data quality isn't the priority but you only want one value in every case because you are looking at too much data. That might be valid for your use-case but there's no reason for making it the default. Showing multiple values when there are multiple values on the same rank also encourages the user to actually make a judgement about which value is best and that's helpful for Wikidata.
In the case you linked above the most precise value was at the time you linked imported from the Russian Wikipedia. We consider that to be a source that's less reliable than Swedish Musical Heritage/Nationalencyklopedin/Musikverkets auktoritetsdatabas which existed as sources for the less precise claim. If a human would set a preferred rank that rank shouldn't be set to the Russian Wikipedia value. ChristianKl () 13:32, 17 December 2017 (UTC)
The user case is: As an Wikidata user writing queries using fields with datetime I would like to have a simple way retrieving dates that is not fragile because of mass upload with data with odd precisions.
The key question I feel is
  1. Napoleon is born in the 1700 century
  2. Napoleon is born 15 aug 1769
Does those statements show 2 different values or one value presented with two different precisions. I vote for one value...
If you comoare those statements with this article then you also have another value 1768 for his birth
We will have the same "problem" with coordinates if people start uploading more precise or less precise coordinates...-- Salgo60 (talk) 21:13, 17 December 2017 (UTC)

Facto Post – Issue 7[edit]

The latest issue of Facto Post is available via this link. Charles Matthews (talk) 13:28, 16 December 2017 (UTC)

Help with “Criterion used”[edit]

I’d like confirmation that I’m using criterion used (P1013) correctly. I want to say that if spinning jenny (Q49015) and water frame (Q1751504) are <subclass of> textile working tools/equipment, as Getty AAT says, then so are spinning frame (Q3274449) and spinning mule (Q1413723) by analogy. Are the qualifier and reference on the subclass statement at spinning mule (Q1413723) the right way to do this? - PKM (talk) 00:52, 17 December 2017 (UTC)

I don't feel that it's clear what information you want to express in that item. ChristianKl () 19:39, 17 December 2017 (UTC)

Polymath Virtual Library ID (Q44280387)[edit]

Is this a property that created wrongly in the Main namespace? --Liuxinyu970226 (talk) 01:41, 17 December 2017 (UTC)

OpenStreetMap Node (Q42375175) seems likely. --Liuxinyu970226 (talk) 01:48, 17 December 2017 (UTC)
I ping Gnjo55 (talkcontribslogs) and Martinbayo (talkcontribslogs) to know their opinion. Pamputt (talk) 10:38, 17 December 2017 (UTC)

What is the difference between motorized bicycle (Q1377570) and moped (Q201783)?[edit]

Aren't both of them simply bike with motor? C933103 (talk) 08:45, 17 December 2017 (UTC)

Not in Germany, there's a huge difference: One is a bicycle with an attached motor (and it has to have pedals, so it can be used without the motor), the other is a small motorbike with 50cm³. motorized bicycle (Q1377570) don't require a driving licence (or any Street-legal vehicle (Q228760) or de:Hauptuntersuchung, moped (Q201783) require all of this. Sänger (talk) 09:39, 17 December 2017 (UTC)
Same in UK, motorized bicycle (Q1377570) and moped (Q201783) are different things. JerryL2017 (talk) 09:53, 17 December 2017 (UTC)
So if I understand correctly, if you put a pair of pedal onto moped (Q201783), then it would become motorized bicycle (Q1377570), and then you would no longer need driving license?
And what would be the difference between them in places where they fall into same regulatory category?
C933103 (talk) 11:36, 17 December 2017 (UTC)
So I'm not sure about regulations in different countries, I guess they will vary. Conceptually though I would say the difference is that a motorized bicycle (Q1377570) can be ridden using the pedals even if the motor isn't running. The motor assists the rider but isn't required. Whereas a moped (Q201783) will always require the motor, even with pedals attached it wouldn't really be possible to propel the vehicle any distance without the motor driving the wheels. JerryL2017 (talk) 12:55, 17 December 2017 (UTC)

Issue with search function?[edit]

I have noticed an issue with some recent items which return "no match found" in the "Search Wikidata" box - using both their label and their Q-ID - although they clearly exist. This also makes it impossible to set some property's value to them. Examples are Pierre Fouchet (Q45825730), no label (Q45825741), district mayor of Tiergarten (Q45825742) or no label (Q45825750). The affected items are not continuous in ID (e.g. Erick Guerrero (Q45825749) works fine). The issue persists both when switching browsers and when logging out.--Pfadintegral (talk) 11:14, 17 December 2017 (UTC)

I can replicate the issue. ChristianKl () 12:33, 17 December 2017 (UTC)

Precision of date[edit]

If I create a statement with date, for example publication date (P577) with a value of 2015, precision year, it shows 2015. But if I create a query, it shows for the same value Jan 1, 2015. Did I create the date wrong or the query service doesn't understand the precision? Or I should make a difference query? :) Samat (talk) 13:47, 17 December 2017 (UTC)

How about providing an example of the item you are talking about and the query you are making? ChristianKl () 13:52, 17 December 2017 (UTC)
You are right. Example query. Samat (talk) 14:10, 17 December 2017 (UTC)
  • ?item wdt:P585 ?date yields 2015-01-01T00:00:00Z without any further information about precision
  • ?item p:P585/psv:P585 [ wikibase:timeValue ?time; wikibase:timePrecision ?precision; wikibase:timeTimezone ?timezone; wikibase:timeCalendarModel ?calendarModel ] . contains full details. The date bound to ?time still has the format 2015-01-01T00:00:00Z as before, but the ?precision variable with value 9 now tells you that this is to be interpreted with year-precision. Doc is at mw:Wikibase/Indexing/RDF Dump Format#Predicates for Time.

MisterSynergy (talk) 14:07, 17 December 2017 (UTC)

Thank you! I think I will need some time now to understand what does this mean for the example query (where some data have year, some month and some day precision) :) Samat (talk) 14:15, 17 December 2017 (UTC)

target for merge of non-human person (Q18170708)?[edit]

This "non-human person" at ca/esWPs would seem to have a match somewhere in our plethora of persons. At the moment it is sitting without items and needs a home or a merge IMO.  — billinghurst sDrewth 14:57, 17 December 2017 (UTC)

non-human (Q13023682)? --Fractaler (talk) 15:48, 17 December 2017 (UTC)
Also, non-human electoral candidate (Q2740853), for example, is "non-human person"? --Fractaler (talk) 15:55, 17 December 2017 (UTC)
What do they mean with "person" in ca/es? Do they count corporations as persons? ChristianKl () 16:19, 17 December 2017 (UTC)
Note: Users who edited Q18170708: @Climent Sostres, Andreasmperu: ^^ --Liuxinyu970226 (talk) 04:18, 18 December 2017 (UTC)

Structured descriptions[edit]

Have WMF/developers ever thought of using structured Wiktionary items (when they exist), or internal Wikidata items, to generate Wikidata descriptions, instead of having the same generic "American actress"/"Wikimedia template"-type descriptions added and (mis-)translated manually over millions of items? (There is currently an English Wikipedia RfC related to this topic, which I have not participated in.) Jc86035 (talk) 14:58, 17 December 2017 (UTC)

This is the community board, for WMF/devs you’d have to ask at Wikidata:Contact the development team. Anyway, what is the problem with the current setting? Wikidata descriptions work well to my opinion and there is no need for changes. —MisterSynergy (talk) 15:31, 17 December 2017 (UTC)
What to do with Wikitonary items is a debate we can have in a year. ChristianKl () 15:37, 17 December 2017 (UTC)
@MisterSynergy: If they worked well enough the WMF wouldn't have needed to disable them on the English Wikipedia.
Aside from the multilingual vandalism/test edits (difficult to find and tedious to correct), descriptions are mostly translated manually and so they may never get added/updated or are mistranslated, descriptions have to be added and updated manually for each of e.g. the three different varieties of English and the more than ten different variations of Chinese on every item, and frequently users just copy the name into the description for some inexplicable reason. There are probably other issues.
It would also be much more efficient, for example, to have a bot add two lexemes(?) which mean "Wikimedia template" to every item for a template, and have real people add translations to the lexemes "Wikimedia" and "template", instead of the current situation where bots go around slowly adding these translations to every single item. Similarly for every combination of demonymgenderoccupationgender and so on (assuming there would be some sort of internal mechanism to indicate grammatical rules for word order etc. for each language).
There's not much point in asking the development team if no one else thinks it's a good idea, and quite often WMF employees will read and comment on this page anyway.
@ChristianKl: There's no harm in thinking about it now, especially to ask if this sort of thing (based on Wiktionary items, not Wikidata text fields) is feasible since it would probably require a sort of rudimentary translation mechanism for at least adjectives and nouns (assuming lexemes shouldn't be created for every combination of descriptions, anyway). Jc86035 (talk) 15:52, 17 December 2017 (UTC)
  • Descriptions were originally there to disambiguate items with identical labels in Wikidata. It seems counterproductive to do this with a highly unifying approach such as the one suggested by you. In some cases it would even be impossible, thus I like the flexibility of individual descriptions even though it is somewhat expensive to manage them.
  • It was much later that someone (at WMF?) realized that Wikidata descriptions are useful for other purposes as well, e.g. to disambiguate search results in mobile apps and so on.
  • The enwiki community fell for a couple of aggressive anti-Wikidata lobbyists, and now they have maneuvered themselves into a very difficult situation with only bad solutions left for implementation. IMO their decision does not really have anything to do with description quality or style.
  • Some tools provide automatic descriptions based on statements in the item (reasonator, User:Yair_rand/WikidataInfo.js, and so on). These descriptions do indeed describe the entities, but they are not at all unique in combination with the label and can’t fully replace the “description” field.
MisterSynergy (talk) 16:08, 17 December 2017 (UTC)
en:MediaWiki:Wdsearch-autodesc.js can't replace the field in every case but at the moment for persons it almost always outplays fixed descriptions. The bigger part of the fixed descriptions also come from bots in a messy way (regarding dates of birth and death and thus hardly improvable by bot again. I think descriptions could benefit from a more systematic proceeding instead of the uncoordinated lonely bot operator actions. They are better than nothing, but much worse than en:MediaWiki:Wdsearch-autodesc.js and this isn't necessary. --Marsupium (talk) 16:32, 17 December 2017 (UTC)
One of the core ideas of these descriptions being shown in the App is that someone at the WMF thought that mobile readers who might not want to edit long Wikipedia pages might be willing to be motivated to add short descriptions texts. This was an idea to turn people who don't edit at the moment into editors who would hopefully edit more as they see how great it is to edit.
Wikidata building a fancy system of automatically created descriptions wouldn't help with that strategic goal. Better descriptions would be nice, but I don't see the strategic value in spending a lot of Wikidata development resources on it. ChristianKl () 02:43, 18 December 2017 (UTC)

Subitem Volume by periodical items[edit]

Consider Q45747062 now classified as a periodical. However this is a specific issue (6 december 1799) and many similar items will follow (9 issues, see Commons:Category:Journal de Bruxelles) Two issues are digitalized in fr:Wikisource. At the higher level Journal de Bruxelles (1790-1800))) is a separate item. To confuse matters there have several periodicals with the same name. The is about another periodical started in 1820. The two level distinction is made in the Wikisources so Wikidata needs to do similarly.

It seems to me that items about periodicals need to be ordened into 'periodicals' and 'volume/issue of periodicals' for the lower level. A better name can be found in English and other languages (In en:Wikisource 'Volume' is often used).Smiley.toerist (talk) 19:36, 17 December 2017 (UTC)

Scientific names[edit]

Why is scientific name, published under a Code of nomenclature (Q23729754) marked as an instance of nomenclature Codes (Q2673092)? And the latter, as an instance of biological nomenclature (Q522190)? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:16, 17 December 2017 (UTC)

It isn't. --Succu (talk) 23:14, 17 December 2017 (UTC)
What isn't? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 00:17, 18 December 2017 (UTC)
@Pigsonthewing: The former of your concerns has not been true since September. @Brya:, as the last person to introduce the statements of concern. Mahir256 (talk) 02:55, 18 December 2017 (UTC)
As Succu points out, it is not, although really it should be. The relationship scientific name, published under a Code of nomenclature (Q23729754) as an instance of nomenclature Codes (Q2673092) seems fine: it is informative. The present relationship "subclass of" "name" is quite uninformative. - Brya (talk) 03:21, 18 December 2017 (UTC)
name (Q82799) is a homonym. scientific name, published under a Code of nomenclature (Q23729754) is the scientific name (scientific word (Q999234)?) of the object. --Fractaler (talk) 08:10, 18 December 2017 (UTC)

Brazil Weather Data[edit]

Greetings!

My name is Caio Silva, I'm a Information Systems student in Universidade Federal Fluminense, Rio de Janeiro, Brazil. I would like to upload some data about the weather in Brazil from a weather institute. They collect data everyday in every state of Brazil. I mailed them asking for permition to upload their data. IF, they let me use their data, I would like to upload them here, to be available for everyone using wikidata and wikipedia. Can you help me do it? I don't really know how to procede.

Thanks!

Sincerely, Caio Silva

@Caiosilva: Is there any plan to use that data ? Daily data is too big for WD which doesn't have the structure to store that kind of data. Snipre (talk) 08:24, 18 December 2017 (UTC)