Wikidata

Wikidata:Project chat/Archive/2017/11

< Wikidata:Project chat‎ | Archive‎ | 2017
This page is an archive. Please do not modify it. Use the current page, even to continue an old discussion.

SPARQL on pywikibot

I seem unable to fetch SPARQL results using pywikibot on PAWS. I'm following the instructions on Wikidata:Pywikibot - Python 3 Tutorial/Iterate over a SPARQL query and got a "WARNING: Http response status 400". (Yes I have put the correct query file in the correct directory.) Thoughts? Deryck Chan (talk) 13:44, 2 November 2017 (UTC)

Your query either is invalid, or times out? What query did you try to run? Matěj Suchánek (talk) 16:12, 2 November 2017 (UTC)
Matěj Suchánek: The query on Wikidata:Pywikibot - Python 3 Tutorial/Iterate over a SPARQL query, as I've written above. Deryck Chan (talk) 17:37, 2 November 2017 (UTC)
Fixed it myself. It seems that either SPARQL or pywikibot or PAWS has changed, such that replacing '\n' with ' ' now breaks the query. Removing that seems to work. Deryck Chan (talk) 21:46, 2 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 08:20, 3 November 2017 (UTC)

Copy labels?

Does someone have a trick to copy the non-EN labels (only) from Amadis de Gaula (Q452075) to Amadis of Gaul (Q42702348)? - PKM (talk) 04:17, 5 November 2017 (UTC)

I can do this in the command line. But I'm not convinced all non-English labels should be moved. Matěj Suchánek (talk) 09:27, 5 November 2017 (UTC)
I guess I'll copy by hand then. - PKM (talk) 19:23, 6 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 18:29, 7 November 2017 (UTC)

Inception vs. date of official opening

Syced is mass-adding inception (P571) to railway stations (example), sometimes to items which already have the opening date in date of official opening (P1619). I think it's incorrect to use inception (P571), since a station can exist before it opens. Should these be changed to date of official opening (P1619)? Jc86035 (talk) 05:09, 6 November 2017 (UTC)

Sorry! The multiples samples I had checked (mostly Japanese stations) did not have this property. I actually did not know about that property :-/ I will revert my additions. Syced (talk) 05:33, 6 November 2017 (UTC)
This QuickStatements v2 command reverts my mistaken edits. I will run it as soon as I can. Thanks for letting know about this, and sorry for the overlook! Syced (talk) 06:04, 6 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 18:29, 7 November 2017 (UTC)

Best practices for naming women

I assume the one for Hillary Clinton (Q6294) is the template for best practices, she is easy because we can do a Google search to find the most common combinations used for her "married name". I will use her entry as an example. I see a mix of naming conventions for other women both in the titles of Wikidata entries and the "family name" field (some use their birth surname, some their married surname, some both) and "married name" field (some use Firstname_maidenname_marriedsurname such as "Hillary Rodham Clinton" or Firstname_middlename_marriedsurname such as "Hillary Diane Clinton" and some just use marriedsurname "Clinton"). Do we use a standard combination for historical people where there will be little to no documentation for common usage? --Richard Arthur Norton (1958- ) (talk) 19:17, 31 October 2017 (UTC)

It's a challenge in that the "standard" varies by culture/period. I personally think "family name" should be the surnames name of the birth or adoption (i.e. maiden) family name, and married name should be the full name as used by the subject where known. But any form that is documented in a source should be recorded. - PKM (talk) 19:12, 1 November 2017 (UTC)

Notice that birth name (P1477) has datatype "monolingual text", while family name (P734) has datatype "item". Deryck Chan (talk) 21:36, 1 November 2017 (UTC)

Why are links to Commons do not show up as links?

Tracked in Phabricator
Task T177698

Items like Robert Völcker (Q42195291) or Martinus van Andringa (Q41269933) have Commons Creator page (P1472) property which link them to Commons. However for me the page name to Commons is displayed as text not link. Usually reloading the page helps with such issues, but not for those links. What can be done to fix them? --Jarekt (talk) 12:54, 1 November 2017 (UTC)

Seems like the AuthorityControl gadget is broken, not only for Commons Creator page (P1472) but also for Commons gallery (P935) and Commons category (P373) --Pasleim (talk) 13:13, 1 November 2017 (UTC)
I do not know how links to commons pages are displayed, but I doubt it is related to AuthorityControl gadget. Unless I am really confused here. --Jarekt (talk) 13:22, 1 November 2017 (UTC)
See phab:T177698, it's pretty unclear what the cause is. Sjoerd de Bruin (talk) 13:24, 1 November 2017 (UTC)
Maybe related: for quite a while, coordinate location (P625) claims are sometimes, but not always linked to geohack. —MisterSynergy (talk) 13:29, 1 November 2017 (UTC)
@Jarekt: purging the page worked for me.  — billinghurst sDrewth 22:22, 1 November 2017 (UTC)
both seemed to have KrBot come in and remove an underscore from the name. Maybe see if other replacements have the same issue.  — billinghurst sDrewth 22:24, 1 November 2017 (UTC)

Participate in Dispute Resolution Focus Group

The Harvard Negotiation & Mediation Clinical Program is working with the Wikimedia Foundation to help communities develop tools to resolve disputes. You are invited to participate in a focus group aimed at identifying needs and developing possible solutions through collaborative design thinking.

If you are interested in participating, please add your name to the signup list on the Meta-Wiki page.

Thank you for giving us the opportunity to learn from the Wikimedia community. We value all of your opinions and look forward to hearing from you. JosephNegotiation (talk) 22:38, 1 November 2017 (UTC)

specific birth/death/event location — what to use other than P969

In obituaries, they often mention specific places of death (street address or house name), for which I have been utilising P969 (P969) to record that factoid. However, I see that this a constraint reserves that for a static object rather than event. That being the case, what is the preferred means for recording the address of an event (birth, death, whatever) where the event is notable, the detail relevant, though the location needs to be recorded in a text field. Thanks for the advice.  — billinghurst sDrewth 23:59, 1 November 2017 (UTC)

Roman numeral system, Roman number, symbol/digit in Roman number

We have a I, V, X, L, C, D or M symbol/digit. Each of them is the letter of the alphabet (heptad (Q29441572) of I, V, X, L, C, D and M), used in combinations to represent a words - Roman numbers (I, X, combinations IX, XI, etc.). This words/combinations is part of the language mathematical system - Roman numeral system --Fractaler (talk) 15:57, 22 October 2017 (UTC)

Roman numerals (Q38918) (numbers in the Roman numeral system) we have. --Fractaler (talk) 17:16, 29 October 2017 (UTC)
Now with these terms we have homonymy (Q21701659) (=chaos): no alphabet (Q837518), no Q42558923, no word (Q42559432) refinement. Is the chaos a good thing? --Fractaler (talk) 14:12, 2 November 2017 (UTC)

Superset poem (Q5185279) or verse (Q1661)?

For set monostich (Q1503358): superset is poem (Q5185279) or verse (Q1661)? --Fractaler (talk) 13:05, 29 October 2017 (UTC)

It's poem (Q5185279). -Valentina.Anitnelav (talk) 20:30, 29 October 2017 (UTC)
monostich (Q1503358) is not a single metrical line in a poetic composition (superset is literary work (Q7725634))? --Fractaler (talk) 08:18, 30 October 2017 (UTC)
This applies at least to the concept described in the German and English wikipedia article and in the English and Spanish description of the item (there are no other descriptions yet). In the French wikipedia article "monostique" seems to denote a type of strophe (Q207804). (I don't know for the other languages - maybe some of the terms denote a type of verse (Q1661))- Valentina.Anitnelav (talk) 09:30, 30 October 2017 (UTC)
Sorry, I did not understand. (In the English and Spanish description) monostich (Q1503358) is a single metrical line in a poetic composition (superset is literary work (Q7725634)) or is not a single metrical line in a poetic composition (is not superset is literary work (Q7725634))? --Fractaler (talk) 09:45, 30 October 2017 (UTC)
According to the English and Spanish description ("poem which consists of a single line"/"poema que consiste en una sola línea") it is not a single metrical line in a poetic composition, but a type of poem that is distinguished by having one line. monostich (Q1503358) is described as a type (or subclass) of poem (Q5185279), and thus also of literary work (Q7725634) (poem (Q5185279) is a subclass of literary work (Q7725634)). - Valentina.Anitnelav (talk) 10:26, 30 October 2017 (UTC)
Then couplet (Q15922520) is not next object? --Fractaler (talk) 13:20, 30 October 2017 (UTC)
I'm generally sceptical if followed by (P156)/follows (P155) are appropriate to express the relationship between monostich (Q1503358), couplet (Q15922520), tristich (Q7844186), etc (what should be the series?). But actually I'm also doubtful if couplet (Q15922520) should be considered a subclass of verse (Q1661) (there seems to be some ambiguity with these items, both across languages and within the same language). - Valentina.Anitnelav (talk) 15:52, 30 October 2017 (UTC)
w:Poetry: Lines of poems are often organized into stanzas, which are denominated by the number of lines included. Thus a collection of two lines is a couplet (or distich), three lines a triplet (or tercet), four lines a quatrain, and so on. These lines may or may not relate to each other by rhyme or rhythm. For example, a couplet may be two lines with identical meters which rhyme or two lines held together by a common meter alone --Fractaler (talk) 08:10, 31 October 2017 (UTC)
So according to this paragraph couplet (Q15922520), tristich (Q7844186) etc. would be a subclass of stanza (Q1756348). -Valentina.Anitnelav (talk) 10:32, 31 October 2017 (UTC)
monostich (Q1503358) also would be a subset of the set stanza (Q1756348)? --Fractaler (talk) 05:51, 1 November 2017 (UTC)
This seems to be a borderline case, but it should be ok.- Valentina.Anitnelav (talk) 07:21, 1 November 2017 (UTC)
Thank you for your discussion! --Fractaler (talk) 07:44, 1 November 2017 (UTC)

poem (Q5185279). Because verse (Q1661) is characterized by rhyme (Q178715) (see sitelinks) but monostich (Q1503358) is a single line so generally isn't characterized by rhyming. Deryck Chan (talk) 10:13, 2 November 2017 (UTC)

Why can not there be a zero rhyme (when rhyme (Q178715) = 0)? And how about stanza (Q1756348) (as superset)? --Fractaler (talk) 14:02, 2 November 2017 (UTC)

seamanship (Q351363) vs marine navigation (Q3337280)

Andreasmperu and I have been in a very slow edit dispute about this question: should we merge seamanship (Q351363) and marine navigation (Q3337280)? Comments welcome. Deryck Chan (talk) 19:29, 31 October 2017 (UTC)

The two seem to be different concepts to me, so I see no reason to merge them. ChristianKl (talk) 19:36, 31 October 2017 (UTC)
@ChristianKl: To me, both are just "the skill to drive a boat to a destination correctly", with cultural slants according to different languages. They have zero overlapping interwikis. It seems that the interwikis have been arbitrarily separated into two sets by title depending on whether they literally translate to "boat / seafarer skill" or "sea navigation". Deryck Chan (talk) 21:15, 1 November 2017 (UTC)
Interwiki's are not the primary concern when it comes to distinguishing Wikidata items. An activity like using an anchor would be in my view part of "seamanship" but not part of "marine navigation". The English article is quite clear that ropework is part of seamanship and it clearly isn't in marine navigation. ChristianKl (talk) 21:28, 1 November 2017 (UTC)

What is navigation?

In first, what is navigation (Q102066)? process of monitoring and controlling the movement of a person, craft or vehicle from one place to another? skill (Q205961) is not process (Q3249551), but is result (Q2995644). --Fractaler (talk) 06:46, 1 November 2017 (UTC)
Isn't it both? You need skill to get to the correct destination, but as you practise navigation you also acquire skill to do it. Deryck Chan (talk) 21:17, 1 November 2017 (UTC)
"practise navigation" = "navigation"? Can we add "navigation", also "practice of navigation"? Fractaler (talk) 06:09, 2 November 2017 (UTC)
I meant "practise" in the sense of "a practitioner does something" e.g. "a doctor practises medicine", not in the sense of "a student practises for his exam" or "Nadal is practising for his next competition". Anyway, unless there are Wikipedias that make that fine distinction, I don't see a need to distinguish "navigation" vs "practice of navigation". Deryck Chan (talk) 13:20, 2 November 2017 (UTC)
Wikipedia is an outdated tool (atavism, rudiment) of modeling the world, which focuses only on humans (unlike the Wikidata). Ie, the person from the Wikipedia article will understand the difference between the terms navigation (skill) and navigation (process). The program needs to be explained by its tool/language (which Wikipedia does not have, and so that the Wikidata appeared).Ok, you have "skill to get to the correct destination" (=it is the ability to navigate). "ability to navigate=navigate? And when you as "a practitioner does navigation" it is the implementation of navigation possible, "process" or "ability"? In addition to a person, anyone/anything else can navigate? --Fractaler (talk) 13:46, 2 November 2017 (UTC)

xR Extended Relationships

Do we have authority control for "xR Extended Relationships" - VIAF (309520354)? Geagea (talk) 08:02, 1 November 2017 (UTC)

@Geagea: That seems like some kind of VIAF placeholder or typo or something. The ID is "XR-VIAFWORKLCn2008059237", which is just "n2008059237" prefixed with "XR-VIAFWORKLC". Jon Harald Søby (talk) 12:44, 1 November 2017 (UTC)
Jon Harald Søby, thanks. We already have the LCCN identifier in On verso of t.p. (Q6595310). Geagea (talk) 17:36, 2 November 2017 (UTC)

Age at death known

Hello, Charles Anatole Maingonnat (Q42555344) died the 21st Feb 1891, aged 54, i.e. he was born between 1836 and 1837. How should I deal with date of birth (P569)? I used "1830s", then precised earliest date (P1319) and latest date (P1326), though it seems utterly complicated. Thanks for your help, Totodu74 (talk) 10:20, 2 November 2017 (UTC)

I think you are doing the correct approach, though. Sjoerd de Bruin (talk) 10:29, 2 November 2017 (UTC)
You could use sourcing circumstances (P1480) with circa (Q5727902), and say 1836, then put the latest date as the extension.  — billinghurst sDrewth 11:13, 2 November 2017 (UTC)
Oh and with the death please qualify with age of subject at event (P3629)  — billinghurst sDrewth 11:14, 2 November 2017 (UTC)

Redirects

Hi there. Most pages here actually are or will soon be essentially redirects to a page with the same title, on a different project (mediawiki.org). Even when content is there, it's either not important (i.e., a logo), or the page is actually historical and not meant to be used. I am just flagging in case this needs consideration. Thank you. --Elitre (WMF) (talk) 11:02, 2 November 2017 (UTC)

@Elitre (WMF): When all of them have been deprecated and you feel that this item no longer serves any useful purpose, you can propose it for deletion at WD:RFD. Deryck Chan (talk) 13:51, 2 November 2017 (UTC)
Thanks, my question was different. There may always be a need for such a page, even when it links only to a few other pages. However, I was curious to understand whether the items that are basically just redirects need to be removed or not. I have no opinions or preferences, or plan to do anything about that page. If the community here thinks it's fine in its current state, that's good to know. --Elitre (WMF) (talk) 14:59, 2 November 2017 (UTC)
There's no general opinion of the community that such links should be removed and there's currently an open RFC with the majority opinion that creation of redirects should be made more easy on Wikidata in the future. ChristianKl (talk) 15:14, 2 November 2017 (UTC)
Well, obviously if a page has no valid sitelinks and is not otherwise notable (since it is a Wikimedia entity) it fits the deletion criteria.--Ymblanter (talk) 16:29, 2 November 2017 (UTC)

Section for Interwiki conflicts

Do we really need https://www.wikidata.org/wiki/Wikidata:Interwiki_conflicts ? Isn't the project chat good enough to solve such issues? ChristianKl (talk) 14:01, 2 November 2017 (UTC)

  •   Keep. We get ~4 requests a month, most of which don't require extensive discussion, so I think it's a useful specialist forum to keep. But we should reduce the archive page frequency from 1 page a month to 1 page a year from 2018 onwards. Deryck Chan (talk) 14:17, 2 November 2017 (UTC)

Other sites section in items

Hi, I making some research on some aspects of the items that are create for structural role. I have noticed that commons category and meta are put in this section, and also that there is some ambiguity between commons category in this section and commons category used as a statements. Can someone link me the discussions where this was standardized? --Alexmar983 (talk) 04:38, 25 October 2017 (UTC) This way if there is a clear rule (like removing commons from other site, or putting both) I fix them, while lurking around.--Alexmar983 (talk) 04:44, 25 October 2017 (UTC)

There is no standardization when it comes to where to put the Commonscat. (There is consensus that it needs to be added to Commons category (P373), but no consensus that it only needs to be added to P373. There is no consensus on what to put in the other section when it comes to Commons.) Mbch331 (talk) 12:48, 25 October 2017 (UTC)
Link? Just curious to see the original discussion, some of them...
So Wikidata:Notability (point 1) point cites commons and a discussion of 2013, that is a consequence of this ambiguity? So for example Q15735643 if it does not exist for the quality of its ID, that is the Wikimedia username, was it created improperly for the commons link originally? And so it is not a valid item in its form for that reason, that commons is on "other sites" section.
I am not saying that you can fix it with some other ID (and I am doing with some of them) I am just saying that personally it is strange to improve such item to be sure it can be kept, when I see no real difference between existing items between putting in the right architecture, in the right "context" the commons category or a category of wikinews (e.g. Q18091124).
Plus, it looks like we really need structured data on commons, maybe with that this issue will find the momentum to be solved.--Alexmar983 (talk) 14:52, 25 October 2017 (UTC)
Talking about Commons category IMHO we have 2 case: 1) item is about a category: we can merge item with other item of categories, if don't exist, we must delete it. 2) item isn't a category item: I move link in Commons category (P373) and delete sitelink. --ValterVB (talk) 17:27, 25 October 2017 (UTC)
My comment was lost. I put it again. It is not about categories per se. So in Q15735643 what do you do? You remove it from the bottom since it is already in the "statements" section?--Alexmar983 (talk) 03:18, 26 October 2017 (UTC)
To put it simply, presence of Commons sitelink does not change the status of whether the item is (not) notable. Matěj Suchánek (talk) 13:46, 26 October 2017 (UTC)
That is what I know too. In that case with no structural need and weak ID, you still have to prove that it is notable.--Alexmar983 (talk) 03:00, 27 October 2017 (UTC)
@Alexmar983: Yes, the sitelink must be deleted, is an item about person not about category --ValterVB (talk) 16:43, 26 October 2017 (UTC)
But than why I was told that there is "no consensus that it only needs to be added to P373"?--Alexmar983 (talk) 03:00, 27 October 2017 (UTC)
removed.--Alexmar983 (talk) 03:03, 27 October 2017 (UTC)

There is clearly no consensus, and I add the category to sitelinks wherever it's appropriate, which is e. g. for many many buildings that never will have a "category" item and shouldn't. It's the only way that a user can see the Wikidata item automatically if stumbling across a category on the Commons. Before P373 is somehow read by the commons and used to include Wikidata as a link automatically, this is absolutely necessary. (Also, the commons must finally change their stupid data model anyway, and do everything with categories and use galleries only as some additional view of a category's topic. But that's another question.) --Anvilaquarius (talk) 07:38, 27 October 2017 (UTC)

Anvilaquarius all of this will change when commons will switch to structured data, probably. Not only because of the similar mentality but because with structured data you have many different way to create different galleries with a click of the mouse (big picture, B/W pictures, quality pictures, pictures used on wikipedias..). Every (often outdated) manual selection will compare poorly IMHO.
In the end, if it is practically kept for buildings, I still don't get what is the main difference about people? The same need or practical aspect than it is behind knowing that the commons pictures of a building are related to a specific database ID (national archive, wiki loves monument) is the same behind a person. I don't see from a data perspective where is the real difference here--Alexmar983 (talk) 11:10, 28 October 2017 (UTC)
I meant: it's useful and non-harmful for buildings (or persons or any other item) that is very unlikely to get a category in one ot the Wikipedias project. For larger towns and cities, e. g., it's very likely that a category exists or will be created in a wikipedia. Only the most famous people and some handfuls of buildings will have their own category in any other project than the Commons. --Anvilaquarius (talk) 14:51, 28 October 2017 (UTC)

For recent numbers and discussion on the links we currently have between Wikidata items and Commons categories, see this thread from September in the Project Chat archive, and its twin from Commons Village Pump.

@ValterVB: Please do NOT remove sitelinks from Wikidata items to Commons categories. They are required in order for Commons interwiki links to work properly.

Also please recognise that Commons categories have a dual nature -- they can be containers collecting related concepts, like categories on Wikipedias. But they also serve as the primary page for a concept on Commons, collecting images about that concept, parallel to articles on Wikipedias.

In the discussions above on both Wikidata and Commons, there was support for the proposition that a Commonscat should sitelink to a category item on Wikidata if such a category item exists; but that if such a category item does not exist, then it is entirely acceptable for a Commons category to be sitelinked to an article item here. (Pinging @Jmabel, Ghouston, Ruthven, Christian Ferrer:, @Zaccarias, ŠJů: who participated in discussions on this point on Commons, and @DarwIn, Jarekt, Mike Peel, billinghurst: @Multichill, Yann, Zaccarias, Mahir256: @Hsarrazin, D1gggg: on Wikidata.) Therefore, please do not remove a Commonscat sitelink unless you are moving it to specific category-type item here; and if you do make such a move, please make sure the new sitelinked item is linked to the old one by a category's main topic (P301) statement. There are templates on categories on Commons that rely on being able to draw data from a corresponding article-like Wikidata item, either via a direct sitelink, or a sitelink and a P301. Please do not break these templates.

As to item notability here, in general, for both the Structured Data on Commons project, in order to understand the information we already have in the category system about images, and also for the ability to more widely roll out the templates noted above, it is a real priority to try to increase the number of Commons categories we can identify to Wikidata (article-type) items. There is no need to create new Wikidata category-type items when we already have article-type items. And there is no need to create new items for Commons categories if they are intersections of concepts that already have Wikidata items. (It will probably be possible to describe such intersections directly on Commons, when the Commons Data system is up and running). But for Commons categories which cannot be described as an intersection of existing concepts, it is essential that such categories have or get a corresponding item here, if we are to be able to record statements about the thing or concept that category represents, in a structured way. This is a clear example of an item here being required to fulfill a structural need. Jheald (talk) 16:24, 31 October 2017 (UTC)

Really, I can't foresee any utility or advantage of linking Commons cats to Wikipedia cats, that's just wrong, pointless and useless, and insisting on such an absurdity keeps breaking the system there in Commons. It's wrong, and it's broken. After a bot was there patching those broken Wikidata links to make them appear as they should be, the situation is more palatable on those cats, but pretty please, do not continue replicating that error there, or at least make sure the patching bot is always running. Thanks,--DarwIn (talk) 16:50, 31 October 2017 (UTC)

Commons cats are more of a "label" that we put on a file in order to categorise it. I hope that in the future we will not have a tree-like structure like we have now, but we will use words/labels as the properties of an item to make the search easier. Actually Commons cats are used in the same way as Commons galleries, and they should be treated in such a fashion. --Ruthven (talk) 17:29, 31 October 2017 (UTC)
Contrary to DarwIn's point of view, the categories are bigger than just Commons alone linking to an article, as some of the categories link across multiple sister wikis, so they are not wikiPedia cats, they are wikiMedia cats. These wikimedia categories maintain benefits there, especially on broader subject matters, so as Jheald explained we have this layered approach. That Commons has a plethora of galleries of little value is part of the problem as well, as they are linked primarily to an article and would be seen to "evict" a more functional category link. Commons simply doesn't have a simple one size fits all approach due to the complexity of all the wikis.  — billinghurst sDrewth 20:39, 31 October 2017 (UTC)
This is a problem with the status quo, that some categories on Commons don't have a site link to Wikidata because a gallery page is already linked to the "article" item and there's no "category" item, e.g., c:Category:Abantiades barcas. The lack of a site link means that templates on Commons that extract data from Wikidata don't work, and there's no link to the Wikidata item in the sidebar. Interlanguage links will also be missing by default. They can be provided by using a "Interwiki from Wikidata" template with an explicit Q number, but explicit Q numbers seem like a bad idea (because there's no indication on Wikidata that the link was made, and the link may break if the Wikidata item is merged). Creating a category item in Wikidata would be one solution, but it's tedious to do manually and creates a somewhat useless item in Wikidata. An easier solution would be to delete the gallery site link and replace it with a site link to the category (gallery site links seem unimportant). Ghouston (talk) 21:21, 31 October 2017 (UTC)
I think galleries are a limited problem, and not worth worrying too much about. Looking at the recent numbers firstly almost all galleries that could have a sitelink already do have a sitelink. Secondly, looking at historic versions of these numbers the number of new galleries is growing really very very slowly. But it is true that out of 94,630 items where there is a Commons gallery (P935) property (tinyurl.com/ybycd7bg), only 20,877 currently have a topic's main category (P910) to a category-item that could support a sitelink (tinyurl.com/ybxpox92), whereas 70,032 have a Commons category (P373).
Picking up the point of @Ghouston:, I think it's important to note that generally the c:Template:Interwiki from Wikidata does not require an explicit Q-number, if there is a category's main topic (P301) that the template can follow from a sitelinked category-item to the relevant article-item. (A mechanism that other templates can also use). Thus -- so long as category's main topic (P301) / topic's main category (P910) pairs exist to link a category-item to an article-item -- it doesn't hugely matter if a Commons category links to a category-item or an article-item here. It's easier and more direct and cheaper on server resources and one less thing to maintain if the Commons category can link direct to an article-item; but it's not the end of the world if it links to a category-item, so long as there is a category's main topic (P301) that a template can then follow to the article item.
And as User:Billinghurst points out, sometimes that category-item link can be very useful for interwiki links to other projects, in particular wikisource. So if some of our current Commonscat <-> article-item sitelinks got changed into Commonscat <-> category-item sitelinks, that might be no bad thing, if it opened up a big new set of Wikisource <-> Commons interwikis; but only so long as those category-items do have P301s to get to the corresponding article-item. (One thing in fact that I noticed in the latest numbers was a big jump in the number of category-items here on Wikidata: up from 2,870,035 to 4,043,980 in the eight months since January. I don't know whether this is a big set of Wikisource categories having Wikidata items created for them; or alternatively, if it is a big set of P31s being added to existing Wikidata items. But it is worth noting that there was no corresponding increase in P301s. If people are creating new category-items on Wikidata, please do also create relevant category's main topic (P301)/topic's main category (P910) pairs if possible, otherwise they're not connected to the article-type items that have most of the useful statements on them.)
But User:Ghouston is quite right, that when there is no category-type item, then that does require an explicit Q-number in the "Interwiki from Wikidata" template, and that is not ideal. Jheald (talk) 23:15, 31 October 2017 (UTC)
Adding "Interwiki from Wikidata" with a Q number gives the interlanguage links, but links to non-Wikipedia projects are still missing, including Wikidata. It doesn't help other templates in any way either, e.g., a template on c:Category:Abantiades barcas has added a category "Biology pages without wikidata link". So it's basically broken without a site link, but with the interlanguage links pasted on so that fewer people notice. Ghouston (talk) 23:25, 31 October 2017 (UTC)
 
Wikidata <-> Commons links
User:Alexmar983, Among other things I do a lot of maintenance on Wikidata <-> Commons links and writing Commons templates that use Wikidata. My approach is mostly to follow plot on the right, which was drawn, in the early days when there were great many parallel proposals. My templates on Commons rely on Commons category (P373), Commons gallery (P935), Commons Creator page (P1472), etc. properties , but sitelinks to Commons are unreliable. As a result, if there is a sitelink to Commons I use it, but adding or removing them should not break anything on Commons or other websites. Interwiki links can be added to Commons pages with the help of c:Template:Interwiki from Wikidata. --Jarekt (talk) 12:02, 1 November 2017 (UTC)
@Jarekt: I would have thought that detecting an incoming P373 to a Commons category page would be (a) quite expensive in terms of server resources, and (b) difficult in Lua. There is a user script I know of and use, wdcat.js that does use this approach; but in general I would have thought it would be much more efficient to use sitelinks or sitelinks + P910, in the way that the "Interwiki from Wikidata" template does, if at all possible? Jheald (talk) 14:05, 1 November 2017 (UTC)
Jheald, "detecting an incoming P373 to a Commons category page" is not just expensive and difficult in Lua: it is AFAIK totally impossible. There is phabricator:T99899 proposing to "Allow looking up Entities by external identifiers", that would make properties function in a way similar to sitelinks. So far I was reluctant to use sitelinks as a mean of connecting from Commons to Wikidata because sitelinks to Commons can shift so unpredictably, but I see your point, maybe I can trust that once there is a sitelink to Commons it might change from article-item to category-item (or vice versa), but it is not going to disappear. So far more stable approach was to hard-wire q-codes on Commons pages, except for the pain of kepting them in synch with Wikidata merges and deletions. Also detecting sitelinks breaks down if page is transcluded, for example you have an article item with a sitelink to Commons category and in that category you have an infobox that relies on the sitelink to detect which item to draw data from. Than if you transclude that category, (and for some crazy reason many Commons pages do that) on some other page than the trancluded infobox can not detect the sitelink. --Jarekt (talk) 14:55, 1 November 2017 (UTC)

I do not really understand most of the technical stuff you are discussing here, and I have no intention at all of interfering with the way Wikidata does things and organizes its stuff. But there is no way things like this scheme on the right can be right (unless I failed to understand it, which, I concede, could be very well the case). The Pablo Picasso category in Commons should always link to the Pablo Picasso article in the wikipedias, that's what we need to organize, classify and curate stuff there. Linking it to the Picasso wikipedia cat is useless and pointless from a Commons perspective (if you really need to do that here, please make sure we don't notice it at Commons, as it only gets on the way of our work there. Not only it doesn't help with anything, it's an hinderance. I don't say this is a kind of biblical rule that should be followed by Faith, but to anyone familiar with Commons, it should be quite obvious that in general Commons cats need to be linked in some meaningful way, preferably by interwikis (as it used to be), with the corresponding articles, as that's what we use in general as a kind of a guide for naming, organizing, etc. As far as I know, we in Commons don't use Wikipedia categories for anything meaningful, as wikipedia articles in general also don't use Commons galleries in any meaningful way - and for some bizarre reason those were precisely the two items you have linked in Wikidata, making it an apparently useless connection. I don't know how this fits with your technical stuff here, but that's what it is.--DarwIn (talk) 15:30, 1 November 2017 (UTC)

As has been expressed previously, this is a holistic model, so please stop thinking in the smaller paradigm of Commons <-> Wikipedia. What about category links between Commons <-> Wikiquote(s), Commons <-> Wikisource(s), etc. Your scheme is too limiting. Wikipedia needs to utilise the other fields and programmatically resolve that linking if it is required.  — billinghurst sDrewth 22:44, 1 November 2017 (UTC)
@billinghurst: I do not understand you, nor what you are talking about. I speak about the way Commons works, and that's what it is, and that's how it is. The fact that you don't like it, don't approve it, or whatever, is irrelevant to me, and, I believe, to Commons folk in general. Just try to not mess up too much there with your things, ok? Thanks.--DarwIn (talk) 00:05, 3 November 2017 (UTC)
Mate. I am an admin at Commons, and just a worker bee here, so I am hardly going to advocate to muck up Commons. It has been tried to be explained that this is bigger than the singular aspects of Commons <-> Wikipedia, to which you refer. Wikidata has to work for all Wikimedia wikis, and no harm is being doing to Commons from its presence or its data. So please don't try to conflate this to an argument as something about damaging Commons, or me and my desires; they are false arguments.  — billinghurst sDrewth 00:58, 3 November 2017 (UTC)

Physical vs. concept of flags

I'm currently visiting the Estonian National Museum, which opened to great fanfare last year, and one of their most treasured objects is the first, original, physical, flag of the country, made of silk in 1884 - which I've just created as an "instance of -> Flag" on Wikidata at first flag of Estonia (Q42522126).
However, this has led me to a problem... how do we differentiate between the concept of "the flag of xyz" (its design, its legal significance, etc) and "a specific physical flag of xyz" which is notable. Currently we list all national/city/battle flags as "instance of -> flag" even though we are generally not referring to a physical item. e.g. flag of Estonia (Q81471).
For context, there are currently 18 results for a query of all instance of -> flag that also are in a museum collection. Another good example of this issue is Star Spangled Banner Flag (Q3979151) (in the Smithsonian museum collection) which is an instance of -> flag of the United States Of America (Q42537), which is itself listed as instance of -> national flag (Q186516).
At first thought, it seems incorrect for the ideas/concepts of national flags to be listed as 'instance of' flag and should instead be 'subclass of' flag since we're nearly always NOT talking about a physical item - but I’m certainly not convinced that’s idea either. Wittylama (talk) 12:58, 1 November 2017 (UTC)

Isn't this what manifestation of (P1557) was created for? If I understand correctly, you can replace instance of (P31) with manifestation of (P1557) in first flag of Estonia (Q42522126). --Yair rand (talk) 13:52, 1 November 2017 (UTC)
Good thought - I've now added the "manifestation of" statements for first flag of Estonia (Q42522126). I've left the 'instance of' statements intact and not modified Star Spangled Banner Flag (Q3979151) for the time being, as I'm not sure if the use of manifestation is supposed to supersede, or be in addition to, the 'instance of' statements.
However, that nevertheless doesn't address the issue that these 'concept' flag items (e.g. flag of Estonia (Q81471)) are being classified as 'instances of' which implies they're physical individual objects (if I'm not mistaken). Wittylama (talk) 14:08, 1 November 2017 (UTC)
Well, national flag (Q186516) is currently listed as subclass of (P279) national symbol (Q1128637), which seems to imply that it's not talking about physical objects. --Yair rand (talk) 14:21, 1 November 2017 (UTC)
Flag, and National Flag, are listed as “subclass of” items, yes. But “flag of estonia” (and other countries) are listed as “instances of” items. Shouldn’t these be “subclass of national flag”? Wittylama (talk) 14:42, 1 November 2017 (UTC)
This is an example of a common issue. I believe either solution is acceptable, but the solution chosen determines the precise meaning within wikidata of the top-level item. If flag of the United States Of America (Q42537) is an instance of national flag (Q186516) (which is a subclass of flag (Q14660)), then national flag (Q186516) and flag (Q14660) are metaclasses, i.e. a class whose instances are themselves classes which have specific physical instances (or manifestations, but the same problem applies if you choose another property instead of P31). If on the other hand it's a subclass relation, then flag of the United States Of America (Q42537) and national flag (Q186516) are the same kind of thing, and you could have additional subclasses like the 'xx star American flag' for the various specific flag designs that changed as the number of states increased, etc. and they would work as part of the same class hierarchy. I think this latter approach makes the most sense, but either way is workable (the 'xx star American flag' can still be a subclass of flag of the United States Of America (Q42537) even while that's an instance of national flag (Q186516), with some consistency). ArthurPSmith (talk) 15:39, 1 November 2017 (UTC)
In German, there is a distinction between „Fahne“ (piece of cloth) and „Flagge“ (arrangement of shapes and colors), though colloquially they’re usually treated as equivalent. However, the items currently linked to the respective dewiki articles don’t seem to reflect that distinction – Q1371499, linked to „Fahne“ (physical), has subclass of (P279)symbol (Q80071), while Q14660, linked to „Flagge” (concept), is described as a “piece of fabric” in English. I also don’t know if the other articles linked to those items reflect the same distinction. —Galaktos (talk) 00:40, 3 November 2017 (UTC)

Automatically adding items that are linked to the watchlist

Given Wikidata's model of items the average number of edits per page is a lot less than Wikipedia. As a result a lot of pages don't have enough people catching vandalism.

Currently, when I create a claim Qaaaa Pbbbb Qcccc, that puts Qaaaa on my watchlist but doesn't put Pbbbb or Qcccc on my watchlist. I think it would be good if it would put Pbbbb and Qcccc as well on the watchlist (with optionally deactivating this behavior in Preferences). ChristianKl (talk) 15:13, 2 November 2017 (UTC)

  Support John Samuel 22:27, 2 November 2017 (UTC)

Inverse for "member of" (P463)

Is there an inverse property for "member of" Property:P463? It doesn't seem right to use "has part" Property:P527 as that is the inverse of "part of" Property:P361, which seems to have a different definition than "member of". Mnnlaxer (talk) 16:37, 2 November 2017 (UTC)

I'm against inverse properties now, we have the query service for this. We only need a better way to show incoming links. Sjoerd de Bruin (talk) 17:50, 2 November 2017 (UTC)
I'm new here. I don't know what the query service is or what incoming links you're talking about. I'm not proposing a policy, just looking for an answer. I'm wondering if there is a better property than "has part" for what I would best describe as "has member". See Q16969866 for the member schools of a collegiate athletic conference. Mnnlaxer (talk) 19:15, 2 November 2017 (UTC)
has part(s) (P527) is not appropriate for that situation. member of (P463) links are not supposed to have any corresponding links going in the other direction. --Yair rand (talk) 19:27, 2 November 2017 (UTC)
Thank you. If I understand correctly, it is redundant to have "part of" statements on the organization, just individual "member of" statements on each member's item. Mnnlaxer (talk) 19:40, 2 November 2017 (UTC)
Also see above (Wikidata:Project chat#Consistency Task Force: part of / has part usage and documentation is inconsistent--Fractaler (talk) 19:38, 2 November 2017 (UTC)

Best practices - coordinate locations

I have an item Narragansett Bay (Q1241943) that has a coordinate location <imported from> EN Wikipedia with no further reference. I have added a GeoNames ID and the coordinates from GeoNames, which are slightly different. In this case should I deprecate the values from EN Wiki or what? - PKM (talk) 19:54, 2 November 2017 (UTC)

I would check them out first. The en-wiki coordinates are sometimes better than GeoNames ones.
I'm also cautious about our policy on importing coordinates from GeoNames. GeoNames is licensed CC-BY. Is Wikidata's relicensing as CC0 of information originating from GeoNames that shows some skill or judgment legally sound? Jheald (talk) 21:07, 2 November 2017 (UTC)
Given the data quality of GeoNames I don't think there's a good reason to replace data that's sourced from EnWiki with GeoNames data. We purposefully don't import the data with bots. The values should only deprecate values when you have reason to believe that the value is wrong. ChristianKl (talk) 21:38, 2 November 2017 (UTC) S
Worth bearing in mind for situations like this, when the coordinates describe a large feature with no well defined central point, that both can be equally correct. In this case, both points are firmly inside the bay, and are about two or three miles apart - the bay itself is 150 square miles. Andrew Gray (talk) 21:33, 2 November 2017 (UTC)
Thanks, all. - PKM (talk) 21:47, 2 November 2017 (UTC)
Should also note that geonames isn't an original source, they compiled data from other sources (and sadly without keeping reference from where each dataset came). For non-US-places a big percentage originally is from Geographic Names Server (Q1194038), so instead adding a coordinate with reference to geonames it'd be best to find and use the actual source instead. Ahoerstemeier (talk) 22:19, 2 November 2017 (UTC)
Thanks for that tip. I got in the habit of using GeoNames because the system prompts for GeoNames ID for geographic objects so I assumed it was the best practice. Guess I know better now. Now if the USGS GNIS site would just respond more often.  :-) - PKM (talk) 23:39, 2 November 2017 (UTC)
The system suggests GeoNames ID because the quantity of the GeoNames database is big because they imported nearly every available database that's compatible with their license no matter the quality of the database. ChristianKl (talk) 00:49, 3 November 2017 (UTC)

can anyone help me help our artist

we created a TQ TARIQ wikidata to help with the google knowledge graph and it was deleted, can anyone help  – The preceding unsigned comment was added by Freesolollc2017 (talk • contribs) at 21:17, 2 November 2017‎ (UTC).

@ Freesolollc2017: Please see our policy, WD:Notability. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:22, 2 November 2017 (UTC)
See also Help:FAQ/Freebase#By_adding_to_Wikidata.2C_I_have_a_free_ticket_into_Google.27s_Knowledge_Graph.2C_right.3F. I don't see any deleted contributions on your account, though. Sjoerd de Bruin (talk) 21:30, 2 November 2017 (UTC)

Harvest templates for latitude+longitude

French railway stations have an infobox with latitude= and longitude=, how to use these values to fill their empty coordinate location (P625)?

I am a fervent user of Harvest Templates but it does not seem to be able to handle that case (I'd be happy if you prove me wrong). Thanks! Syced (talk) 05:47, 2 November 2017 (UTC)

@Pasleim:, who maintains Harvest Templates. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:37, 2 November 2017 (UTC)

  Done (there were just 24 of them, most railway stations having already been harvested for a long time).

HarvestTemplates often balks for properties with a large existing base, and moreover tries to implement the unique constraint by retrieving all existing values (and times out here). A workaround for the latter limitation is to set a breakpoint and hijack the overkill constaint. Nono314 (talk) 08:09, 3 November 2017 (UTC)

This wouldn't help, though. HT has never been able to harvest coordinates. Matěj Suchánek (talk) 08:16, 3 November 2017 (UTC)

Why are only registered users allowed to view Wikidata in other languages?

Hi,

I use wd from time to time and contribute things when I find something that matters. But as a German user it would be much easier if I could use the German frontend. Some time ago I could add ?setlang=de at the end of the url which was unneccesarily complicated but that way I could use the german frontend. For quite some time this doesn't work anymore and if you go to the language switcher it says that language selection does only work when logged in. Why do you make it so hard for non-English-speakers? 91.221.58.29 10:38, 2 November 2017 (UTC)

But how could we make German by default? If you are a registered user, you can cosse it. If you are not - well, may be we might look at the geolocation of the IP and choose the preferred language based on geolocation, but this is presumably difficult and I guess sometimes politically dangerous (Belgium? Switzerland?).--Ymblanter (talk) 10:49, 2 November 2017 (UTC)
I have no idea how you came to the idea to make German the default. I only want the old language-switcher. 91.221.58.29
This is phab:T134592. Daniel, can you give a quick summary? Sjoerd de Bruin (talk) 10:54, 2 November 2017 (UTC)

So if I understand Ymblanter correctly, the URL-extension language switcher has been deactivated because of political problems with Belgium und Switzerland? And phab is the removal discussion? Without a language switcher wd is way to complicated for me, sorry. but thanks for the explanation. 91.221.58.22

In the spirit of trying to understand you, what's holding you back from registering an account? ChristianKl (talk) 11:19, 2 November 2017 (UTC)
FWIW, the unnecessarily complicated way still works if you use ?uselang=de. And IIUC the bigger problem with using the IP’s location is that it doesn’t play well with caching. --Lucas Werkmeister (WMDE) (talk) 13:46, 2 November 2017 (UTC)
By the way, the "switcher" was removed in 2015 because of some technical issues. Matěj Suchánek (talk) 08:20, 3 November 2017 (UTC)

Question about brands and brand owners

A given car manufacturing company may produce cars of more than one brand or marque. E.g. Jaguar Land Rover produces Jaguar cars and Land Rover cars. Therefore each Jaguar car model has the statement brand=Jaguar (e.g. Jaguar F-Pace). My question is how should these brands/marques be added to the car company itself? Currently Jaguar Land Rover has a brand statement with two values: "Jaguar" and "Land Rover", which does not seem right. I searched for "owned brand" and "brand used" and a few other variations, but drew a blank. Any advice would be appreciated. Thanks. DeFacto (talk) 21:57, 2 November 2017 (UTC)

Jaguar Land Rover (Q6122893) owns Land Rover (Q26777551), with start time 2013? Ghouston (talk) 03:30, 3 November 2017 (UTC)
OK, sounds good. I was imagining something more specialised and more complicated for some reason. Thanks for that. DeFacto (talk) 16:49, 3 November 2017 (UTC)

Wikibooks cookbook tagging scheme

Does anyone know if there are efforts to describe individual recipes from the Wikibooks Cookbook? ~★ nmaia d 16:54, 3 November 2017 (UTC)

Talks or presentations in a conference

I would like to add the talks/presentations made during WikidataCon 2017 (Q37807682). What is the best approach to do this? Should I use property has part(s) (P527)? or is there any other property available to represent talks/presentations? John Samuel 20:32, 3 November 2017 (UTC)

The individual talks can link to WikidataCon 2017 (Q37807682) but I don't see a reason why WikidataCon 2017 (Q37807682) has to link to all the presentations directly. ChristianKl (talk) 21:04, 3 November 2017 (UTC)
In that case, how to link to WikidataCon 2017 (Q37807682)? What do you suggest? Using part of (P361)? John Samuel 21:10, 3 November 2017 (UTC)
I looked at examples and found Tim Berners-Lee: The year open data went worldwide (Q22946133) which does use part of (P361). I'm however not really happy with it and will create a new property proposal. ChristianKl (talk) 21:33, 3 November 2017 (UTC)
Something like 'presented in' would be interesting. Please add the link here, if you make a proposal. Thanks. 21:36, 3 November 2017 (UTC)
I created Wikidata:Property proposal/Creative work#presented in ChristianKl (talk) 22:59, 3 November 2017 (UTC)

Sandboxes

What's the best way to setup a Sandbox page for a project to experiment with different entry techniques? Basically, I'd like to have the functionality of a Wikidata item entry to interactively enter information but not have it muck up other things. Thanks in advance! Madcoverboy (talk) 22:16, 3 November 2017 (UTC)

We have https://test.wikidata.org/wiki/Wikidata:Main_Page . ChristianKl (talk) 23:14, 3 November 2017 (UTC)

Sources for constraint violations reports

I've poked around a bit, but was not able to find any SPARQL source code of the various constraint violations reports. Hints very much appreciated - Jneubert (talk) 09:17, 3 November 2017 (UTC)

Every property has its talk page with links to miscellaneous reports. What exactly are you looking for? Matěj Suchánek (talk) 17:04, 3 November 2017 (UTC)
As far as I know, most constraint checks don’t use SPARQL. The source code for the WikibaseQualityConstraints extension, which powers the checkConstraints gadget, is on GitHub (specifically, all the checkers are here, and the SPARQL queries used are here). The source code of KrBot, which creates the periodic database reports, isn’t public AFAIK. --Lucas Werkmeister (WMDE) (talk) 17:46, 3 November 2017 (UTC)
Thank you very much! I'm currently working on variations of the unique constraint and single constraint reports, which take into account mapping relation type (P4390) (particularly, for control of n:m relations to external identifiers). Your links will help me understand how the gadget works - which is a really great extension to the user interface, BTW! Jneubert (talk) 15:57, 4 November 2017 (UTC)

historical motorcycle manufacturer (Q15081032)

I fail to see why this should be a separate item from motorcycle manufacturer (Q15081030). Whereever one needs to denote that it is "historical", just use start time (P580)/end time (P582). Thoughts? Jon Harald Søby (talk) 13:34, 3 November 2017 (UTC)

Much has been said on this problematics:
I would advocate using former entity (Q15893266) a year ago but now I do the contrary. Wikidata:Project chat/Archive/2017/06#Building under construction offered a solution. Matěj Suchánek (talk) 17:18, 3 November 2017 (UTC)
There are 2 types of motorcycle manufacturer (Q15081030): 1) existing and 2) non-existing (historical motorcycle manufacturer (Q15081032), fictional motorcycle manufacturer, etc.) --Fractaler (talk) 12:37, 4 November 2017 (UTC)

Result of a battle

Is there any property to mention the result or outcome of a war or battle or military operation etc.?  – The preceding unsigned comment was added by Krishna Chaitanya Velaga (talk • contribs).

number of deaths (P1120), number of injured (P1339), number of casualties (P1590), number of missing (P1446), has effect (P1542), end cause (P1534). - Kareyac (talk) 16:13, 4 November 2017 (UTC)

industrial railway (Q256324), field railway (Q1402344), and tramway (Q7833250)

I am trying to determine the difference between industrial railway (Q256324), field railway (Q1402344), and tramway (Q7833250) and determine which which wikidata item should a specific Japanese article belongs to.

As I understand, Q256324, Industrail Railway in English, refer to railways that are specifically made for industrial purpose.

Q1402344, Feldbahn, which can probably be interpreted as field railway, refer to railways that are also being used by industry, but are often simpler or in narrow gauge?

Q7833250, as I understand, is also referring to railway being used for industrial purpose and are simpler than regular railroad, but what are their difference from the three items above?

And then there is the Japanese Wikipedia article, ja:専用鉄道 Senyou Tetsudou, which is currently linked to Q1402344, and the article title can be literally translated as Dedicated Railways. The article currently introduce two different things together:

  • Rail lines specifically built by specific industry owner for their own use and to connect with the rail network, according to the Tetsudou law.
  • Railway that are for self use in boarder sense that are not restricted to legal definition, also including rail lines that are build within individual property and also rail lines are not connected to rail network.

Which of the above three wikidata items would be a better fit to connect the wikipedia article to?C933103 (talk) 13:50, 4 November 2017 (UTC)

Possible new property for material handling

We already have properties for things like product or material produced or service provided (P1056), but as far as I can see, there is no way to claim what item or material a given item interacts with.

Generally this is referred to as "material handling" in English.

So, for example, a bulldozer (Q131677) handles soil (Q36133), crushed stone (Q1262699), etc. A pallet jack (Q830071) handles pallet (Q505996).

Essentially, it would be useful to claim that this item is designed to, or is typically used to interact with another item.

Can anyone see a way to use existing properties, or is a new one needed? Danrok (talk) 15:06, 8 November 2017 (UTC)

See designed to carry (P3349). Thierry Caro (talk) 16:09, 8 November 2017 (UTC)
Thanks. I think that should cover it. Danrok (talk) 18:13, 10 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 08:49, 11 November 2017 (UTC)

Regex constraint in NLP ID (old) (P1695)

Hi! We've just had a discussion (in the main Facebook group for Polish speaking Wikimedians) about issues with P1695 (which is the NLP ID - authority control from the Polish National Library). I always edit with the "check Constrains" gadget turned on and quite often when I add values for this property I get a warning that the number provided breaks the property consraints, because it doesn't match the regex. One of my plWiki collegues (more proficient in computer studies that I am) has just run some checks and he says that, for example, in Marlena Happach (Q30902341), this warning shouldn't appear as the NLP number does match the regex. Any ideas what's wrong? Thank you in advance. Powerek38 (talk) 10:02, 10 November 2017 (UTC)

There’s a U+200F RIGHT-TO-LEFT MARK character at the end of Marlena Happach (Q30902341)’s NLP ID (old) (P1695) value – it’s invisible, but it makes the regex match fail. (Try editing the value, setting the cursor to the end of the input box, and then pressing the left arrow button.) --Lucas Werkmeister (WMDE) (talk) 10:13, 10 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 08:49, 11 November 2017 (UTC)

ORES problems

I noticed, that old ORES, that I used frequently, was disabled in favor of New filters for edit review. But I am not able to set up these filters to replicate old ORES results, as all edits (including blatant vandalism) are marked as "very likely good" and "very likely good faith", so there is nothing to highlight.--Jklamo (talk) 09:48, 3 November 2017 (UTC)

Was it disabled? I thought ORES had been made compatible with either interface... Matěj Suchánek (talk) 17:01, 3 November 2017 (UTC)
ORES is no longer on my list of Beta features.--Jklamo (talk) 23:47, 5 November 2017 (UTC)

Template talk:Authority control at Commons displaying Wikidata authority control data needs merging

Please comment at commons:Template talk:Authority control on displaying Wikidata authority control data at Commons in categories. There are two templates that can be merged if we modify one to display the Q-number rather than just display the Wikidata symbol. Please come and comment. You can see an example of the problem here: Commons:Category:Ransford Dodsworth Bucknam --Richard Arthur Norton (1958- ) (talk) 02:40, 6 November 2017 (UTC)

Paired and unpaired organs

human eye (Q430024) (human ear (Q7362), human kidney (Q9377), etc.) is a paired organ? human eye (Q430024) is half (Q39373172) or dyad (Q29431432)? --Fractaler (talk) 09:54, 13 October 2017 (UTC)

Human -> has part -> kidney -> quantity -> 2 ? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:43, 13 October 2017 (UTC)
For the human eye I created pair of human eyes (Q41890371) for you. In general it's helpful when you reference the FMA IDs when creating items like this. Feel free to create more of the items. ChristianKl (talk) 17:00, 13 October 2017 (UTC)
@Fractaler: A dyad is both of the organs at the same time. It's what the FMA calls "set". Do you see an issue with describing the content this way? ChristianKl (talk) 21:10, 13 October 2017 (UTC)
pair of human eyes (Q41890371) (set of human eyes or pair of human eyes) is anatomical set (Q27058130)? anatomical set is a pair (group of 2)? --Fractaler (talk) 16:50, 14 October 2017 (UTC)
If you go up the subclass tree you find that it subclasses "Anatomical set". Being an anatomical set itself doesn't mean that it's a pair. See http://xiphoid.biostr.washington.edu/fma/fmabrowser-hierarchy.html?fmaid=55652 . I wouldn't have an issue with adding an additional "antomical pair" class if you think that's necessary. ChristianKl (talk) 14:35, 16 October 2017 (UTC)
Thank you for this useful links, at once there is a weight of questions (material anatomical entity=material anatomical object? Anatomical set = group of anatomical object? Why Foundational Model of Anatomy (Q1406710) has not link to http://xiphoid.biostr.washington.edu/fma/index.html? Etc.) About anatomical pair (paired organ). They were created, but then they were deleted without discussion and I do not know what to do now. --Fractaler (talk) 14:17, 17 October 2017 (UTC)
It's a nature of a set that all it's parts are of the same class. If you have two hands and a liver that's a group but no set. To me the phrase "paired organ" suggests that there's an organ that has the attribute of being "paired". That's from an ontological standpoint different then having an pair item that contains the left and right version of the organ. ChristianKl (talk) 15:40, 20 October 2017 (UTC)
What kind of set (Q36161)? "1) two hands, 2) liver" is the list set (extensional definition (Q5421961), no definition). Set "two hands as an organ" and set "liver as an organ" are subsets of the set "organs" (intensional definition (Q1026899), has definition). And set (Q36161) is mathematical term. Not in mathematics it is better (for the elimination of homonymy) to use its non-mathematical analog (for example, group (Q16887380)). --Fractaler (talk) 14:43, 22 October 2017 (UTC)
In refering to the way FMA uses the word set. ChristianKl (talk) 11:49, 30 October 2017 (UTC)
What is a set (FMA)? Is anatomical set (Q27058130)? anatomical set (Q27058130) has no definition. What should come here, and what not? A set of letters without definition = 0. --Fractaler (talk) 13:28, 30 October 2017 (UTC)
human eye (Q430024) is half (Q39373172) the human binocular organ. The human binocular organ is a visual, binary (paired) organ. The binary organ is intermediate between the unary/unpaired organ (the last) and the ternary organ (previous) (in accordance with the Valentin Dogiel (Q2406712)'s law of oligomerization of Q42101062 - oligomerization of homologous organs (Q4333796)). Binary organ can not consist only of 1 part (only 1 human eye (Q430024), only 1 upper extremity (Q841423), only 1 human leg (Q6027402), only 1 kidney (Q9377), human lip (Q15173), human nostril (Q858590), etc.), it is always an anatomical structure of two parts. --Fractaler (talk) 14:43, 22 October 2017 (UTC)
If you want to understand this better, go and actually look at external sources like FMA yourself. ChristianKl (talk) 10:07, 31 October 2017 (UTC)
Found: The second principle of modeling declares three of these parts, Organ, Cell and Biological macromolecule as units of structural organization of the whole (i.e., the Body). Other classes of the Anatomy Taxonomy subsume concepts that refer either to aggregates of these units (e.g., Organ system, Anatomical set, Anatomical cluster), or are parts of these units (e.g., Body part, Organ part, Cell part). So, organ system (Q188193), anatomical set (Q27058130), anatomical cluster (Q28654910) are an aggregate of units of structural organization of the whole? --Fractaler (talk) 11:49, 31 October 2017 (UTC)
paired limb/fin segment (Q41813561) is a paired organ? --Fractaler (talk) 12:19, 29 October 2017 (UTC)
I don't see a good reason for such a classification. No existing ontology for anatomy classifies it that way or has a "paired organ" category. If you want to invent a new way to categories anatomy that's different from the existing ontologies you should provide a good justification why you think the existing way of doing this isn't enough or provide references that back up the categorisation you favor. ChristianKl (talk) 18:35, 29 October 2017 (UTC)
in Russian. May be in English for парные и не парные органы use other terms? It is not my classification. We are studying this biological, medical concept even at school.--Fractaler (talk) 20:13, 29 October 2017 (UTC)
There's no reason to link to Google. It would make more sense to link to actual Russian anatomical texts and use them as references when you make you claims. ChristianKl (talk) 20:49, 29 October 2017 (UTC)
Ok. In Russian, for example: 1) journal “Фармацевт Практик” #02′ 2015, biologist's article "Why does a person have paired and unpaired organs?"; 2) elementy.ru: Why does a person have some organs - paired (for example, lungs, kidneys), and others - in a single copy?, Liver - unpaired organ, ..., Body structure: On the head are sensory organs: unpaired - nose, tongue; paired - eyes, ears, balance organ; According to the structure scheme, unpaired organs in vertebrates are placed along the middle line, but they can move to the right or to the left; Pairs are usually located in relation to this median line on the right and left, etc. From the school textbook (where the high school student is introduced to the basic concepts of a paired / unpaired organ, give examples), to the university textbook (where the student is explained in more detail the evolution of these organs, its direction, etc.). I think this biological concept also should be in English, it simply has some other name (sorry if I incorrectly translated this term and misled). Is it necessary to ask biology specialists from a biological project to confirm that this concept exists in biology? --Fractaler (talk) 06:24, 30 October 2017 (UTC)
In English the sentence you refered to with "have" doesn't imply identity but imply that being "paired" is a property of the organ. To me your arguments don't suggest that you understand the difference of something being an identity and being a property well enough, that I would be convinced that Russian does have different concepts about this then English anatomical categorizations. Maybe Russian speakers who are interested in anatomy want to comment? @Infovarius, Роман Беккер:. ChristianKl (talk) 08:53, 30 October 2017 (UTC)
  Notified participants of WikiProject Biology: which (in English) term ("set 1") combines the following items: liver (Q9368), spleen (Q9371), tongue (Q9614), etc.? Which (in English) term ("set 2") combines the following items: 2) lung (Q7886), pair of eyes, pair of kidneys, pair of ears, etc.? Sets "set 1" and the "set 2" must have a superset. --Fractaler (talk) 16:38, 6 November 2017 (UTC)

Bots

Pinging other crats, @Lymantria: and @Vogone:. I had to remove the flag from User:ScorumMEBot, which I approved in September. The bot task was to create items on statistics of Japanese soccer league, apparently using these results on the website affiliated with the bot owner. It turned out that the bot first was working properly, but recently they have changed urls at the website, and the (correct) links turned into the main page of the website which promotes some cryptocurrency, aka spam. I mass-deleted recent additions of the bot and checked that the previous ones are fine (though the refs are hosted at the same website, and I have my doubts even about them). This was discovered at the English Wikipedia by an opponent of Wikidata, and quickly escalated in a serious of personal attacks against me, but this is not the point. The point is that there was no community input for the bot task, and I approved it after waiting for some time, getting no objections, and checking that test edits are good (though I should have insisted that the bot makes at least 50 edits, as specified by the policy; however, I would likely have approved it after 50 edits anyway since the edits were good). Generally, RfP for bots attract very little attention, and we will likely have more such cases in the future. Of course we can require that there is a minimum number of votes needed for approval of a bot task, in which case virtually no tasks will be approved. I would appreciate some community input on the issue before we run into more troubles.--Ymblanter (talk) 11:42, 31 October 2017 (UTC)

Thanks for the report. I am not aware of any enwiki discussions, so from what I see within Wikidata, your reaction to the situation was appropriate in my opinion. I also do not see that we should change the bot approval procedure. Open projects such as Wikidata are to some extent vulnerable by bad-faith editors, but we do have the tools to deal with them—just as you did in this situation. The same could have happened to the Wikipedia projects. —MisterSynergy (talk) 12:47, 31 October 2017 (UTC)
Thank you for your reaction. I actually think that this particular case has been already handled, there is nothing to be done unless the bot owner reapplies. However, greater participation at the approval stage would have likely resulted in understanding that the website itself might be not an optimal reference venue, and the request might have been rejected or stalled forever. PS. You do not want to be aware of the enwiki discussion, but currently the most likely development of the situation will be that an RfC is held later this year, and as a result of this RfC any usage of Wikidata in the article namespace is going to be prohibited. This incident will be used as a justification of the thesis that the Wikidata community tolerates spam.--Ymblanter (talk) 13:12, 31 October 2017 (UTC)
Although I support your request for more community input, I don't think more input would have prevented the case at hand. Good actions by you, thank you. Lymantria (talk) 15:15, 31 October 2017 (UTC)
MisterSynergy said what I had on my mind when reading your message: The same could have happened to the Wikipedia projects. Matěj Suchánek (talk) 13:40, 31 October 2017 (UTC)
We created a template for bot requests and this template requires a link to a community discussion justifying the request. I think if no discussion in a wikiproject/in the Project chat/in a RfC/in a property deletion can be provided then the task should not be approved and bot operators should ignore the task which will be deleted after 6 months.
For bot flag request, the best is to accept only requests supported by several contributors.
Your example shows the importance of good selection in data import to avoid unreliable sources and this can be done only if several contributors with a good knowledge of the subject agree on the sources selection. Too many persons use the first online database which can be unreliable instead of working to extract data from an official database. Snipre (talk) 16:56, 31 October 2017 (UTC)
And now Ymblanter is being harassed by Fram in the form of an arbcom case, see en:Wikipedia:Arbitration/Requests/Case#Crosswiki_issues. This user has been on a crusade against Wikidata for quite some time, this is just another sad chapter. Multichill (talk) 23:03, 31 October 2017 (UTC)
Can someone explain to me what the enwiki folks mean with refspamming in this context? Isn't it good when all data that's provided is backed up by references? ChristianKl (talk) 13:50, 1 November 2017 (UTC)
They say the references themselves are spam (one could also take the results directly from the website of the J-League, it is just nobody volunteered to run such a bot). Also apparently the bot-owner came to the English Wikipedia and made a template importing these references from Wikidata (and was indefblocked).--Ymblanter (talk) 13:56, 1 November 2017 (UTC)
I guess en:WP:REFSPAM explains it. Although we don’t know for sure, this case is very likely refspam in the sense described on that page, and not acceptable here at Wikidata as well. For this reason User:Ymblanter handled the situation correctly by removing all affected references and even by deleting a couple of items created by the bot. However, it is sad to see that some enwiki editors escalate this situation to something big, although this could have easily been solved here without much trouble. —MisterSynergy (talk) 14:09, 1 November 2017 (UTC)
If the references point to the sales page for cryptocoins than I understand the idea of calling it refspam. On the other hand, if the references point to the page that lists the statistics, I don't see the reason to call it refspam. It sounds to me like some enwiki people consider do consider it to be illegitimate. I.e. items like Q38073048. With that item I see the problem that it doesn't have instance of (P31). It also problematic that it doesn't link to the football team in question or references the year. But I see no problem with the references of the items when they point to https://stats.scorum.com/football/tourneys/1103-2/j-league/season/2017 ChristianKl (talk) 16:49, 1 November 2017 (UTC)
As far as I know we don’t really know what happpened to that website. We just see that a fresh user added a couple of seemingly fine weblinks to Wikidata, and that some of them later turned out to be bad. Whether the bot operator personally or someone else in their vicinity has intentionally made them bad—we don’t know that for sure, but chances are high that this was not an accident on Scorum’s side and thus intentional refspam from the very beginning. Anyway, experience shows that very likely something fishy is/was going on here, and we should better consider to have been tricked by the bot operator and/or the organization their work for. Looking at individual items or links is not helpful for a discussion about the situation in my opinion. If at all we should consider to remove all remaining Scorum links, even if the data still looks healthy at this time. —MisterSynergy (talk) 17:04, 1 November 2017 (UTC)
It is probably better to replace them with links to the official website of the J-League. The site is here, though I can not immediately locate the results of 2017 games, and I am not sure how stable URLs are.--Ymblanter (talk) 17:15, 1 November 2017 (UTC)
I agree. Here are the remaining links via Special:LinkSearch. The other .me domain (which turned out to be malicious) is no longer linked to my knowledge. If I find time later this day or tomorrow, I will try to work on that task. —MisterSynergy (talk) 17:20, 1 November 2017 (UTC)
If you read the ArCom case there seem to be people who argue that even without the switch in content on the URL we shouldn't have approved the bot because of refspam. ChristianKl (talk) 17:20, 1 November 2017 (UTC)
There are lots of Wikipedia editors who find problems with anything which is related to Wikidata, although they obviously have little understanding how this project works. I wouldn’t worry about them, even if enwiki might soon decide not to use Wikidata in article namespace. It would be primarily their problem, not ours. —MisterSynergy (talk) 17:27, 1 November 2017 (UTC)
And there's the real problem. If enwiki decides not to use Wikidata in articles, then it's everybody's problem. It's enwiki's problem because we lose all the advantages of importing data from a central repository; it's other wikis' problem because we will lose a lot of the development work on integration that would otherwise take place on enwiki; and it's Wikidata's problem because we've failed an important part of our primary aim: to make the sum of human knowledge freely available to every person on the planet. You'll note that I say "we" in each case, not "they", because we're all part of the same community. If the Wikidata community doesn't focus more on the output of the resource (i.e, how it can be used), instead of being solely focused on input, we'll end up with the biggest unusable database ever invented. --RexxS (talk) 16:31, 6 November 2017 (UTC)
@Ymblanter: It seems to me like you approved the bot when the example items didn't have a instance of (P31), don't link to our items about the football clubs in question and the usage of the properties violates 3 constraints. Could you clarify whether you think items should be created under those circumstances by a bot? ChristianKl (talk) 17:20, 1 November 2017 (UTC)
Actually I think they have P31 which is Q27020041, and indeed I have a gadget which shows constraint violations, but it says the constraint violation is that the items should be in P31 which is a subclass of Q13406554. I believe this is an error, and Q27020041 must be a subclass of Q13406554. They also link to the clubs. Could you please give an example of an item which went wrong? I though I have checked all survived items yesterday.--Ymblanter (talk) 17:28, 1 November 2017 (UTC)
number of wins (P1355) has the constraint that it's supposed to be used as a constraint. In not very far into the sport modeling, but having it created that way suggests to me that it might be more worthwhile to not have a separate item for the session of each team, but store this kind of information in the item of the club in question. This seems to be a data-modeling question for the Wikidata project sports. While searching for it, that seems to be a bit strange and doesn't have a participant list. I think it would be good to ask   Notified participants of WikiProject Association football in a case like this. If there's a decision that number of wins (P1355) should be used as a top-level property the constraint can be removed.
Q38571489 seems to be one of the first demo items that was created and it still doesn't have instance of (P31) or links to other relevant Wikidata items like the team. Currently, all identifying information seems to be in the titel of the item which means it's useless for any kind of automated usage. ChristianKl (talk) 18:15, 1 November 2017 (UTC)
Indeed, you are right. I suggest that we delete all such items rather than try to repair them.--Ymblanter (talk) 18:21, 1 November 2017 (UTC)
I agree with deleting them given their current form. ChristianKl (talk) 19:10, 1 November 2017 (UTC)
Sports results in general are difficult to model in Wikidata. There is no established concept how to do it and to which level and precision it should be done, but some types of sports have enthusiastic editors who make the best out of the complicated situation. The method that was chosen here (season item per team) is very untypical, but I am not surprised about that. Maybe we should not only update the links, but also the model and merge all information into one season item. —MisterSynergy (talk) 18:27, 1 November 2017 (UTC)
Add: 1993–94 Cypriot First Division (Q2697925) is an example item how this information could better be modeled. One item per season, not per team and season. —MisterSynergy (talk) 18:35, 1 November 2017 (UTC)
Yes, I would prefer it to be modeled like 1993–94 Cypriot First Division (Q2697925) and it would likely to make sense for a sport interested Wikidatian to write this down in some Wikiproject. Either in the football one or in a more general sports one. Having a common data-model will make it much easier for data reusers. ChristianKl (talk) 19:09, 1 November 2017 (UTC)
I have tried and tried and tried to have manual pages for football (or sports in general) but we always have problems to agree for basic things like what "is" a season... If someone want to create that pages, he will have my help. This may help Wikidata:WikiProject Association football/Discussion about properties. Xaris333 (talk) 19:23, 1 November 2017 (UTC)

@ScorumME, ScorumMEBot: Being banned on the English Wikipedia doesn't mean that your bot account has to be banned here and the account ScorumME is still unbanned here (Wikidata and English Wikipedia make separate decisions about banning). If you want your bot to get unbanned, can you give us your perspective of what happened here? ChristianKl (talk) 18:21, 1 November 2017 (UTC)

I actually unbanned the bot when I took the flag away.--Ymblanter (talk) 18:22, 1 November 2017 (UTC)

  Info I proposed to delete all remaining 19 items, see Wikidata:Requests for deletions#Bulk deletion request: Japanese Football Championship 2017 items. —MisterSynergy (talk) 12:24, 2 November 2017 (UTC)

number of English Wikipedia items on Wikidata with and without description in English

Are there statistics available on either the number of English Wikipedia items and the number of them which have descriptions in English, or the percentage that have descriptions in English. Approximate value will be sufficient. Cheers, Pbsouthwood (talk) 10:54, 2 November 2017 (UTC)

Does anybody know the answer? I am pretty sure I have seen statements like "56% of Wikidata items have descriptions" but I have zero ideas where I should look them up.--Ymblanter (talk) 16:55, 2 November 2017 (UTC)
There are lots of canned statistics linked from Wikidata:Statistics but I haven't located these specific ones. I'll look around. I think SPARQL queries for these would likely time out but we could try that approach too. ArthurPSmith (talk) 19:07, 2 November 2017 (UTC)

The numbers are too big to be queried with the query service. Some older statistics can be found at User:ValterVBot/Labels and descriptions (Feb 2015), User:Pasleim/Language statistics for items and User:Pasleim/Sitelink statistics (both fresh, but not exactly what was asked here) and likely some others at Wikidata:Database reports. Since I did not find the exact requested numbers, I downloaded two data sets with SQL at Toolforge and compared them locally on my machine:

  • 7.415.123 items with enwiki sitelink
  • 26.893.359 items with English description
  • 5.176.179 items with enwiki sitelink and English description (i.e. ~69.8% of all items with enwiki sitelink)

Interestingly, the penetration depth of English descriptions of entire Wikidata (~38.1M items) is at ~70.6% and thus very comparable to the number that only involved items with enwiki sitelinks. HTH, MisterSynergy (talk) 06:06, 3 November 2017 (UTC)

Thank you all for your responses. I will look at the data and see if I can work out what it all means. Cheers,Pbsouthwood (talk) 11:30, 5 November 2017 (UTC)
Am I correct in my understanding that 7.4 million wikidata items with enwiki sitelink are combined articles and other namespace pages on enwiki which are considered valid data items, such as for example policy and help pages, and that there is no practical way of estimating the accuracy and usefulness of the page descriptions?Pbsouthwood (talk) 15:56, 5 November 2017 (UTC)
@Pbsouthwood: Yes, items from all enwiki namespaces are included here. You asked for the “number of English Wikipedia items”, so I saw no requirement to limit the query to particular namespaces. However, I could try to evaluate this for main namespace only as well if you tell me that such an estimation would be more useful for you.
Regarding “accuracy and usefulness of the page descriptions”: by which measure do you want to estimate the quality of descriptions? I am not sure whether this is doable at all, but maybe you could give further input what you think about… Thanks, —MisterSynergy (talk) 16:17, 5 November 2017 (UTC)

Okay, I went ahead and assume that you are interested in ns0 data only, @Pbsouthwood. I excluded all ~86.5k linked enwiki redirects from the evaluation, but they in fact don’t really influence the results much. Numbers are:

  • 5.486.860 items with enwiki sitelink to ns0 (excluding the ~86.5k enwiki redirects in ns0); this means that 99.7% of enwiki articles are connected to Wikidata
  • 27.044.945 items with English descriptions (+151.586 within three days)
  • 3.309.337 items meet both conditions (enwiki sitelinks in ns0 and English description)

So the ratio for the article namespace is 60.3%. —MisterSynergy (talk) 17:06, 5 November 2017 (UTC)

Thanks MisterSynergy, That is more useful. I didn't originally ask because I just didn't think of making the distinction. I doubt whether there is a way to judge accuracy and usefulness other than by inspection by a person with some understanding of the topic. Many of the current descriptions that I have inspected are not particularly useful to identify whether the article is one which the reader was searching for. For example, the description "organization" may be true but does not necessarily help much, particularly when the article title is "Technical Diving International", whereas the description "diver training and certification agency" will be more useful. This will usually depend on the person who writes the description, and how much effort they put into reading the article first. This would usually be done best by the editors of the article, not by a relatively random data accumulator, so ideally the articles should have short descriptions on the home wiki which Wikidata can collect, rather than the other way round. Maybe some day there will be software that can do this, bu I don't think yet.Pbsouthwood (talk) 18:51, 5 November 2017 (UTC)
@Pbsouthwood: As far as I know the descriptions were originally not invented for what they are used right now in Wikipedia apps and Wikipedia search function. The Wikidata-centric purpose of them is to disambiguate items with identical labels to human editors. That’s why there cannot be two items with identical label and identical description in a given language at the same time (ignoring empty labels and descriptions, of course; this means that label+description pairs are always unique per language). You can find information and rules about how to use descriptions at Help:Description. By for not all of them are perfect, of course, some of them are even problematic, but in general descriptions are not regarded as a problematic field in Wikidata given the purpose they have within this project.
At some state someone recognized that the growing amount of descriptions in Wikidata can be considered useful in some sense for Wikipedia as well (Apps, search, etc.) and this information was included outside of Wikidata. To my experience the descriptions do indeed outperform any comparable information from Wikipedias that was available before, so that this extra use of descriptions is fully legitimate in my opinion. However, this does of course create potential for conflicts between different Wikimedia projects when descriptions are suddenly meant to serve possibly contradicting purposes. It hasn’t been a big problem until now, though, and I am personally very optimistic that descriptions can indeed be used for different purposes in different projects.
Exchange of descriptions between Wikidata and Wikipedias would be rather simple. We use tools a lot, and crawling Wikipedia articles for template information to import it to Wikidata is a very simple process. The other way round is also possible, yet not as automatable as imports to Wikidata.
A last personal comment on the Enwiki—Wikidata relationship trouble that is going on mainly in Enwiki these days: I have experienced Wikidata as a very open and welcoming community with much less hierarchy , bureaucracy and complications than large Wikipedias (I know this mainly from comparisons to German Wikipedia, but it applies to other large language editions as well as I learnt meanwhile). We would be very open to enwiki editors that want to improve, say, English descriptions. Project boundaries do not mean much for many Wikidata editors, since many of us naturally have a strong cross-wiki edit behavior. We thus more feel as part of the one and only Wikimedia community of volunteer editors, i.e. part of the same community as Wikipedia editors. So there is not much “them and us” to my opinion which would make it difficult for newbies and Wikipedians to integrate here… —MisterSynergy (talk) 19:41, 5 November 2017 (UTC)
What you say rings true to me. I have no fight with Wikidata when WMF chooses to use this content. It is their (WMF) responsibility to use the material appropriately, and I think they overstepped the mark in this case, which is not good for Wikipedia or Wikidata and is causing unnecessary resentment among Wikipedians, who have had this display hidden from them, and no direct facility to make corrections where necessary. The big issue here is that the projects must be allowed to choose when and how they use Wikidata, It must never be imposed on them from outside. Cheers, Pbsouthwood (talk) 11:12, 6 November 2017 (UTC)

Hello everyone

Hi all,nice meeting you all,i am new in the forum and to WIKIDATA,but i am interested to organize a workshop in my community Buea,Cameroon.I will need advice from you all. Thanks  – The preceding unsigned comment was added by Linus Ayangwoh Embe (talk • contribs) at 3. 11. 2017, 08:21‎ (UTC).

This presentation can help you. However, you are new to Wikidata and want to organize a workshop? Matěj Suchánek (talk) 17:00, 3 November 2017 (UTC)
As an addition, go to https://query.wikidata.org and play with the examples for a few days until you understand most of the queries. That will give you some insights about Wikidata as a consumer, and you can talk about that in your workshop. Cheers! Syced (talk) 07:30, 6 November 2017 (UTC)

Items that are defined as properties ?

see Q33073144 and also Wikidata property for New York Times semantic concepts (Q23673786), Q30570134, Q42298191, Q42399821... found through Wikidata:WikiProject Properties/Reports/Identifier properties without subject item !!

How can Q-items be properties ? are those errors in creation or complete nonsense ? --Hsarrazin (talk) 13:03, 3 November 2017 (UTC)

Not all users are aware of the property proposal process and think that they can create properties that way. We should delete these items and help the creating users to find their way to Wikidata:Property proposal. —MisterSynergy (talk) 13:19, 3 November 2017 (UTC)
@JakobVoss: You seem to moved Wikidata property for New York Times semantic concepts (Q23673786) to be instance of (P31) Wikidata property for authority control (Q18614948). Can you explain the point? ChristianKl (talk) 13:59, 3 November 2017 (UTC)
@Hasive: It seems like you created an item instead of writing a property proposal. Do you need help with doing so? ChristianKl (talk) 14:03, 3 November 2017 (UTC)
I deleted the items created by IP/ accounts for which those were the only activities and pinged people involved with the others. ChristianKl (talk) 14:05, 3 November 2017 (UTC)
Hi ChristianKl, Thanks and sorry! It was my mistake to created an item instead of writing a request. BTW, a property proposal created already here: Bangladesh_Parliament_ID --Hasive • talk • 09:18, 4 November 2017 (UTC)
Okay, I deleted the item. ChristianKl (talk) 11:50, 4 November 2017 (UTC)
@ChristianKl: Please list these deleted items by label, so that we can see what they were about, and make sense of this discussion. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:21, 6 November 2017 (UTC)

Person, Jr. vs. Person Jr.

Wikipedia is harmonizing on "Person Jr." instead of "Person, Jr.", should we too? --Richard Arthur Norton (1958- ) (talk) 13:34, 3 November 2017 (UTC)

I feel that discussion would be better at Wikidata:WikiProject_Names. ChristianKl (talk) 14:08, 3 November 2017 (UTC)
This feels like something where harmonisation (rather than reflecting local practice) is perhaps going to be a bit confusing - I'm surprised WP are doing it. Andrew Gray (talk) 20:49, 3 November 2017 (UTC)
One reason I was against it at the vote in Wikipedia was because the other projects are not going to do it, so Commons, Quote, Data and the other projects are not following it. As people add in names in the body of articles they still add them under the old convention and leads to inconsistency and constant patrolling. --Richard Arthur Norton (1958- ) (talk) 16:35, 4 November 2017 (UTC)
@Richard Arthur Norton (1958- ): There are ~300 Wikipedias. Do you mean English Wikipedia? Please link to the relevant discussion/ policy. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:18, 6 November 2017 (UTC)
w:Wikipedia:Manual_of_Style/Biographies#Generational_and_regnal_suffixes spearheaded by w:User:Dicklyon. "Omission of the comma before Jr., Jr, or Jnr, and Sr., Sr, or Snr, is preferred. The comma can be used in cases where it is clearly and consistently preferred for a particular subject in current, reliable sources (most likely a living subject whose own preference is clear and consistent). Articles should be internally consistent in either omission or use of the comma for any given person's name.." The vote was against allowing a bot to change all, so they are doing one at a time manually. The key to changing all is the word "current", it isn't based on a simple Google count of "Person, Jr." vs. "Person Jr." --Richard Arthur Norton (1958- ) (talk) 14:06, 6 November 2017 (UTC)

Heads-up: Mass creation of ORCID authors, and linking to them from publications

@Magnus Manske: Somewhere I read that they create ISNI based on these. If this is the case, you might want to cross-check if we already have items for these people in the VIAF import.
--- Jura 07:08, 6 November 2017 (UTC)

Bot-added coordinates precision

I don't know if this has been discussed before, but there seem to be at least some bot-added coordinate location (P625) values with custom precision "0.0097195045176106" or similar like this one. Is this normal; should they be changed to the more usual ones like 1', 1", 0.0001°? (I changed that one to 1".) Jc86035 (talk) 14:54, 5 November 2017 (UTC)

Inputing this data via the API isn't trivial because the API sometimes converts the data into floating point numbers if it's done wrongly. Magnus's QuickStatement2 still does it wrongly. It would make sense to alert the owner of the bot that his bot might be malfunctioning. ChristianKl (talk) 15:00, 5 November 2017 (UTC)
@ChristianKl: The edit linked was made a few years ago, but pinging Multichill. Is there a way to query all items which have custom coordinates precision (has someone done it already)? Jc86035 (talk) 15:30, 5 November 2017 (UTC)
Not so easy...If you're adding coordinates via pywikibot (so - more or less - with API) and to get "normal" coordinates (seconds doesn't have any decimal numbers), not "awful ones" (seconds with very much decimal numbers), you need to set precision "0.00027777777777778". --Edgars2007 (talk) 08:19, 6 November 2017 (UTC)
It is unfortunately not well-defined what the precision of a coordinate exactly should express, thus different approaches have been taken. In the last contribution of Wikidata:Project chat/Archive/2017/10#Geographic coordinates: precision and import tools (after the outdent) I have addressed this problem, and there are also links to past discussions. —MisterSynergy (talk) 08:32, 6 November 2017 (UTC)

Wikidata weekly summary #285

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Discussions
  • Other Noteworthy Stuff
    • If you want to review what happened during the WikidataCon, here is the list of all the sessions, with slides, notes and videos, when available
    • Release of wikidata-edit v2.0.0: brings qualifiers and references support, and a few breaking changes. See changelogs
    • Release of wikidata-cli v5.0.0: brings qualifiers and references support, multi-entity summary|label|description|description requests, a new `edit-item` command, and a few breaking changes. See changelogs

You can see all open tickets related to Wikidata here.


Merge?

Can someone confirm that I should merge racing automobile driver (Q15958185) and racing driver (Q378622). The merge does not work, and I do not see what the error is, the Wikipedia articles have no overlap. --Richard Arthur Norton (1958- ) (talk) 16:06, 6 November 2017 (UTC)

One is subclass of the other one, the former is more specific than the latter (racecars × vehicles). Matěj Suchánek (talk) 17:33, 6 November 2017 (UTC)
Other riders (for example, on water) does not happen? What is the name of the racer (on water)? --Fractaler (talk) 17:46, 6 November 2017 (UTC)
I would either change the description or the label of racing driver (Q378622) since "racing drivers" could apply to any driver of a vehicle in racing, but the description specifies "motorsport vehicle". Not all racing drivers drive a motor vehicle, there are non motorized vehicle races. Amqui (talk) 18:09, 6 November 2017 (UTC)
+ to WD the item "Non motorized vehicle race"? --Fractaler (talk) 18:14, 6 November 2017 (UTC)

Properties created without formal discussion

What should be the value for the properties created without formal discussion? For example, currency (P38), head of government (P6) have values for property proposal discussion (P3254) as 'no value'. But much recently created properties like rating (P4271) have no related statements. What's the best option? ':o value' or not creating any such statement? I personally feel that specifying 'no value' is a better way. Refer Property_talk:P3254 John Samuel 13:02, 5 November 2017 (UTC)

  • I think there was a discussion for this property. @ChristianKl: probably just deleted the subpage name when creating the property talk page (and forgot a couple of other things they should have done). Maybe they wants to revisit their creation and finish it.
    --- Jura 13:18, 5 November 2017 (UTC)
  • I didn't delete anything. User:Andrawaag created the property proposal and didn't have the subpage information from the start. Various people have supported the proposal in the form it existed without the "|subpage = rating" and when @Jura1: commented on the proposal to express the need for a description they ignored the need for this information. A way to fix this problem from repeating would be to change the property proposal template in a way that makes it visible when the "|subpage = ?" information is missing. ChristianKl (talk) 13:33, 5 November 2017 (UTC)
  • The rating (P4271) property meanwhile has a P3254 claim.
  • There are 99 properties without any property proposal discussion (P3254) claim: 
    SELECT ?property ?propertyLabel WHERE {
  ?property a wikibase:Property . # all properties
  MINUS { ?property rdf:type wdno:P3254 } # remove those with "property proposal discussion: no value"
  MINUS { ?property wdt:P3254 ?discussion } # remove those with any other "property proposal discussion" claim
  SERVICE wikibase:label { bd:serviceParam wikibase:language '[AUTO_LANGUAGE],en' }
}
    Try it!
    The oldest is BNMM authority ID (P3788), where it was for some reason removed; the second oldest is narrower external class (P3950) (created on 14 May 2017).
  • On 12 May 2017 I ran an automatic QuickStatements job to add the P3254 value for all existing properties, based on the value found in the {{Property documentation}} template on the property talk page.
  • no value is indeed the right way to do it if there was no discussion at all.
  • We have 60 properties with P3254:no value right now (query). Very likely most of the values have been added by me at around 12 May 2017 as well. —MisterSynergy (talk) 13:43, 5 November 2017 (UTC)
  • Generally, I believe that the task of copying over the information from the template to a statement should be done by a regular running bot. In the absence of a bot that copies over a claims I don't think the manual task of copying over the statements should be a constraints that limits the creation of fully supported new properties. ChristianKl (talk) 14:14, 5 November 2017 (UTC)
  • The principle of a Wiki is that everybody does the work he considers to be important. If you consider that work important to be done, feel free to do it. I personally consider it important to work on keeping the time between the creation of a property proposal and it getting support and it's creation to be small and like to invest my energy into that task. I consider that to be important because when properties aren't created that are needed for people to do projects they want to do, it discourages them from editing Wikidata. Theory of constraints therefore lets me focus my energies on that task. ChristianKl (talk) 14:33, 5 November 2017 (UTC)
  • I think I told you the same the last time. Copying over the information from the template creates enough information to make the property useable. Since the last time we have SPAQRL queries listed in Wikidata project properties for people who want to get various things filled. At the moment it's hard to know exactly what work is done by bot. MatSuBot does some work but mostly that work is done by human editors before they bot actually gets around to it so I didn't get around to documenting it's behavior at Wikiproject properties. ChristianKl (talk) 14:50, 5 November 2017 (UTC)
    • Apparently people conclude that ChristianKl creates properties without prior discussion. It appears that the documentation you add to talk pages is incomplete. If you leave the property otherwise incomplete and the original proposer doesn't complete it, it tends to get messy. So, you might want to revisit your property creation and complete as needed.
      --- Jura 17:29, 5 November 2017 (UTC)
  • I add most of the missing "|subpage = ?" statements when I notice that they are missing. In this case, there's a property where I didn't notice it. There are two ways to focus attention. One is the lean manufacturing way where you ask how you can fix the problem on a systemic level. In this case, this perspective suggests changing the property proposal template in a way to indicate that "|subpage = ?" is missing. The other is about fixing errors individually when they appear. In dealing with this particular case I edit the "|subpage = ?" into the rating property and I create a request to change the template.
I don't think anything that happens in this discussion is "mess" in a negative way. Creating situations where people think "here's something that can be improved on Wikidata" is healthy for the project. This discussion might be an effect of me creating the Wikidata project properties with the list of missing property proposal discussion (P3254).
Where there's life and activity things are messy. Creating the appearence of everything being sterile on the other hand doesn't encourage action. ChristianKl (talk) 18:08, 5 November 2017 (UTC)
There are a series of steps to be done when creating them. Many are essential, some not. If you systematically ignore them and expect others to clean up after you, there isn't really any thing healthy about it.
--- Jura 07:05, 6 November 2017 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ @MisterSynergy: I'm pretty sure Canadian Register of Historic Places ID (P477) don't have a discussion for is creation eater. --Fralambert (talk) 14:41, 5 November 2017 (UTC)

  Info there were some 60+ cases remaining which I all solved for now. Since I even found some properties without any statements at all (not even P31), I’d like to express the wish towards property proposers and property creators that they spend more effort into a proper setup at the very beginning. Property creation is an activity that requires special knowledge and rights, and we do it roughly two times a day. This really needs to be done properly and carefully in order to avoid poor data at a later point. —MisterSynergy (talk) 18:29, 5 November 2017 (UTC)

Thanks MisterSynergy. A number of properties lack vital information and need to be further improved. John Samuel 18:41, 5 November 2017 (UTC)
You're welcome at Wikidata:WikiProject Properties! Please add SPARQL queries to detect poorly documented properties at Wikidata:WikiProject_Properties/Reports so we can spot them and improvide property documentation. -- JakobVoss (talk) 09:41, 6 November 2017 (UTC)
  • @MisterSynergy: What kind of scenario happened where poor data was caused at a later point? If you list examples where damage is done, I'm happy to think about how to change processes to reduce/eliminate damage.
I find the number of 2/day curious. The weeks where that was the status quo, I think I didn't create any and backlog was building up. Currently, I'm relatively happy with our process given that the creation backlog shrinks over time.
When it comes to the task under debate at this discussion of copying over property proposal discussion (P3254) I don't see harm caused when it isn't copied over soon after property creation and I see no harm in letting the amount rise to the point where systematically solving the task with a bot makes sense. I especially don't see why doing so should be a blocking task.
Given that we have software support for constraints, I see a case to prioritise their creation higher. ChristianKl (talk) 13:58, 6 November 2017 (UTC)
I don’t have lists of examples at hand, but I have seen several properties in bad shape—particularly the identifier region P3000 and later. You typically find either only very little use of them, or very diverse use beyond the originally proposed property scope. Repairs are often possible, but not always simple or straightforward to perform. I thus prefer a clean definition of properties from the very beginning to at least an extent that an average Wikidata editor is able to use it properly within items, and data users know what to expect. Unfortunately there is no policy to enforce that. The number of 2/day was estimated by dividing ~4000 properties that we have at the moment by the ~2000 days of Wikidata age. —MisterSynergy (talk) 17:02, 6 November 2017 (UTC)
We might have been at 2/day before I started prioritizing the creation of new properties. Not creating properties in a reasonable amount of time was a constraint that reduced their creation. I invite you to participate when discussing new properties and being part of Wikidata:WikiProject Properties if you want to understand what's the status quo.
I agree that having a clean definition of a property before it gets created is good. Process-wise getting a clean definition is what property proposal discussion is about and currently copying that information to a new property happens every time. The problem seems that the property proposal dicussion doesn't contain all information that's needed for a "clean definition". A big thing is that it doesn't contain information about the constraints when that's part of a clean definition. Ideally, we should also have consensus about what constraints a newly created property should have after creation. We can agree on what information should exist and add corresponding fields to the property proposal template. ChristianKl (talk) 22:25, 6 November 2017 (UTC)

The Community Wishlist Survey 2017

Hey everyone,

The Community Wishlist Survey is the process when the Wikimedia communities decide what the Wikimedia Foundation Community Tech should work on over the next year.

The Community Tech team is focused on tools for experienced Wikimedia editors. You can post technical proposals from now until November 20. The communities will vote on the proposals between November 28 and December 12. You can read more on the 2017 wishlist survey page. /Johan (WMF) (talk) 20:16, 6 November 2017 (UTC)

Merge required

Q39719918 and Q133923 are the same taxon. Aloe aristata is the earlier name, Aristaloe aristata the newer one, accepted by recently updated sources, such as WCSP. So they need merging, which I have no idea how to do on Wikidata. Peter coxhead (talk) 10:26, 30 October 2017 (UTC)

This is not possible. The problem is at the Wikispecies end.--Ymblanter (talk) 14:16, 30 October 2017 (UTC)
I added basionym (P566) to Aristaloe aristata (Q39719918) to establish the relationship between this two names. --Succu (talk) 19:08, 30 October 2017 (UTC)
I find it very puzzling that a merge is not possible, given that Q39719918 and Q133923 are indisputably the same taxon. @Ymblanter: what is the problem with Wikispecies? This correctly has only one taxon, species:Aristaloe aristata, to which species:Aloe aristata redirects. Why must there be two Wikidata items for one taxon? Peter coxhead (talk) 21:29, 30 October 2017 (UTC)
I missed that it was a redirect. Now I removed the species item from Q133923, and attempt to merge them, but I keep getting an unspecified technical error which I hope someone here will comment on.--Ymblanter (talk) 21:40, 30 October 2017 (UTC)
@Ymblanter: Two items which link to each other can't be merged - you'll need to remove the links first. These two seem to be crosslinked with basionym (P566) on one and subject has role (P2868):basionym (P566) on the other. Andrew Gray (talk) 12:32, 31 October 2017 (UTC)
Thanks Andrew. I will leave it then to someone more competent in taxons than I am. All sitelinks are now in one item.--Ymblanter (talk) 12:41, 31 October 2017 (UTC)
  • NO, a merge is not required. Having a link between different taxons and linking them through a basionym is the right way to do it. Each taxon is linked to a set of literature and for a correct understanding of that literature the correct description is the original description. From a Wikipedia point of view moving to the later description is an option however it does not necessarily reflect the articles. Thanks, GerardM (talk) 06:19, 1 November 2017 (UTC)
@GerardM: but then the entries in Wikidata are wrong to say "instance of: taxon". There's only one taxon in this example, with two alternative scientific names under consideration. The correct entity relationship diagram underlying this part of Wikidata is not reflected in the way it has been set up. Articles in wikis, Wikispecies entries, and Commons categories and galleries are about taxa. They need to be linked together via a Wikidata instance of a taxon. Names are different entities, and, as you say, have their own literature, types (if heterotypic synonyms), entries in IPNI, etc. Taxa and names have a 1:N relationship.
If Q39719918 and Q133923 are kept separate, as different names although the same taxon, then presumably w:en:Aristaloe should be linked to Q39719918 and w:de:Aloe aristata should be linked to Q133923. But then they won't show up in each other's articles. To keep them linked to each other, we have to make an arbitrary choice of which name to use, and there's no guarantee that editors in different wikis will make the same choice. Peter coxhead (talk) 16:14, 7 November 2017 (UTC)
A definition: "the name applied to a taxonomic group in a formal system of nomenclature". So it is a name in a system. When a name is given it is given as part of a formal system. These names change as the system is formally overhauled. Your notion about articles is something that is not relevant as Wikidata is not there to serve Wikipedia. It is also there to allow for the linking to the literature that exists for all those names. Fundamentally there may be a link between these taxons but it is not true to say that the description is the same. It is not. We are not beholden to Wikipedia and your argument is exactly what prevents collaboration with Wikispecies. So no, your argument is wrong on many levels. Thanks, GerardM (talk) 17:16, 7 November 2017 (UTC)

dead language (Q45762) vs. extinct language (Q38058796)

I'm not sure, even the simply merging is not possible, why on earth both can't be linked with said to be the same as (P460) but instead (?) different from (P1889), where are differents between Nynorsk nn:Utdøydd språk - nn:Dødt språk, and between Portuguese pt:Língua morta and pt:Língua extinta? --Liuxinyu970226 (talk) 12:12, 7 November 2017 (UTC)

It seems dead language (Q45762) to be still in usage, but not as native to anyone; and extinct language (Q38058796) fully out of any usage. - Kareyac (talk) 12:39, 7 November 2017 (UTC)

Nicolo' Figlia -> Nikollë Filja merge

How can some articles be merged into one set of wikidata? In partcular, can someone help me with the merging of Nikolle Filja (Albanian spelling) with Nicolò Figlia (Italian spelling)? Thanks! --1l2l3k (talk) 17:08, 7 November 2017 (UTC)

See Help:Merge. Matěj Suchánek (talk) 18:27, 7 November 2017 (UTC)

Q42882548 - self promotion ?

Looking at the data in the item, and the history and creator, this looks like self-promotion. Is it legit ? --Hsarrazin (talk) 00:14, 12 November 2017 (UTC)

@Hsarrazin: I'd guess it's self-promotion, but keeping it would depend on whether you consider an AllMusic identifier sufficient notability and authority control for Wikidata. Mahir256 (talk) 02:48, 12 November 2017 (UTC)
An identifier is in and of itself no reason to point to notability. We do support Twitter and Facebook too. Thanks, GerardM (talk) 07:30, 12 November 2017 (UTC)
Personally, I do not really care. I stumbled upon it through my weekly maintenance of items with sex or gender (P21) with no instance of (P31), and because the name seemed strange… ;) --Hsarrazin (talk) 10:07, 12 November 2017 (UTC)
  Deleted by Multichill (talk • contribs • logs). Matěj Suchánek (talk) 20:25, 13 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 20:25, 13 November 2017 (UTC)

Strange edit incorrectly attributed to me

This edit appeared in my contributions, and I didn't do it. Does anybody know what might account for it? Have I been hacked? Is there a bug? Thanks! --TimK MSI (talk) 17:44, 12 November 2017 (UTC)

You have tried Wikidata:Tours? --ValterVB (talk) 17:47, 12 November 2017 (UTC)
Haha, that was it, thank you! :-) TimK MSI (talk) 17:48, 12 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 20:25, 13 November 2017 (UTC)

Deleting item without discussion

In which cases does the administrator have the right to delete the item without any discussion (with the creator of the item, with other users related to the item, etc.)? --Fractaler (talk) 12:29, 29 October 2017 (UTC)

Many items are deleted without discussion if they have no statements and no sitelinks or backlinks. Maybe you have more info you can share on this? ArthurPSmith (talk) 13:47, 29 October 2017 (UTC)
If an admin think that an item isn't notable he/she can delete item wthout discussion. Naturally errors are always possible, but the fix of errors for obvious cases re very quickly. --ValterVB (talk) 13:57, 29 October 2017 (UTC)
@ValterVB: I think we are arguing about Fractaler's recent creations such as Q38791127 (if you scroll up you find a post with references three deleted items. ChristianKl (talk) 12:17, 30 October 2017 (UTC)
Sorry, more info: 1) if the item was discussed before (for example, on "Requests for deletions" and has not been removed)? Or 2) there were links to the items, and item referred to other items (but then all the links were deleted without any discussion by the administrator)? Should the creator of the item have the right to discuss the removal of the item he created? --Fractaler (talk) 14:18, 29 October 2017 (UTC)
maybe the admin wasn't aware of the previous discussion? If the links don't seem valid in some way I can see this happening. Can you be a little more specific? Or raise it on the admin noticeboard? Perhaps talk to the admin involved? ArthurPSmith (talk) 14:46, 29 October 2017 (UTC)
I thought the administrator before deleting has full information about the item. Thanks for the idea, I'll try --Fractaler (talk) 15:03, 29 October 2017 (UTC)
Does the administrator, who behaves in this manner, have the right to ignore the request to join the discussion of the item he deleted? --Fractaler (talk) 14:15, 6 November 2017 (UTC)
I'll answer with a question: Is the administrator who deleted an item necessary for reviewing its deletion? (The answer is no.) Matěj Suchánek (talk) 17:39, 6 November 2017 (UTC)

Rights of item and rights of item's creator

Does the creator of the item have the right to prove the rights to the existence of the item (created by him)? Are there any Wikidata's pages on this subject? --Fractaler (talk) 09:40, 30 October 2017 (UTC)

No. The creator of the item should provide sources about the existence of the item. Obviously there is a point to asking for sources but there is an end to it. Thanks, GerardM (talk) 10:07, 30 October 2017 (UTC)
Sorry, I did not understand. Should provide sources where? --Fractaler (talk) 11:39, 30 October 2017 (UTC)
@Fractaler: When you create an item, you are supposed to provide references for the claims that you make on that item. ChristianKl (talk) 12:09, 30 October 2017 (UTC)
@ChristianKl: Thanks, I know about it. I mean the rights of item and rights of item's creator (do they exist, where they are listed, etc.). --Fractaler (talk) 13:16, 30 October 2017 (UTC)
There were no references on the items you recently created and that got deleted. ChristianKl (talk) 13:34, 30 October 2017 (UTC)
Items that were created without references were based on the WP, Wiktionary, school, university, etc. information and were obvious to me. But the speech in this topic about the rights, is not it? --Fractaler (talk) 14:00, 30 October 2017 (UTC)
The fact that you consider something obvious doesn't establish notability. Wikimedia projects do not provide good sources for our purposes. Instead of trying to write down what you consider to be obvious you could go and search for prior art where people invested nergy in figuring out how to model a domain.
There's no policy document or other consensus that declares such a right and given the amount of items that get deleted those admins who do the bulk of the deletion work would see such a right as a problem for their work. ChristianKl (talk) 16:26, 30 October 2017 (UTC)
Copyright applies here, but in practice the effect of that is very limited:
  1. Any edit is licensed under a pretty open license.
  2. The creative aspects of any one statement are quite limited; only a bunch of edits together may be of some significance.
Brya (talk) 17:31, 30 October 2017 (UTC)
Data of the Wikidata are data for internal, external, internal and external use? If an item is deleted, the rights of users (internal, external, internal and external) that use this item are violated in this case?
A bunch of edits is possible for an existing item.
So:
  1. policy document or other consensus - there's no.
  2. policy document should indicate: item's creator has no rights to protect the right of existence of the item created by him. Cause: given the amount of items that get deleted those admins who do the bulk of the deletion work see such a right as a problem for their work.

All right? --Fractaler (talk) 08:06, 31 October 2017 (UTC)

No. Like anyone else the creator of the item has the right to put the information in its own database and publish it independently from Wikidata. Managing data outside of Wikidata is common practice so if there is no consensus here, just store it anywhere else. -- JakobVoss (talk) 19:48, 7 November 2017 (UTC)
Not copyright, but right of Wikidata user. The Wikidata creator of the Wikidata item has the right (has the opportunity/capability) to put the information in Wikidata. Then, The Wikidata creator of the Wikidata item has not the right (has not the opportunity/capability) to protect the Wikidata item (in the Wikidata, not outside the Wikidata)? --Fractaler (talk) 06:27, 8 November 2017 (UTC)

Ghostwriting academic and academic papers written by ghostwriters

Some academics allow their reputation to be used for publications that are paid for and are as a consequence not trustworthy. I have created a qualifier to Martin Keller obviously it needs a reference. The paper(s) needs a similar statement. Please advise. Thanks, GerardM (talk) 10:14, 3 November 2017 (UTC)

I don't think "significant event" is an good way to model this. If he's ghostwriting specific papers, we need a way to mark him in relation to those papers. ChristianKl (talk) 14:07, 3 November 2017 (UTC)
The point is that it casts doubt on all the papers produced. So it is certainly significant. The known papers involved need to be marked as such as well. What do you propose. NB I do not care for a negative advise, I want to known how to move forward. Changing existing data is easy. Thanks, GerardM (talk) 15:04, 3 November 2017 (UTC)
An event is something that happens at a point in time, I don't think this example fits. Futhermore, the mission of Wikidata isn't to critize individual people and do activism against individuals. Labeling people as engaging in "malpractice" can become easily become libelous and we don't have good processes to do quality control about it. ChristianKl (talk) 16:43, 3 November 2017 (UTC)
Publishing a paper IS an event. At the time it was written by ghostwriters.. There is a source and the negative impact of the research is documented. Now when we do not criticize people, we cannot write about Mr Weinstein either. But even worse, we cannot truthfully report on the many papers that are imported in Wikidata. That would be absolutely awful absolutely worse. Thanks, GerardM (talk) 17:55, 3 November 2017 (UTC)
Currently, there's no statement on the item for Weinstein that includes any allegations against him. If you think we should handle Weinstein differently than we are currently doing that's another discussion.
When it comes to the ghostwritten papers, I agree that it makes sense to have a way to represent their authorship situation. I however don't think that we have to apply any judgemental terms such as "malpractice" to the case. The information might be marked in a qualifier for "author". Wikidata talk:WikiProject Source MetaData would be the right forum to discuss which qualifier might be appropriate.
It's also factual wrong to say that Martin Keller engaged here in ghostwritting. He appearently didn't do any writting. We generally don't have "published a paper"-events and don't think we should. You also didn't subclass any "publish a paper" event class. ChristianKl (talk) 21:29, 3 November 2017 (UTC)
When Wikidata is only a stamp collection of bio-medical and other "scientific" publications we have a severe problem. As it is information from Wikidata is used on the use of chemical substances for medicinal use. They often are not fit for use as is the case of paroxetine in juveniles. Your formalistic attitude means that because of "technical" considerations you remove data. Data that can easily be changed after a discussion. Without the ability to clearly state that an FDA approved substance is proven to be dangerous based on research and is not allowed in other countries we cannot have this stamp collection. So when you move on "technical" considerations, I move to remove all bio medical literature because its connection to Wikipedia makes it dangerous to have. Thanks, GerardM (talk) 06:30, 5 November 2017 (UTC)
Feel free to fill request for deletions for various biomedical scientific papers that the Wikicite people ingest on the grounds that they are dangerous if you think that will do anything besides embarrising yourself.
If you want to ingest information than either ingest it within the realm of the properties we have within the constraints for which the properties got created or create a property proposal for a new property. Bending an existing property and violating it's constraints isn't the way to go. ChristianKl (talk) 12:22, 5 November 2017 (UTC)
While it appears that it would make sense to mark ghostwritten papers, does it make sense to mark the writer himself? That would like saying a person is a rapist/murderer/drug addict/robber/criminal/rebel in wikidata property which are not necessarily true even when people brought up evidence for that, controversial, and can be unrepresentative of the human.C933103 (talk) 08:51, 5 November 2017 (UTC)
Absolutely. People died because of the effects of their use of paroxetine that was sanctioned because of this study. So yes, this label fits among the labels you used: "rapist, murderer, drug addict, robber, criminal, rebel". NB rebel is an ill fit in this list. Yes, these people need to be stigmatised for their bad behaviour. They are what is wrong in science. Thanks, GerardM (talk) 10:13, 5 November 2017 (UTC)
Part of Wikidata core principles is being neutral. Engaging in stigmatizing people for behavior you don't like is to leave the role of being neutral. If you want to violate core principles of the project my objections are far more than just technical. ChristianKl (talk) 12:48, 5 November 2017 (UTC)
There is no doubt about the facts and consequently "neutrality" is not an issue. Words like criminal, murderer we use in Wikidata. These people are known for what they have done. There is no difference. The stigma attached is the same stigma Mr Weiner has to suffer from. He has not been convicted yet, the process in this case is done. Thanks, GerardM (talk) 13:27, 5 November 2017 (UTC)
I don't think there are label for e.g. rapist in wikidata. There are label for those who committed the crime of rape but that is not calling the people rapist, and also those who get labelled as committed the crime of rape or other crimes under convicted of (P1399) are usually those that have been trialed as far as I understand which doesn't seems to be the case here.C933103 (talk) 14:27, 5 November 2017 (UTC)
How about use participant in (P1344) or position held (P39) together with medical ghostwriter (Q6806538) if there are enough evidence to support such behavior existed? But then the problem is there are apparently people, including authors themselves, are still trying to claim it is not ghostwriting. Is there something like BLP guideline in wikidata? C933103 (talk) 14:38, 5 November 2017 (UTC)
The act of ghostwriting is about writing a document without attaching your name to it. In this case the charge is that he added his name to the paper but didn't write it. To the extend that someone in this case was engaging in ghostwritting it's not Keller. ChristianKl (talk) 17:15, 5 November 2017 (UTC)
ghostwriter (Q623386) is about "writing a document without attaching your name to it", however, according to explanation on English Wikipedia linked entry, medical ghostwriter (Q6806538) also include some other acts that does not fit with this description. C933103 (talk) 17:40, 5 November 2017 (UTC)
ITYM "ghost-written academic" (a ghostwriting person would be a person who works as a ghostwriter), and have modified ghostwritten academic (Q42590222) accordingly. I have also applied it to Martin Keller (Q6775864) using occupation (P106). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:19, 6 November 2017 (UTC)
Enwiki entry for medical ghostwriter (Q6806538) also used the term "guest-authoring" to describe the situation.C933103 (talk) 05:36, 7 November 2017 (UTC)
"Guest authoring" does not do this malpractice justice. It is a form of fraud. Thanks, GerardM (talk) 09:40, 7 November 2017 (UTC)
As the article describe, there are some people who think it is not a problem.C933103 (talk) 21:02, 7 November 2017 (UTC)
In that case they should have no problem when their unproblematic practice is identified. Thanks, GerardM (talk) 06:41, 8 November 2017 (UTC)

Which property to indicate in what electoral district a city is located?

Are electoral districts considered as "administrative territorial entity" or is there another property, or should we create another one? Thanks, Amqui (talk) 17:30, 6 November 2017 (UTC)

Why don't we just create a specific one then or re-scope the one existing "electoral district" property that right now is only used to show which electoral district someone represents? Note that in a lot of countries this would need more than one, for example in Canada with provincial and federal electoral districts, but that could be in the same property with proper qualifiers. Amqui (talk) 17:48, 6 November 2017 (UTC)
This could get a bit messy - potentially you'd have to have it on every populated place in a country with complex boundaries like the US. I've been trying to come up with a good solution to this for the UK (here electoral districts are mostly made up of smaller admin entities, but not always perfectly corresponding) and it's really difficult to do. I think we might end up having to wait until we can define the district as a shapefile and search for all points inside that shapefile, or something. Andrew Gray (talk) 20:39, 7 November 2017 (UTC)

Also I don't know how to include the following statistics and information about a city that are all easily structured data:

  • Median age of the population
  • Number of private dwellings (could use "has part: dwellings with quantity as qualifier")
  • Mother languages by percentage of the population
  • Languages known by percentage of the population
  • The level of education by percentage of population (i.e. 25% with highschool diploma, 35% with university diploma...)
  • The domain of studies by percentage of population
  • The location of studies by percentage of population
  • The religions of the population in percentage (with proper references of course)
  • The ethnicities of the population in percentage (with proper references of course)
  • The major access highways/roads to the city, and the next cities on that road in each direction (that last part could be a qualifier of the first part)

Any ideas? Thanks, Amqui (talk) 17:48, 6 November 2017 (UTC)

Maybe for statistical data, we could have a generic "Statistics" property that we could use. For example "Statistics: Mother language with qualifiers Language and Quantity", but that's not pretty and it cancelled other existing properties such as Population and Unemployment Rate... Amqui (talk) 18:06, 6 November 2017 (UTC)

  Comment Can I ask a provocative question. Why would we wish to determine or specify this in our data sets? There are always going to be particular issues with such declarations. Certain cities are in multiple electorates, and depending on how you define cities they can contain multiple electorates. I can just see it getting messy and ugly. Example:

So which city are we talking about? On which date are we talking about that particular "version" of that city and that electorate? Are we talking Melbourne, the conceptual "greater" city as per the Wikipedia article, or the one prescribed by the local government act? Either way there is no exact match for city and electorate. There is no complete overlay that one will fit within the other. While the local government area remains (reasonably) static, the respective state and federal electorates change based on population changes, at least incrementally on a 10 year cycle. Want to see the complexity (at least for Australia), start at http://stat.abs.gov.au/itt/r.jsp?ABSMaps and look at the drop down, and see how the areas can be defined for 2016, 2011 and 2006.

With regard to statistics ... aren't we better off linking to the specific datasets that exist and contain that data rather than trying to bring such complex datasets into our system and overlay them against our people generated systems? We simply do not have a good standard to enforce an alignment of our place creations to government organisations. For Melbourne (greater) you have statistics like http://stat.abs.gov.au/itt/r.jsp?RegionSummary&region=2GMEL&dataset=ABS_REGIONAL_ASGS2016&geoconcept=ASGS_2016&measure=MEASURE&datasetASGS=ABS_REGIONAL_ASGS2016&datasetLGA=ABS_REGIONAL_LGA2016&regionLGA=LGA_2016&regionASGS=ASGS_2016 for a point of time, and that will differ significantly from 2011 and 2006, so putting that into an item is along the lines of "you want to do what?!?" Surely we don't want that level of data in our systems, well not curated by us.  — billinghurst sDrewth 22:52, 6 November 2017 (UTC)

OK I get your point about electoral districts.
About the statistical data set, if we want to be useful to Wikipedia in all languages at the same time, which in my opinion is one of the main function of Wikidata, then it makes sense to include those data here. You say it would be hard to curate the data on Wikidata because the change at every census, but isn't it even harder to have to curate them on every language versions of an article on Wikipedia? If I can change them at one place instead of 10 or 20, I say it is easier than harder. Amqui (talk) 23:44, 6 November 2017 (UTC)
That sounds like we need to get hold of dataset and see what is there and can be imported and generally explored and presented. Looking at the Australian stuff, I see geography over time explained and data packs. All beyond my me! :-)  — billinghurst sDrewth 10:44, 7 November 2017 (UTC)

My answers to the original question: Yes, maybe, probably. Jane023 (talk) 10:52, 7 November 2017 (UTC)

An unhappy birthday

Many of us celebrated Wikidata's recent 5th birthday, but two other birthdays recently passed with less fanfare:

have now both been open for more than a year. Can someone please put them out of their misery? The latter includes multiple proposals, all of which seem to meet the criteria for creation.

There are over 40 other property proposals that have been open for over 150 days; including several that have not been edited for over 100 days.

were created 147/148 days ago, and have not been edited since then. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:44, 7 November 2017 (UTC)

If there is no discussion and the proposal lacks information, I see no reason to push property creation. We have already too many properties that were created but have little been used since because nobody was interested in filling in the data. I'd ask for more details (examples, subject item, domain...) to check whether somebody is willing to invest more time in the property and take care making use of it. By the way: is there a list to check these old proposals? -- JakobVoss (talk)
Even if property creation takes a lot of time, the community conversation is very important. I periodically check newly proposed properties, but unlike generic (or creative work) properties, sometimes it's very difficult to give an opinion (support/oppose) on proposals because they are out of my domain of expertise. I am sure that many among us here do check periodically. What JakobVoss pointed is also correct. But, is the delay in property creation one of the reasons behind their ultimate low usage? We need to reflect on both these problems. John Samuel 20:18, 7 November 2017 (UTC)
Community conversation is indeed very important. But in many of these cases, like the latter two I cited, there is no discussion ongoing. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:18, 7 November 2017 (UTC)
Who said anything about "Pushing creation"? Proposals can as easily be closed as "not done". The list is at Wikidata:Property proposal/Overview and is updated daily by a bot. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:18, 7 November 2017 (UTC)
It is a little provocative to assign happiness or otherwise to a proposal that is moribund and unsupported. While it may be unfortunate they are still open, prescribing emotion is just asking for an argument. Presumably they have been left open due to uncertainty or lack of information. While a property creator can mark them as not done; similarly, any community member can suggest that they are not progressing and suggest they be closed as not done. So often it is more helpful to not just tell us the problem, but what are the reasonable solutions.  — billinghurst sDrewth 00:21, 8 November 2017 (UTC)

Taiwanese push car railways (Q381727) and handcar (Q3008463)

Q381727 have 4 articles in 4 different wikipedia; the entry in zhwiki and enwiki are about Taiwanese pushcar railways, however the entry ind enwiki and itwiki are about pushcar railway in general.

Q3008463 have 6 articles from 6 different wikipedia; the entry in zhwiki, enwiki, frwiki and nlwiki are about pushcar(handcar), but jawiki and kowiki entry there are about pushcar railway in general.

How to properly link these entries with each others? If a simplistic way to drop enwiki and itwiki from Q381727 and drop jawiki and kowiki from Q3008463 to create a new wiki entry then that would lost much interwiki linkability from the process.C933103 (talk) 18:26, 7 November 2017 (UTC)

@C933103: It more sounds like these just need to be fixed. We would need wikidata items for

  • Taiwanese pushcar railways
  • pushcar railway in general
  • pushcar(handcar)

and then we move the interwiki links as appropriate. The wikidata items need to be to have their "instance of" or "subclass of" allocated appropriately to show their hierarchy. [Noting that there are tools and gadgets that allow you to push interwikis, and data fields, between entities to make this easier.]  — billinghurst sDrewth 00:13, 8 November 2017 (UTC)

@billinghurst: I have created Taiwanese push car railways (Q42832982), but then there are a few problems, 1. how to facilitate interwiki links across items especially on wikis without the Template:Interwiki Extra ? 2. Also, I don't know which property I should use to link Q381727 to Q42832982 or vice versa, there are no property I could find that would represent the concept of "national variant of" or "regional variants being". C933103 (talk) 05:12, 8 November 2017 (UTC)

Locations of Würzburg radar (Q1849277)

Will there be any limitations using location (P276) for the different places Würzburg radar (Q1849277) was erected during WW2? Breg Pmt (talk) 20:37, 7 November 2017 (UTC)

Add qualifier point in time (P585) -- JakobVoss (talk) 20:54, 7 November 2017 (UTC)

Mayors of locations in New Jersey

I want to cluster all the Wikidata entries for mayors of locations in New Jersey like Mayor of Hoboken, New Jersey (Q6797722). Should I just create a new entry called "Mayors of locations in New Jersey" and make "has part" Mayor of Hoboken, New Jersey (Q6797722) and Mayor of Newark, New Jersey (Q20899061). Has anyone clustered entities like this before? For instance there is no cluster of the 50 governors of US states: Governor of New Jersey (Q3112728). --Richard Arthur Norton (1958- ) (talk) 05:26, 3 November 2017 (UTC)

list of current United States governors (Q11250358) is the item representing the Wikipedia List of the US governors. In general there tends not to be a Wikidata item for such groupings unless there is corresponding Wikipedia page for them, in which case is a list of (P360) is used instead of has part(s) (P527). When you say you want to cluster them — what do you actually want to achieve? If you're wanting to produce lists of them, that tends to be done as a SPARQL query, such as:
SELECT ?item ?itemLabel ?position ?positionLabel ?start
WHERE 
{
  ?item wdt:P31 wd:Q5 ; p:P39 ?ps .
  ?ps ps:P39 ?position ; pq:P580 ?start .
  FILTER NOT EXISTS { ?ps pq:P582 [] }
  ?position wdt:P279 wd:Q889821 .
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}
ORDER BY ?positionLabel
Try it!
--Oravrattas (talk) 08:26, 3 November 2017 (UTC)

I created mayor of a place in New Jersey (Q42688505) and will populate it based on the query. I think we should have them easy to find without having to know how to form a query. --Richard Arthur Norton (1958- ) (talk) 23:39, 4 November 2017 (UTC)

I have deleted P31 because item isn't a "Wikimedia category". --ValterVB (talk) 08:59, 5 November 2017 (UTC)
I changed the label to singular and added a description.
--- Jura 13:23, 5 November 2017 (UTC)
@Oravrattas: Can you add the "Wikidata SPARQL query equivalent" for me at: mayor of a place in New Jersey (Q42688505) like was done for list of presidents of the United States (Q35073).
The best I can do is 
SELECT DISTINCT ?item ?itemLabel ?itemDescription 
WHERE
{ 
  ?item wdt:P31 wd:Q5; wdt:P39 [ wdt:P279* wd:Q30185 ; wdt:P1001/wdt:P131* wd:Q1408 ] .
  SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en" }
}
Try it!
However, that times out, presumably because of the double *. Perhaps someone else can come up with a better version. --Oravrattas (talk) 10:03, 9 November 2017 (UTC)

Gadget / userscript editor feedback wanted

Tracked in Phabricator
Task T95403

Hi everyone! I’m considering to make a change to the HTML output for statements, and I’d like to gather some feedback from people who work on gadgets and user scripts :)

The problem is that any gadget that appends to the value of a statement (e. g. checkConstraints) changes the HTML in a way that Wikibase’s own JavaScript doesn’t expect, and sometimes the appended elements bleed into Wikibase’s own elements, causing e. g. phab:T167869 and phab:T169866. (To clarify: the tasks mention checkConstraints specifically, but at least the first task also affects other gadgets.)

My proposed solution is to change the layout of the HTML slightly, from the current 

<div class="wikibase-snakview-value-container" dir="auto">
  <div class="wikibase-snakview-typeselector">...</div>
  <div class="wikibase-snakview-value">...</div>
</div>

into the following: 

<div class="wikibase-snakview-value-container" dir="auto">
  <div class="wikibase-snakview-typeselector">...</div>
  <div class="wikibase-snakview-body">
    <span class="wikibase-snakview-value">...</span>
    <span class="wikibase-snakview-indicators"></span>
  </div>
</div>

There is a new element for “indicators” on a snak, and the value itself is now a span inside of a new div wrapper. The indicators are hidden while the statement is edited, and cleared on save. Gadgets can subscribe to the wikibase.statement.saved hook to populate the indicators again after a statement has been saved, using the new value. A simple example gadget using this technique is at User:Lucas Werkmeister (WMDE)/colorIndicator.js.

Here’s a list of some gadgets and user scripts I’m aware of that could use this “indicators” area:

Existing gadgets that use something like $( '.wikibase-snakview-value' ).append( … ) will continue to work, though they could be changed to append to the indicators instead. However, gadgets that select something like div.wikibase-snakview-value will break, since the element is no longer a div.

Do you see any problems with this new HTML layout, or do you want to suggest any improvements? --Lucas Werkmeister (WMDE) (talk) 11:29, 3 November 2017 (UTC)

In case of a change it might be a good idea to have a look at insource:/wikibase\-snakview\-value/ and notify script maintainers. —MisterSynergy (talk) 11:44, 3 November 2017 (UTC)
That’s a great idea, thank you! It looks like there are a lot of results with div.wikibase-snakview-value :/ --Lucas Werkmeister (WMDE) (talk) 12:00, 3 November 2017 (UTC)
I endorse this. The same enhancement would also be useful for sitelinks. Matěj Suchánek (talk) 11:53, 4 November 2017 (UTC)
Thank you! The sitelinks suggestion sounds useful too, I’ve filed phab:T179961 for it. --Lucas Werkmeister (WMDE) (talk) 17:56, 7 November 2017 (UTC)
Update: Instead of turning the .wikibase-snakview-value into a <span>, we’ll keep it as a <div> and make it display: inline-block via CSS, so this means that all existing gadgets and user scripts should (famous last words) continue to work – though it would of course still be a good idea to migrate them to the indicators area where applicable. --Lucas Werkmeister (WMDE) (talk) 17:56, 7 November 2017 (UTC)
Further update: Actually, we’ll use display: inline instead of display: inline-block. Also, this has been merged, so it should be deployed next week, hopefully. --Lucas Werkmeister (WMDE) (talk) 16:35, 9 November 2017 (UTC)

First cosmonaut/first astronaut

Yuri Gagarin (Q7327)? --Fractaler (talk) 15:46, 6 November 2017 (UTC)

Yes --Jarekt (talk) 17:04, 6 November 2017 (UTC)
How to tell the user about this? --Fractaler (talk) 17:38, 6 November 2017 (UTC)
I am sorry, but can you explain what is the question? Are you looking for a property counting in which order people went into space? I would be surprised if we had one. But you can use {{Autodescription | Q7327 }}: "Soviet pilot and cosmonaut, first human in space (1934–1968)". If that is not your question than please explain. --Jarekt (talk) 17:44, 6 November 2017 (UTC)
Ok. Some bot is looking for the one who was the first cosmonaut. What tool does the Wikidata offer for such user? Description? Bot should be able to understand the text of the description? --Fractaler (talk) 17:53, 6 November 2017 (UTC)
Hm, perhaps we should have a property for "first to..."? I'm unsure of what kind of scope we could have with this. --Yair rand (talk) 20:18, 6 November 2017 (UTC)
SELECT DISTINCT ?human ?humanLabel ?mission ?missionLabel ?startDate WHERE {
  ?human wdt:P31 wd:Q5 .
  ?human wdt:P450 ?mission .
  ?mission wdt:P619 ?startDate

  SERVICE wikibase:label {
    bd:serviceParam wikibase:language "en" .
  }
} ORDER BY ?startDate
LIMIT 1
Try it!
You can remove the limit to find next astronauts. Paweł Ziemian (talk) 20:29, 6 November 2017 (UTC)
Thanks for query, but if bot just ask: "First cosmonaut is ...?", it must use such query? --Fractaler (talk) 14:11, 8 November 2017 (UTC)
I do not understand the problem. Bots are used for boring and repeatable tasks. Maybe there is a task, which is hard to define in simple query and bot is required to scan a large amount of items to get simple result like 42, but for finding the first astronaut the query above is enough. Paweł Ziemian (talk) 19:18, 8 November 2017 (UTC)
A bot is just a program. It like a child. This child see phrase "first cosmonaut" and want to know/understand what is it. It have not any services,only Wikidata. --Fractaler (talk) 20:20, 8 November 2017 (UTC)

Is there an easy way to find the administrative region in different countries from coordinates for 100s of items?

Hi all

Myself and Navino Evans have been working to import the UNESCO Biosphere Reserves into Wikidata. We have coordinates for all the sites and have created a map on Monumental, however Monumental also requires located in the administrative territorial entity (P131) to be present, There are over 650 Biosphere Reserves but currently only 80 appear on the map. I understand that we can't just use the country for located in the administrative territorial entity (P131), is there some lookup service or some other way to populate the property with something more accurate than country?

Thanks

--John Cummings (talk) 09:53, 7 November 2017 (UTC)

As I recall, part of your problem is that many of these overlap countries and thus have multiple jurisdictions. No idea how to get around that one, sorry. Jane023 (talk) 10:36, 7 November 2017 (UTC)
Jane's point notwithstanding you might be able to use mySociety's 'MapIt' tool. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:06, 7 November 2017 (UTC)
Not sure how much MapIt can do outside the UK.
OSM's Nominatim can turn coordinates into multiple levels of administrative areas, worldwide. The Nominatim service at MapQuest offers up to 15,000 free lookups a month. Jheald (talk) 11:26, 7 November 2017 (UTC)
There's also a global version of MapIt: http://global.mapit.mysociety.org/ --Oravrattas (talk) 20:29, 7 November 2017 (UTC)
Ideally that's something we could do directly inside Wikidata as we now have support for geoshapes… But I have no idea if it is possible to query for all geoshapes containing a particular position. I suspect we would need an add-on to the query service for that. Of course that would require storing all the required geoshapes in Commons and linking the appropriate items to them. @Lucas Werkmeister (WMDE):, any thoughts? − Pintoch (talk) 12:08, 7 November 2017 (UTC)
Hm, I don’t really have any useful thoughts on this. It would be nice to have, but I have no idea if it’s technically possible – I guess WDQS would need to essentially hold a copy of all geoshape files on Commons, so that it could search them efficiently. --Lucas Werkmeister (WMDE) (talk) 15:27, 7 November 2017 (UTC)
While we do have geoshape support, we only have 479 usages of it, so you wouldn't get very far by implementing it at this point. An OpenStreetMap based solution will do much better at the moment. ChristianKl (talk) 11:19, 9 November 2017 (UTC)
Wasn't there a game by Magnus for this?
--- Jura 13:25, 7 November 2017 (UTC)
You mean this one? Matěj Suchánek (talk) 14:12, 7 November 2017 (UTC)
Matěj Suchánek, this would be perfect (for doing it manually, still kind of hoping there is an automatic way of doing it). Does anyone know how I could set this up for Biosphere Reserves? --John Cummings (talk) 14:31, 7 November 2017 (UTC)
I do have a tiny problem with this issue. Two of the three Biosphere Reserves in my country are whole provinces and the remaining one is a municipality. I'm not sure if located in the administrative territorial entity (P131) should be set to the provinces and municipality themselves or if we need to specify the higher administrative entity that encompass these reserves. —seav (talk) 19:09, 8 November 2017 (UTC)

Sharing economy subclasses

I noticed using wdtaxonomy that the subclass tree for sharing economy (Q2277143) is this: 

sharing economy (Q2277143) •29
├──crowd funding (Q348303) •45 ×2 ↑
├──carpool (Q749649) •30
├──car sharing system (Q847201) •29 ×4 ↑
├──bicycle sharing system (Q1358919) •29 ×86 ↑↑↑
│  ├──??? (Q17332642) •1
│  └──dockless bicycle sharing system (Q42556547) ×5
├──Flight sharing (Q18207130) •3
└──Rideshare (Q19698372) •3

I'm not sure any of these are really valid subclasses of sharing economy (Q2277143). I see systems and services like this as "part of" the sharing economy. I've defined the "part of" property on some other items, e.g. redistribution market (Q41795373). What do others think? Pauljmackay (talk) 08:02, 8 November 2017 (UTC)

They are more like examples of sharing economy? Other than the most common bike sharing car sharing and home rental (like airbnb), there are everything from sharing furniture to sharing kitchen to sharing boyfriends to sharing knowledge that are considered as sharing economy. As sharing economy is just a model where people are making greater good by sharing their resource that can be applied onto different entities, and the concept wouldn't be reduced if a particular element is removed, I think they are more of an example of instead of part of. C933103 (talk) 14:30, 8 November 2017 (UTC)
If they are examples of sharing economy (Q2277143), why not just use instance of (P31) (instead of subclass of (P279))? − Pintoch (talk) 09:59, 9 November 2017 (UTC)
Because thats not valid. Any of these individual sharing activities is not an "instance" of a sharing economy. The combination of many activities of that sort forms the economic activity that comprises the sharing economy. Pauljmackay (talk) 10:10, 9 November 2017 (UTC)
Then I guess they should not be linked at all! :) − Pintoch (talk) 16:18, 9 November 2017 (UTC)
I was wondering if facet of (P1269) would be more appropriate than part of (P361), but am not sure if facet of (P1269) is only intended for Wikipedia articles that are a specialisation of another more general article on a topic. Any ideas? Pauljmackay (talk) 17:18, 9 November 2017 (UTC)

Firework show and firework

firework (Q127933) links to the Wikipedia article and represents firework shows. Should there be a separate item for "firework" that represents an actual single firework? That would then be the logical superclass for all firework types such as catherine wheel, cake, etc. Pauljmackay (talk) 08:06, 8 November 2017 (UTC)

Firework is class (Q17519152)? spark (Q911679) is member of a group (Q36809769) firework? What is firework show? --Fractaler (talk) 09:46, 8 November 2017 (UTC)
I'm not sure that's the best representation. I see a firework as a single object, not a group. A box of fireworks would be a group. And a firework show is an event where fireworks are launched. Pauljmackay (talk) 15:55, 8 November 2017 (UTC)
Firework has states? Firework is a 1) fire/launched single object or 2) not-fire/launched single object (is launched or not)? Of what is 1 and 2? And box of fireworks, is launched or not? And when launched, box of fireworks is a group (Q16887380) or a system (Q58778)? --Fractaler (talk) 08:43, 9 November 2017 (UTC)

Open*Facts

There is a family of crowdsourced databases about consumer products, namely Open Food Facts, Open Pet Food Facts and Open Beauty Facts, and they seem like a natural fit as a Wikidata property. However, it seems the English Wikipedia doesn't generally have articles on specific food/pet food/beauty products, and seems to focus more on brands.

Can you think of examples of 1:1 relationships between Wikidata items and entries in those databases? ~★ nmaia d 10:36, 8 November 2017 (UTC)

I do not reply directly to the question but we already have Open Food Facts food category ID (P1821). Pamputt (talk) 06:59, 9 November 2017 (UTC)

Take the Subject of Licenses Seriously: a 2017 Community Wishlist Survey Proposal

Hello Wikidatians,

Although I think every Wikimedian is concern, and even people reusing data from Wikidata beyond Wikimedia, but sure people in Wikidata might feel especially involved, there is a 2017 Community Wishlist Survey proposal regarding the topic of licenses. Your feedback and other contributions on this proposal are welcome, and for the sake of global visibility, it would be better to maintain all that on the proposal Community discussion section.

Cheers, Psychoslave (talk) 09:38, 9 November 2017 (UTC)

Having the discussion at that place is a clever way to have it in at a place where it doesn't affect policy as the Community Tech team doesn't have any powers affectign policy. ChristianKl (talk) 11:05, 9 November 2017 (UTC)

Confusing help page

Cleaning up the Help namespace I stumbled upon this unlinked page Help:Special:MyLanguage/Help:Authority control. Can it be delete or moved to some proper place? It looks like this was accidentally created at least under its page name. -- JakobVoss (talk) 08:24, 13 November 2017 (UTC)

It was moved from somewhere else. Feel free to move it back or around.
--- Jura 09:06, 13 November 2017 (UTC)
  Moved to Help talk:Authority control (based on investigation of histories). Matěj Suchánek (talk) 20:27, 13 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 07:48, 15 November 2017 (UTC)

List of values

Hello. How can I have a list with all items that are using FIFA country code (P3441) with all the values and the items number. For example:

Item Value Item id

Germany national football team GER Q43310

Xaris333 (talk) 14:40, 13 November 2017 (UTC)

Hi. Is it that? - Kareyac (talk) 14:49, 13 November 2017 (UTC)

Yes! Thanks!! Xaris333 (talk) 15:45, 13 November 2017 (UTC)

This section was archived on a request by: Matěj Suchánek (talk) 07:48, 15 November 2017 (UTC)

Property proposal - iconography

I have proposed a new property Iconography to record the attributes and symbols of saints, mthological characters, etc. Please suggest ways to improve this proposal if you can. - PKM (talk) 21:50, 13 November 2017 (UTC)

Oops, this property was proposed and created in August as iconographic symbol (P4185). - PKM (talk) 22:11, 13 November 2017 (UTC)
I marked it accordingly. ChristianKl (✉) 12:02, 14 November 2017 (UTC)
Thank you! - PKM (talk) 21:36, 14 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 07:48, 15 November 2017 (UTC)

Disentangling of overseas territory (Q1211529)

The item is a bit a mess re. site links and class (most sitelinks are disambiguation pages, the Englisch page has been merged in the past, etc.) Since site links are involved, and I have no experience with these, it would be great if somebody could have a look at my plan to disentangle the item at Talk:Q1211529. Jneubert (talk) 13:29, 14 November 2017 (UTC)

This section was archived on a request by: Jneubert (talk) 13:56, 15 November 2017 (UTC) - with thanks to Jura!

Help with documenting Wikidata

I have been looking at Wikidata since June, reading bits and pieces here and there from Wikimedia sources, mailing lists, and YouTube videos. I now have random notes scattered over three computers in two states. Beginning with the workings of the database I would like some help in understanding and documenting what I have been seeing with a view toward updating the Wikidata and Wikibase articles on en.wiki and editing some documentation for here.

Wikidata is an amazing database to someone who last studied databases 20 years ago. I've been learning about graph databases and JSON blobs. As I understand it:

  1. The Wikidata project modified core Mediawiki to store structured data as "pages".
  2. Two types of entities are stored, item (Main namespace) and property (Property namespace). This gives us versioned entities with easily accessible histories.
  3. The project also developed two extensions, Wikibase Repository to display and edit the data and Wikibase Client to retrieve data for use in an associated wiki.
  4. Wikibase was delivered as an empty database 29 October 2012 and began to be filled with information from the infoboxes in the various language Wikipedias.
  5. This provides a rather unconventional database that can be searched but is not in a form to be easily queried. The current version is exported in weekly dumps in several formats for anyone to use; one version is archived at the Internet Archive.
  6. The Wikidata Query Service provides an interface to query the RDF version using SPARQL.

So far what I have is mostly original research on my part. I would like to find documents that describe these parts of the project in detail so that I can use them as sources. Thank you in advance for your help. StarryGrandma (talk) 19:10, 8 November 2017 (UTC)

There are a bunch of scientific papers on Wikidata that will provide high quality references. Just put Wikidata into Google Scholar. ChristianKl (talk) 23:46, 8 November 2017 (UTC)
@StarryGrandma:
Nice summary by the way. I would have added that an unconventional panoply of editing tools from fast editing to gamification in the true wiki-spirit.  — billinghurst sDrewth 01:06, 9 November 2017 (UTC)
ChristianKl, billinghurst, thank you both. I've looked through Google Scholar and found some useful papers, even some independent of the developers. The links to the archive and history are great. The earlier versions available in the histories are very helpful. I will look at the tools. I'd forgotten I did do some of the challenges during the migration of Persondata. StarryGrandma (talk) 01:22, 10 November 2017 (UTC)

Order in the wikidata

We have decade (Q29442417) (group of 10) and we have dozen (Q605704) (group of 12). Should there be a connection between them, a logical transition - Q29468245 (group of 11)? --Fractaler (talk) 08:32, 9 November 2017 (UTC)

Wikidata:Notability said: An item is acceptable if: 3) It fulfills some structural need, for example: it is needed to make statements made in other items more useful. Item "group of 11" is needed to make statements made in other items more useful? --Fractaler (talk) 09:00, 9 November 2017 (UTC)
In French onzaine does exist. --Psychoslave (talk) 09:40, 9 November 2017 (UTC)
And in English eleventh. But for some reason it is considered the norm in Wikidata to delete such items altogether without discussion. Why? --Fractaler (talk) 10:14, 9 November 2017 (UTC)
If you want to have items like that that won't deleted you can bring referenes (Wikitionary itself isn't the best reference). Apart from that the English Wikitionary doesn't say that eleventh is a word that means a group of 11 things. ChristianKl (talk) 10:59, 9 November 2017 (UTC)
@Andreasmperu: Given that you deleted it, it would make sense for you to reply as well. ChristianKl (talk) 11:10, 9 November 2017 (UTC)
I don't understand, how can I bring reference to an items that does not exist? Is "source itself isn't the best references" a reference for "bring references" or not? Is Wiktionary (Q151) a source that is banned in Wikidata? And I never said "eleventh is a word that means a group of 11 things". I said: And in English eleventh. Item "eleventh" also (as a very large number of the other item like this) was deleted without debate. Is this the norm for administrator behavior? --Fractaler (talk) 11:29, 9 November 2017 (UTC)
eleventh (Q28469721) isn't deleted. ChristianKl (talk) 12:01, 9 November 2017 (UTC)
eleventh (Q28469721) is ordinal number (Q191780), but "deleted eleventh" said: eleventh is "one of eleven equal parts of a whole". --Fractaler (talk) 12:14, 9 November 2017 (UTC)
Admins deleting items with no discussion, and refusing to recreate them when such behaviour is challenged, is not a "norm", but it does happen too often, and the admins collectively, via their noticeboard, have a history of failing to act to remedy this. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:12, 9 November 2017 (UTC)
Is it possible to get at least a list of items (with a Labels, description, links to other items)? --Fractaler (talk) 13:31, 9 November 2017 (UTC)
@Fractaler: I don't understand this whole series of "group of x" that you seem to have added. I have never heard (in English) the word "decade" used to represent a generic "group of 10 things", nor most of the others in that collection. You have no references or sitelinks for any of them. What conceivable structural need do they fulfill? It seems to me they only lead to confusion by adding multiple uncorrelated new items for things that we already may have items for. For example, how is monad (Q39604065) different from singleton (Q1165112)? Both have the meaning of a set or series consisting of 1 member. These should at least be merged. I have no opinion on the usefulness of a "group of 11" or higher numbers, but such a sequence surely needs to stop somewhere reasonable. ArthurPSmith (talk) 12:46, 9 November 2017 (UTC)
@ArthurPSmith: It is all from Wiktionary (Q151), source itself isn't the best reference: wikt:triad, wikt:tetrad, ..., wikt:decade, ..., etc. singleton (Q1165112) is n-tuple (Q600590). monad (Q39604065) is group (Q16887380). group (Q16887380) is n-tuple (Q600590)? --Fractaler (talk) 13:01, 9 November 2017 (UTC)
A singleton is also identified in wikidata as a "set". A "group" is surely the same thing as a "set"? A "tuple" is an ordered set, but order obviously is not an issue when there's only 1 element. So anything that is a "group of 1 item" is also a "set of 1 item" and a "tuple of 1 item"; the meanings of these terms are clearly the same. ArthurPSmith (talk) 13:07, 9 November 2017 (UTC)
What do you mean "identified"? Description? What is description?. What is the superset of the set singleton (Q1165112)? The superset is n-tuple (Q600590). There are dictionaries, which while Wikidata's users use in the mixture. Singleton is a math term. monad (->group) is not math term. --Fractaler (talk) 13:24, 9 November 2017 (UTC)
singleton (Q1165112) is described as a "set" in the English description, but it also has the statement subclass of (P279) finite set (Q272404). As to "math term" - "monad" is not a term used in ordinary English at all, so I'm not sure what you're getting at there. There is also monad (Q12370437) from philosophy but that has a very different meaning. ArthurPSmith (talk) 15:57, 9 November 2017 (UTC)
Of course in Wikidata there are a lot of homonyms. But now all my not-homonyms was deleted without disscussion, also as describes. Monad was because I did't found other term for "group of 1", and only monad was, in Wictionary. No problem, have you sinonym? --Fractaler (talk) 20:29, 9 November 2017 (UTC)

Wait, so the suggestion is like, because there are US nautical miles (Q39380159) that equal to 1852 meters, and then there are mile (Q253276) that equals to 1609 meters, therefore we should create all the other arbitrary units in between them from a unit for 1851 to 1850 to 1849 to 1610 meters? C933103 (talk) 17:23, 9 November 2017 (UTC)

There are users, which do 1)machine-readable data (Q6723621), and which do 2)Q28777989. So, you do 1) or 2)?--Fractaler (talk) 20:29, 9 November 2017 (UTC)

Multiple items for the same concept

It seems that sometimes it's necessary to have multiple items for the same concept in Wikidata, because a Wikipedia has decided to have multiple articles. An example is human (Q5) and Homo sapiens (Q15978631). I don't think any information would be lost in Wikidata if they were merged into a single item, since one is just the common name of the other. (Human is also used to refer to Homo (Q171283), but that's not at issue here because human (Q5) explicitly refers to Homo sapiens). Unfortunately, this means that information is collected in two places about the same thing, and other items will link randomly to one or the other. This will make it harder to run queries. I suppose somebody will disagree with me that these two items actually refer to the same thing. But in general, is there any procedure by which an item can be marked as the "main item", taking both names as aliases, and the other marked as "created for the benefit of Wikipedia" This would be similar to the concept of category items. Ghouston (talk) 21:17, 7 November 2017 (UTC)

Basically characters who are easily identified as fictional you also have various mythical figures. If you trace ancestor trees of some people you get back to Greek gods or Hindu gods. The Bible has people who are very old and should probably not show up when you seek for the oldest people.
I don't think our current handling is optimal but I do think there's a reasonable argument to be made for multiple items. ChristianKl (talk) 22:05, 7 November 2017 (UTC)
Anything fictional isn't a real thing, by definition. Harry Potter isn't a person, and doesn't have a human tooth. In the same way, I wouldn't expect a fictional president of the USA to turn up in a query of US presidents. Ghouston (talk) 22:15, 7 November 2017 (UTC)
I think that a famous painter can paint a picture of a human tooth without that tooth having to belong to any particular human. An anatomy textbook for dentist might also show a fictional tooth to teach prospective students that doesn't belong to any particular human. Do you think we should say that the anatomy textbook doesn't show a human tooth? ChristianKl (talk) 22:32, 7 November 2017 (UTC)
It would be a drawing of a tooth that depicts human tooth (Q561) rather than any specific instance of tooth. But this seems off topic, Harry Potter isn't described in Wikidata as human (Q5) or Homo sapiens (Q15978631), and I'm still no better informed about which one human tooth (Q561) should link to. Ghouston (talk) 22:51, 7 November 2017 (UTC)

Here's another example of how the current setup seems wrong. The Dutch Wikipedia has a single article nl:Mens which is site linked to human (Q5). The Spanish Wikipedia has a single article es:Homo_sapiens which is site linked to Homo sapiens (Q15978631). Shouldn't they both link to the same Wikidata item, since they are about the same topic? Ghouston (talk) 23:12, 7 November 2017 (UTC)

  • As far as I'm concerned we should use redirects to get the interwiki links working.
To make thinks more fun we have more items than just human (Q5) and Homo sapiens (Q15978631). We also have anatomically modern human (Q5891007) and Homo sapiens sapiens (Q3238275). I'm not strongly commit to the current status quo, but solving it well isn't an easy problem. ChristianKl (talk) 01:18, 8 November 2017 (UTC)
@Brya, Succu: Do you have any opinions here? ChristianKl (talk) 01:20, 8 November 2017 (UTC)
Given that we don't have living specimens of other subspecies of humans to study, it's probably never going to be all that clear exactly which properties would uniquely apply to Homo sapiens sapiens (Q3238275) or anatomically modern human (Q5891007). I think it would be best to assign everything to one Human / Homo sapiens item, and use other items when they are really necessary. E.g., an item "Homo sapiens (taxonomic classification)" for en:Homo sapiens to link with. Ghouston (talk) 02:32, 8 November 2017 (UTC)

I found permanent duplicated item (P2959) and Wikimedia permanent duplicate item (Q21286738) which I think answer my original question: if a Wikipedia does have two articles for the same topic, one of its Wikidata items can be marked with that. Ghouston (talk) 03:23, 8 November 2017 (UTC)

permanent duplicated item (P2959) - a great idea for solving the problem of homonymy (Q21701659). As in Wikipedia. We have homonym (Q160843) (Q5). And this homonym will allow the user to choose next navigation page of the Wikidata hierarchy: Qhuman (general dictionary), Qhuman (religion), Qhuman (philosophy), Qhuman (biology), Qhuman (medicine), etc. --Fractaler (talk) 06:58, 10 November 2017 (UTC)

Wikidata software restrictions

In the current version of the Wikipedia engine, there is no possibility to switch between the types of dictionaries (general, of trivial name (Q913170), professional - religious, philosophical, scientific, etc.), there is only an opportunity to switch between language dictionaries (English dictionary, French dictionary, German dictionary, etc.). Now the user can not select the dictionary he needs (how he chooses the language) for an item that models a part of the world: Q5 (human:general dictionary), Q5 (human:religion), Q5 (human:philosophy), Q5 (Homo sapiens:biology), etc.; Q190 (God:religion), Q190 (god:science), Q190 (God:philosophy), etc. Until this restriction is eliminated, Wikidata editors (Q28859214) are doomed to problems of this kind. --Fractaler (talk) 08:42, 8 November 2017 (UTC)

SPARQL and Wikipedia usage

I read https://meta.wikimedia.org/wiki/2017_Community_Wishlist_Survey/Wikidata/Tracking_category_in_Wikipedia_for_missing_labels . Is it possible via SPARQL to filter for items that are used within Wikipedia? ChristianKl (talk) 10:54, 9 November 2017 (UTC)

Somewhat possible. The information about used items is available via API. However, I didn't manage to filter for those items that need a label:
SELECT * {
  SERVICE wikibase:mwapi {
    bd:serviceParam wikibase:api "Generator" .
    bd:serviceParam wikibase:endpoint "cs.wikipedia.org" .
    bd:serviceParam mwapi:gcmtitle "Kategorie:Údržba:Doplnit štítek na Wikidatech" .
    bd:serviceParam mwapi:generator "categorymembers" .
    bd:serviceParam mwapi:gcmlimit "10" .
    bd:serviceParam mwapi:prop "wbentityusage|info|pageprops" .
    bd:serviceParam mwapi:ppprop "wikibase_item" .
    bd:serviceParam mwapi:wbeulimit "max" .
    # this doesn't work bd:serviceParam mwapi:wbeuaspect "L.cs" .
    ?title wikibase:apiOutput mwapi:title .
    ?item wikibase:apiOutputItem mwapi:item .
    ?item2 wikibase:apiOutputItem "wbentityusage/wbeu/@id" .
    # this doesn't work ?aspect wikibase:apiOutput "wbentityusage/wbeu/aspects/aspect/@???" .
    ?num wikibase:apiOrdinal true .
  } .
} ORDER BY ?num
Try it!
Matěj Suchánek (talk) 16:11, 10 November 2017 (UTC)

Self-referencing not identifed as self-referencing  ?

Hi all,

Every now and then, I check the constraints violations, in particular the self-referencing. Most of te thime, they are identified as such (qv. Special:Diff/580531184) and the job is much more easy; but for some reason, sometimes it's not tagged (qv. Special:Diff/590060916; there is other cases but I don't remember them). Does somebody know why?

Cdlt, VIGNERON (talk) 14:47, 9 November 2017 (UTC)

The abuse filter was disabled due to performance reasons, see here. Sjoerd de Bruin (talk) 15:43, 9 November 2017 (UTC)
Ok, I'm sad but thank you for the explanation Sjoerddebruin. @Matěj Suchánek: would it be possible to reactivate it somehow? maybe, by limiting it to direct value of properties (where there is the less false positive) and exclude qualifiers? Cdlt, VIGNERON (talk) 16:12, 9 November 2017 (UTC)
Hopefully, I can make a fix soon. Could you explain first when it's fine in qualifiers (and references)? Matěj Suchánek (talk) 17:50, 9 November 2017 (UTC)
@Matěj Suchánek: there is a lot of very weird and different cases where self-reference is ok (or kind of ok). One symbolical case is the Tree That Owns Itself (Q1421130) but you also have time-travellers killed by their younger/older self and other stuff like that ; fortunately, these cases are quite rare. Meanwhile, in qualifiers and references, you have more strange and complex cases ; in particular in movies and books (where credits or other informations about the objects is stated in the object itself), for instance today I stumble upon Kari Hakli (Q5402288) (where the image (P18) is an self-portrait (Q192110), so author (P50) is himself). These cases are not exactly correct nor incorrect, they are in a grey area. I suspect that these cases re more common and could affect the performances (maybe wrongly). Cdlt, VIGNERON (talk) 20:42, 9 November 2017 (UTC)
If we know it's an self-portrait (Q192110) (I would use instance of (P31) for this qualifier, though), why should we also store the information about its author? Matěj Suchánek (talk) 08:26, 10 November 2017 (UTC)
@Matěj Suchánek: I don't know, as I said « These cases are not exactly correct nor incorrect » and the question here is not about correctness but about frequency (based on my experience, it's more frequently clearly incorrect in direct properties) and if it really impact the performance of the abuse filter. Cdlt, VIGNERON (talk) 08:55, 10 November 2017 (UTC)
Here an other case more or less correct and frequent : The Cheetah Girls 2 (Q587876) where references for cast member (P161) are the credits of the movie itself. Cdlt, VIGNERON (talk) 09:22, 10 November 2017 (UTC)
Could be, yeah. But it's still "self-referencing". Anyway, I hope the filter will now be faster. Matěj Suchánek (talk) 09:26, 10 November 2017 (UTC)
Plus, see these recent changes to see that most of them are in qualifiers and references. Cdlt, VIGNERON (talk) 12:30, 10 November 2017 (UTC)
Tree That Owns Itself (Q1421130) is homonym (Q160843), is 1) white oak (Q469555) (term of biological dictionary), 2) object that owns, proprietor (Q16869121), owned by (P127) (term of legal dictionary) 3) object that is owned (term of legal dictionary). 2+3="object that does not have a classification by time or space" (simultaneously, in one place it is both the owner and the object of ownership). --Fractaler (talk) 11:45, 10 November 2017 (UTC)
Yes, and? I didn't understand anything you wanted to say. Anyway, this is a exceptional and folkloristic case, maybe Samsung Card (Q624076) would be a better example to talk about. Cdlt, VIGNERON (talk) 12:30, 10 November 2017 (UTC)
Sorry for my bad English. Samsung Card (Q624076) is also homonym (Q160843): 1)business (Q4830453), 2)proprietor (Q16869121), 3) object that is owned. --Fractaler (talk) 18:39, 10 November 2017 (UTC)

theological concepts and interpretations of quantum mechanics

As a beginner in Wikidata, I tend to have all kinds of stupid questions. Today I found a couple more and I'd appreciate if someone else would answer them. Both emerged when I tried to classify Berkeleian interpretation (Q3800537). First, I looked into all the interpretations we have (there seems to be the usual mess across different Wikipedias) and stumbled on the theological interpretation (canon law) (Q6056448) which had no instances or statements. I'm not a native speaker of English, so I have no idea how someone else might classify it, but I would say that unless there would be a class of concepts of canon law (which I haven't found), it should be an instance of both legal concept (which I inserted, Q2135465) and theological concept (which doesn't seem to exist). So, should I create an item for the theological concepts/concepts in theology/whatever you name it (I don't have the feel of the naming habits here yet) or would that be somehow superfluous and should be solved by some other way? There also does not seem to be a separate item for the concepts in logic which in some cases in philosophical logic, I suspect, some people would claim would perhaps not fit best under mathematical concepts (e.g. interpretation in model theory, Q17101792). Should there be one? Second, I finally managed to find the "interpretations of quantum mechanics" (Q899137) but it only had one actual interpretation referring to it, many-worlds interpretation (Q40590). That one used the "subclass of" property (P279) to link to the Q899137, so I imitated it in the Berkeleian interpretation and Copenhagen interpretation (Q46079) as well. Yet later I started to wonder whether that was correct or had I just propagated a mistake. I mean, all particular interpretations are named in singular, which in my mind would hint at them being singular items, not subclasses. So, how should the relation be defined? And last but not least, what should the "interpretations of quantum mechanics" (Q899137) itself be instance of? --Oop (talk) 11:43, 5 November 2017 (UTC)

That's a tough one. Here a few thoughts, but also the reminder that there is no single right answer, and also that many answers are not necessarily wrong. We are talking about identity of abstract objects here, and this is known to be a hard problem for ontology in general, and also a woefully underexplored one (see for example the Stanford Encyclopedia of Philosophy on the topic of [https://plato.stanford.edu/entries/sortals/#SortForAbstObje Sortals for abstract concepts).
It certainly makes sense to regard the Copenhagen interpretation (Q46079) as an instance of the "interpretations of quantum mechanics" (Q899137), and thus make all such interpretations instances of the latter. This has the great advantage that one can ask a query engine for all instances of interpretations, and thus get all the results back.
Now one issue is that complex abstract concepts like the Copenhagen interpretation are often actually not a single abstract entity. In fact, there might be interpretations that are "subinterpretations" of the Copenhagen interpretation, or that have historically developed from it. Such a subinterpretation would make a lot of sense to also mark an instance of the Copenhagen interpretation? But alas, we already said that the Copenhagen interpretation is an instance of the QM interpretations. This would thus require to make the Copenhagen interpretation a subclass of the QM interpretations. But if we then remove it from the instances - because things surely shouldn't be instances and subclasses of one thing at the same time - our original query will break! Also, what does this even supposed to mean?
So, one way to regard this is to use only the subclass of relation between them, and to regard the actual instances as "the understanding of that interpretation in the head of a single scientist at a specific time". We wouldn't have items of those in Wikidata, usually. But if we regard this to be the instances, then we have a pretty workable model for why we would be using "subclass of" to connect the articles about the interpretations. These articles are all types for the understandings in the scientists heads.
In the end, what is more important than getting it right - because, as said, there is no single right or wrong - is to gain a shared understanding of the domain and formalize that. We want to be able to answer specific queries, and for that, the data needs to be structured in a specific way. So what we should do is, if we get into a murky area like this, and there starts to be some prolonged discussions about it, to agree on the use cases we want to answer, consider how these can be turned into queries, and then edit the data accordingly. If there is no discussion, we can, as we do in many cases, just go ahead and do what feels right, but sometimes this won't work out, and then we need some way to get to an agreement.
Instance of and subclass of lend to particularly vicious discussions. We should all consider more humility in discussing those, and realize that most of it is happening in our head. We should find models and that are useful and that we agree on in order to solve our tasks at hand - because it is easy to argue one way or the other, and both of them can be right. But we want to get through Wikidata with the minimal ontological commitment necessary to get our job done, so that we can work on more issues together.
If it turns out to be insufficient, because a new use case shows up, the beauty of Wikidata is that, unlike natural language text, it is often much easier to write a few queries and a few bots and get a situation fixed again.
I hope this wasn't too much preaching. Cheers, --Denny (talk) 23:46, 5 November 2017 (UTC)
(Long time, no see; please pass my regards to Q.) Concerning purely Wikidata, I'd say it might be often best to take a pragmatist approach, as all the questions need answers that work, and that in turn comes often down to just having an agreement how to solve a particular problem. From purely metaphysical point of view, perhaps one deeply non-Platonist approach would be to say there is even no essential difference between subclasses and instances, because they're both just ways of describing a certain relation between entities, and the description is mostly in our head(s) - as, also, is quite often the separation of entities. But we probably need some kind of difference to make the practical system work. So, I can keep using the subclassing for the qm interpretations and if someone feels an overwhelming urge to improve their searchability, they can turn subclasses into instances. Still no clue about canon law and instancing Q899137 itself, though. I mean, I could make every uninstanced term in physics a "concept" of some kind, but I doubt that's what is the actual practice here. More preaching, please. --Oop (talk) 09:29, 6 November 2017 (UTC)
I would prefer using subclass of (P279) at this point. I think that's more in line with the subclass/instanceOf distinction that we set as norms with our constraints. ChristianKl (talk) 14:57, 6 November 2017 (UTC)
There is one major difference between subclassing and instance of, which is in the formalization, and that difference has a major effect in how to use it: subclass of is usually regarded as transitive - and in many Wikidata queries this is represented by the * on P279 - whereas instance of is not (but instance of is transitive over subclass of). And that difference is worth preserving, but it only appears in actual queries (which is why I was pointing out that we need to write down the queries we want to have answered).
If we think that both instance of and subclass of are the same, we run into the classical problem of "Fred is a bald eagle, bald eagle is an endangered species, therefore Fred is an endangered species". By keeping the P31 and P279 carefully separated, we are avoiding this problem, as Fred is an instance of bald eagle, bald eagle is an instance of endangered species, but a subclass of bird. Therefore we could infer that Fred is a bird, but not an endangered species anymore.
For abstract concepts, as said, it is slightly different, because the question of identity is very different. Unless we can tie identity of abstract concepts to something, we are running in the problem that the differentiation between subclass of and instance of is hard, because we actually do not know what our entities with identity are. I hope that makes any sense. --Denny (talk) 16:05, 6 November 2017 (UTC)
Though (whether for more or less abstract concepts) I believe, as I have stated elsewhere, that our choice for these relationships helps to more precisely define the meaning we are associating with their wikidata items, rather than being a matter of some absolute truth. If Copenhagen interpretation (Q46079) instance of (P31) interpretation of quantum mechanics (Q899137) that says something different about what interpretation of quantum mechanics (Q899137) means than Copenhagen interpretation (Q46079) subclass of (P279) interpretation of quantum mechanics (Q899137). Both may be justifiable, but in the first case "interpretations" represents a grouping of theoretical outlooks on the physical universe, while in the second case "interpretations" is more representing the action of interpreting. And in turn that difference in meaning should weigh into the P31 and P279 relationships that interpretation of quantum mechanics (Q899137) is given (it has neither at the moment, but it probably should be an instance or a subclass of physical theory (Q9357058) (or one of its subclasses)? ArthurPSmith (talk) 20:18, 6 November 2017 (UTC)
Yes, that is often the case when doing ontological work. We are pinning down the exact meaning of the item. It is a bit problematic though, because we usually do not create the items for the other meanings, and we are, at the same time, rather cavalier about where exactly the statements are, and prefer them on a single item instead of spread out on a number of items with very closely related meanings. I am not saying that this is bad - I really am not, it might be actually the only way that this can work, who knows? A little bit of conflation can go a long way. --Denny (talk) 16:00, 7 November 2017 (UTC)
@Denny: While the initial draft of Wikidata included the idea of being agnostic about a lot of ontological issues, the adoption of constraints means that even if we make some expection to the constraints we have a developing standard of what we mean with subclass/instance. Given the dynamics it feels to me like we will settle sooner or later on mainly using the subclass relationship to specify how abstract classes relate to each other. ChristianKl (talk) 21:52, 6 November 2017 (UTC)
Maybe. I rather like to think that queries will be the great unifier and force us to have more conformity throughout the knowledge base. And constraints can often be regarded as a special case for queries. Insofar, and this I anticipated, we will slowly make more and more commitments. But what I think is important is to not make them prematurely, but when needed. --Denny (talk) 16:00, 7 November 2017 (UTC)
A person who writes a query doesn't have necessarily the power to change around a huge amount of how items are classified. On the other hand, if we decide on specific constraints, then enforcing those constraints works as a community process. ChristianKl (✉) 14:13, 11 November 2017 (UTC)

Min, average and max values for the same property

Hello everyone,

We were talking on the French Bistro about how to fill some elements (like the speed (P2052) of Saturn (Q193)) that have data available for, for example, minimal (minimum (Q21067468)), maximal (maximum (Q21067467)) or even average (Q202785) values of the same property. For the need of an example, I used P794 (P794) for the mean speed of Saturn (Q193) around the Sun, but the property seems deprecated for most uses.

With so much uncertainties, it has been suggested by Tubezlob we expand the debate here in order to "unify and set in stone the way to specify a maximum, minimum or mean" value for the sake of internal consistency and to facilitate external uses (especially inside infoboxes)."

So how do you think we should do it? :) J. N. Squire (talk) 18:20, 6 November 2017 (UTC)

I think this is a perfect example where "as" is appropriate. Amqui (talk) 18:26, 6 November 2017 (UTC)
Sometimes determination method (P459) is used, e.g. on Müggelsee (Q694789). --Pasleim (talk) 07:24, 7 November 2017 (UTC)
Saturn moves more slowly when it is further away from the sun in its elliptical orbit and speeds up nearer the sun. For an example of what to do see 4 Vesta, an asteroid whose apparent magnitude varies as seen from earth. The entry has both minimum and maximum magnitude values, qualified with the items minimum and maximum. You have already qualified the mean speed as average. Also qualify Saturn's speed as orbital speed. I understand the property proposal for average orbital speed was turned down with the suggestion to "use speed (P2052) with qualifier of (P642)=orbital speed (Q200924)". StarryGrandma (talk) 20:40, 8 November 2017 (UTC)
Instead of average (Q202785) may be to use mean (Q2796622)? --Infovarius (talk) 21:41, 9 November 2017 (UTC)

Beta label editor's problem.

I was trying to edit a Traditional Chinese entry with the beta label editor by typing in the language code zh-Hant, however it rejected me by saying no such lang code, and it only accept me when I type zh-hant instead. However, per IETF standard, zh-Hant is the preferred way to capitalize such a code. Can that be fixed for the beta editor?C933103 (talk) 06:13, 8 November 2017 (UTC)

Actually, I think we don't have any capitalized language codes in Wikidata. So if that is wrong, Wikidata is what should be fixed. Perhaps there were some discussions in the past... Matěj Suchánek (talk) 08:51, 11 November 2017 (UTC)

Marital status

Following a twitter conversation about the use of "NoValue" statements in spouse (P26) to indicate "not currently married"; and also noting that this issue has been been raised before on Project Chat regarding an edit war on Franz Kafka; I was wondering:
Is there value in a "Marital Status" property, corresponding to the item marital status (Q1282093)?
Both for Privacy and Notability reasons, it might be very useful to be able to state that a person is married (or divorced, etc.) without having to ALSO have an item for their spouse. This is not dissimilar to number of children (P1971) in that regard. Also, it appears (via both examples linked above) that we are sometimes using spouse (P26) as a stand-in for being able to query marital status. I believe the only way you can currently ask for "is this person married" is to query for a Spouse statement that does not have an end date and/or has a "preferred rank" (if they've been married more than once). This also implies that a Spouse->NoValue statement needs to be added for any period before first marriage, between any other marriages, and after a divorce. Furthermore, number of children (P1971) (and also unmarried partner (P451) only covers the relationship in a true/false sense and doesn't cover many other related types of status, including: betrothed, engaged, married, divorced, de-facto, civil union, widowed... "Marital status" (also sometimes called civil status) seems to be is a relatively common type of 'vital statistic' that is frequently stated in everything from government forms to trashy magazines. What do you think? Wittylama (talk) 12:36, 9 November 2017 (UTC)

[Pinging the people who were involved the previous discussion: Matěj Suchánek, Pigsonthewing, Vrenak, Koxinga, ChristianKl, Jura1, VIGNERON etc. :-) Wittylama (talk) 12:46, 9 November 2017 (UTC)
  • If you don't want to create an item for the spouse, you could use "somevalue". Personally, I only add "novalue" in P26 to people with P570.
    --- Jura 12:51, 9 November 2017 (UTC)
Not sure.
One thing for sure, according to Wikidata:Notability there is no problem of notability (qv. mainly point 3 some structural need and partially point 2 too). And as all data in Wikidata are supposed to have publicly available references, I guess privacy shouldn't be much of an issue either.
Meanwhile, I am concerned about the creation of item just for the sake of fulfilling structural need. If we push this logic to its end, we should have items about all peoples that have existed (or at least all we have sources, does civil registry count as a sources? if so, all people born, married or died in France for the last 2 century should have an item, that several hundreds of millions of people).
Cdlt, VIGNERON (talk) 12:54, 9 November 2017 (UTC)
For people who were married to one of the persons on this list, I think there is a structural need.
--- Jura 13:02, 9 November 2017 (UTC)
Creating an item for the Spouse of a Notable person is allowed because of the "structural need" rule, but that should not be required merely to be able to state the fact that the Notable person is married. In exactly the same way that we should not have to create items for their children to be able to say they have kids (c.f. number of children (P1971)). Wittylama (talk) 13:11, 9 November 2017 (UTC)
  • Our notability guidelines say that the structual need of linking a person via the "spouse" category makes them notable.
I don't think that there's a rule that causes us to have all items that qualify under our notability guidelines. Notability just means that we don't delete an item when somebody created it. When it comes to creating millions of items in bulk that needs a bot request and the corresponding discussion.
As far as I'm concerned "Spouse->NoValue" means that the person isn't married. "Spouse->UnknownValue" means that the person is married but Wikidata doesn't know who the spouse happens to be. ChristianKl (talk) 13:05, 9 November 2017 (UTC)
Using unusual variations of the Spouse property in order to clarify that the person is/isn't/was married seems to me to be a clever hack. It also doesn't account for things like "engaged" which is quite important info for things like modelling royal families (and could be qualified with "to -> Qitem"). Wittylama (talk) 13:11, 9 November 2017 (UTC)
It's not a hack, it's the core idea of what "NoValue" is supposed to do.
If you want model engagement you can currently do it with unmarried partner (P451) or significant person (P3342) and qualify it appropriately. It's a statement with one qualifier in the same way that having "martial status" "engaged" with a "to -> X" qualifier would work. The complexity would be the same. ChristianKl (talk) 13:28, 9 November 2017 (UTC)
Unmarried Partner is indeed a good (indeed - preferable) way to link to the Qitem for the fiancee (etc). Good point, thanks. Nevertheless, I still feel that the [marital/civil] "status" ought to be able to be expressed in data without necessarily requiring a Spouse property. For example - take Sean Penn (Q44221). He has three "spouse" statements (two people with start and end dates, and a "no value" with a start date the same as the second Spouse's end date); and one "umnarried partner" statement (currently without start or end dates). That is: "married twice, divorced twice, currently not married. In a stable relationship." To take this structure to its fullest, should we ALSO have a "Spouse -> no Value" for the time-period between the two Spouse dates? What about before the first one starting at birth until the first marriage? Would it not be simpler, and cleaner/neater, to indicate the Spouse(s) in one property [a list of linked names, and with their dates], and the Status in a different one [a chronological list of known civil status] - rather than trying to infer the latter from the former? Wittylama (talk) 14:17, 9 November 2017 (UTC)
I personally don't see how this is a privacy issue, if the information 🛈 wasn't known then it would be original research and could be reverted as such, otherwise spousal data is already publicly available knowledge somewhere before it was placed on a Wikimedia project. I just don't see how adding less information would benefit the project. 🤔-- 徵國單  (討論 🀄) (方孔錢 💴) 13:58, 9 November 2017 (UTC)
The relationship to privacy was suggested in this tweet, but it is not the core point of the question, just a potentially useful side-effect. If a notable person is publicly known to be married but not to whom, then yes, we could equally well say "spouse -> UnknownValue" this is true. The point is that it seems (to me at least) to be answering the question of "is this person married?" in reverse... I would feel that the structure of the statements should be in the order: "this person is married. This is their spouse." but currently we are stating "This person has a spouse. There is no end-date. Therefore they are married". Wittylama (talk) 14:17, 9 November 2017 (UTC)
Not really sure if the currently situation is optimal when trying to determine a history, but I don't think adding another property to keep in sync makes it easier.
--- Jura 14:26, 9 November 2017 (UTC)
In biographical dictionaries there are many examples of knowing to whom a person is married (John Doe married 1. Bertha, daughter of Sir John Bloggs; 2ly. Beatrix, daughter of Lord Farquhar) Regularly, I lack the will (or maybe the distraction) to create the person or chain of people that are referenced. With Wikidata, the problem is that I cannot reference that they were married, even when I know a name, without going through the full creation process of otherwise non-notable person. I will comment that I have seen people avoid this by using significant event (P793) to record a marriage event without the detail.  — billinghurst sDrewth 21:34, 9 November 2017 (UTC)
Supplementary question. @Wittylama: On Guy Ritchie (Q192990) the question was about ranking? Ranking marriages itself is an interesting concept, and preferring a marriage seems unusual. Did the conversation cover why one was or should be preferred? We can use series ordinal (P1545) to show the order of marriages rather than applying a preference.  — billinghurst sDrewth 21:43, 9 November 2017 (UTC)
Wikidata is designed to encourage the creation of items for those people and I think it's good that it's structured this way. From an usability perspective it could be easier to create the relevant items but given that Wikidata focuses on linked data you say that Bertha is the daughter of Sir John Bloggs by having items for both. ChristianKl (talk) 13:40, 10 November 2017 (UTC)
Creating more anchorless and unnotable people is something that I decline to do. I don't keep up with the important tasks, let alone the nonentities. This is not a glorified family tree, there are better sites and means of managing those people and references properly.  — billinghurst sDrewth 14:58, 10 November 2017 (UTC)
Our notability policy is specifically writting to declare those people notable through them fulfilling a structural need. I also don't see a problem with Wikidata including family trees. Most of the existing websites for family trees use unfree licenses (and that includes Wikitree). Wikidata's way of being able to specify which statements on an item are backed by which references also seems better to me than most of the other websites that list sources of a page without saying which statements are backed by which source. There's also a lot of synergy by having a database that describes humans in multiple context. If you have citiations from someone who lived in the 18th century and who used two different names, having family data can sometimes help to find out that two people are the same.
I don't see a problem when you don't put the information about "Bertha is the daughter of Sir John Bloggs" into Wikidata, but part of being a structured database means that this information would be added through new items. ChristianKl (talk) 16:00, 10 November 2017 (UTC)
  •   Comment at the end of this discussion it would be really helpful to have a summary of the preferred means. If there is a difference where comment is needed, then highlighting that difference would be good. We have been less than perfect in summarising these discussions, and moving to help pages. We seem to keep expecting people to read long blown-out conversations to work out how to enter data, and some (many?) days I just want to do the data entry.  — billinghurst sDrewth 21:34, 9 November 2017 (UTC)
  • In most cases I would expect that currently new users look at the way similar items are structured and try to follow the example. Even if we would have help pages it's hard to make them visible to a user when they would need it. ChristianKl (talk) 15:10, 10 November 2017 (UTC)
  • I just looked at whether there's a tooltip for "No Value" and "Unknown Value". Maybe we could show an explanation of the meaning of the terms as a tooltip? Do you thinkt hat would help? ChristianKl (✉) 18:58, 10 November 2017 (UTC)

Merge sport association (Q1531833) and sports governing body (Q2485448)?

Technically it would be simple to merge sport association (Q1531833) and sports governing body (Q2485448), but since there are plenty of backlinks to both items, I would like to ask the community for confirmation that those items should indeed be merged. Thanks for advice, MisterSynergy (talk) 12:42, 10 November 2017 (UTC)

I   Support merging, there's no different in the actual organization definitions, if both are not mergeable then why we merged two former organizations to one WBSC? --Liuxinyu970226 (talk) 05:01, 11 November 2017 (UTC)

electoral district (P768)

As long as electoral district (P768) is a qualifier how can I use it in a template like Template:Infobox officeholder (Q5830052)? Which, in Norwegian wikipedia, is collecting the value from Wikidata. Breg Pmt (talk) 23:04, 10 November 2017 (UTC)

You need to use LUA if you want to use Wikidata properly in infoboxes. Sjoerd de Bruin (talk) 23:07, 10 November 2017 (UTC)
Thats ok for Properties, but as far as I can understand it is not possible to use Qualifiers only Properties. Ref [Jon Mannsåker https://no.wikipedia.org/wiki/Jon_Manns%C3%A5ker] using [Infoboks officeholder https://no.wikipedia.org/wiki/Mal:Infoboks_person_med_verv].where Properties on wikidata are giving values to the infobox and where electoral district (P768) can not be collected from Wikidata as it is a qualifier Pmt (talk)
@Pmt: This is a known limitation of the "statements" parser function. I have raised qualifier support as a feature request at meta:2017 Community Wishlist Survey/Wikidata/Further functionality for "statement" parser function. For now you'll need to use no:Modul:Wikidata to get qualifiers. Deryck Chan (talk) 11:44, 11 November 2017 (UTC)
@Deryck Chan: Thank you very much. Pmt (talk) 11:52, 11 November 2017 (UTC)

Servant of God

Can anyone merge Q27808195 and Q869974? 77.253.40.96 09:39, 11 November 2017 (UTC)

No, they should not be merged as they are two different concepts. — Ayack (talk) 10:13, 11 November 2017 (UTC)

distinction: podcast (distribution) -- podcast (program)

The concept podcast (Q20899) has for the last years described the means of distribution. Last year I created podcast (Q24634210) a subclass of broadcasting program in analogy to radio program or television program. This conceptual distinction was introduced as a fix so that individual podcasts could be classified as an instance of a program for broadcasting rather than an instance of a technical distribution channel. Podcast has recently been modified and also made a subclass of broadcasting program. I reverted this for now. Though I did it before fully apprehending that a large number of podcasts have been in the meantime been classified as an instance of audio podcast (Q24633474) (itself a subclass of podcast as distribution). Obviously there is a problem with communicating the distinction. Likely it's the name since many will naturally think of "podcast" or "audio/video podcast" as the correct object class rather than of "podcast program".

So I wonder should the conceptual distinction be resolved by merging both concepts, distribution and transmitted content, in Q20899 (podcast)? It would be a bit like radio and radio program being just one object, but maybe it's appropriate for podcasts. If instead the conceptual distinction should be preserved, how can the difference between podcast and podcast program be better communicated?

--Toaem (talk) 12:29, 11 November 2017 (UTC)

Help with merging

Could somebody please merge Q43216992 (Catherine Neill) with Q26262898 (Catherine Annie Neill)? I tried to look up how to merge them myself but I am unfamiliar with Wikidata (I mainly stick to the English Wikipedia) and the instructions were a little complicated for me. Thanks for your help. 97198 (talk) 10:11, 16 November 2017 (UTC)

  Csigabi (talk) 10:13, 16 November 2017 (UTC)

This section was archived on a request by: Matěj Suchánek (talk) 10:24, 17 November 2017 (UTC)

Suggestion: Automatically provide the functionality for Template:Interwiki extra (Q21286810) for all items with said to be the same as (P460) or permanent duplicated item (P2959), and make said to be the same as (P460) and permanent duplicated item (P2959) mirror to target wikidata entry

  • said to be the same as (P460) and permanent duplicated item (P2959) are used to represent the concept of same as or concepts that are identical, so when for example someone added Q12345 as a P460/P2959 to Q24680, then there should be a mechanism to automatically add Q24680 as P460/P2959 back to Q12345.
  • Because items stated as P460/P2959 are supposed to be same as each other, I think it would make sense if the function of the template Template:Interwiki extra (Q21286810) are to be automatically applied onto all the entry with this property.

C933103 (talk) 06:00, 5 November 2017 (UTC)

How do I set a link between a Wikipedia article and a Wikimedia Commons category?

What parameters should I use if I want to connect a media category on Wikimedia Commons with an item on Wikidata that mostly concerns Wikipedia articles? If I click on "set link 🔗" on Wikimedia Commons it requires me to make a new item, but I want to add these links to existing items. Is this wanted or is it preferred that I just continue linking the Wikimedia Commons categories on the bottom of Wikipedia articles? -- 徵國單  (討論 🀄) (方孔錢 💴) 12:04, 11 November 2017 (UTC)

@Donald Trung: See the section above titled "Proposed change to WD:N regarding Commons categories" for the current iteration of the discussion regarding what to do about these. Mahir256 (talk) 22:35, 11 November 2017 (UTC)

Changes to the global ban policy

Hello. Some changes to the community global ban policy have been proposed. Your comments are welcome at m:Requests for comment/Improvement of global ban policy. Please translate this message to your language, if needed. Cordially. Matiia (Matiia) 00:34, 12 November 2017 (UTC)

Defined Daily Dose (DDD) - question about this and other unit-based properties

Hi, I was informed about this by @Doc James: We now have a property defined daily dose (P4250) for "Defined Daily Dose" (defined item is defined daily dose (Q1182681)), which you can observe in use for amoxicillin (Q201928). He wants to import a few thousand of these now, which is good, but I think this is a wrong property. Since the DDD is its own "thing" but is now a sub-website of the World Health Organization collaborating centre (Q7950347), it should be an external reference like any other external ID. Perhaps the actual <number> in <units> should go into a generic "Dosage" property for age/bodyweight with this DDD as the reference for the generic "70 kg male dosage". I see parallels with how we use inventory number (P217) for inventory number (which by the way, I think is much less likely to get vandalized than the DDD). I think the value of external IDs is that you can include the formatter url so if the url for J01CA04 changes you only need to change it once. Doc James wants to upload 1000's of DDDs but how should he watchlist the changes? I suggested a listeria list with just the url and dose and then he can replicate the list in userspace on all wikipedias and ask his network of doctors to watchlist those. He wants an email alert if a value is changed on Wikidata. Is this possible? Jane023 (talk) 10:33, 7 November 2017 (UTC)

Thanks User:Jane023 I am not sure the benefit of have the DDD be a url for a website (if that is what you are proposing)?
Yes advice on how to watchlist is appreciated. Doc James (talk · contribs · email) (if I write on your page reply on mine) 15:09, 7 November 2017 (UTC)
No I totally get that you want to be able to have this key amount in the basic query as could be used in e.g. an infobox on Wikipedia. My concern is about using the full url as a reference. I am thinking maybe you need a two-step process - one ID in external ids (maybe as part of a larger external id family for sub-websites of the World Health Organization collaborating centre (Q7950347)?), and the other in the "normal" statements section. Also you will need to request a bot anyway to upload units as they can't be uploaded with quick statements. Jane023 (talk) 17:42, 7 November 2017 (UTC)
I am not sure if I understand what you are saying correctly, but if you are talking about the code then it seems like it is an ATC code given under Anatomical Therapeutic Chemical Classification System (Q192093). As far as I understand, that is something entirely different from DDD so I am not sure about what User:Jane023 you are trying to say.C933103 (talk) 17:38, 9 November 2017 (UTC)

@C933103: My point is that if you make a property for something supported on an external website, then it should be treated like an external identifier with a formatter URL (P1630), which both makes updates to urls easier AND it makes checking individual values easier. You seem to be objecting to the way defined daily dose (Q1182681) is defined. You can discuss that with James. Jane023 (talk) 10:01, 12 November 2017 (UTC)

Proposed change to WD:N regarding Commons categories

d:Wikidata:Notability currently contains the following regarding Commons categories:

  • An item with only a sitelink to a category page in Wikimedia Commons is not allowed on main article items. However, it is allowed to link Wikimedia Commons categories with categories in other Wikimedia sites in items.

I believe this is unhelpful. As I understand it, the original intention of this clause was based on the idea that in general Wikidata items should reflect "clearly identifiable conceptual or material entities", and therefore to exclude main items that correspond to Commons categories that reflect intersections of such entities.

If that is the intention, I believe it would be better to say so directly, eg:

  • An item may be site-linked to a Commons category, but such items should only be created if they relate to a distinct identifiable conceptual or material entity. Items should not be created for Commons categories that can be described as an intersection of existing entities.
    A Commons category should sitelink to a category-type item on Wikidata if such a category item exists; but if such a category-type item does not exist, it is acceptable for a Commons category to be sitelinked to an article-type item.

This would make much clearer that sitelinks to Commons categories are welcome. In fact, it is a current urgent priority to identify as many of the entities behind Commons categories with Wikidata items as possible, to prepare the way for Structured Data on Commons: in order for information to be described by the Structured Data project, it has to have a corresponding entity on Wikidata -- it is important that those entities exist, and WD:N should be positive about creating them.

As for the second sentence above, there seemed to be support for it as a proposition in the threads last month on both Commons Village Pump and Wikidata Project Chat examining the most recent numbers on links between Commons and Wikidata. It essentially is a reflection of current reality: there are now over 120,000 more Commons categories linked to article-type items than category-type items here (539,071 vs 412,231), a trend which has been steadily increasing; and many more Commons categories which could be so linked. Such sitelinking makes templates work better and more efficiently; it also means interwiki links can be migrated to Wikidata; it is also more stable, because if pages get merged or split the sitelinks automatically move too. It's also a reflection that there is really no desire or value or point in creating a category item here, if its only function is to point to an article item: better (more reliable, less indirect, less potential for approximation fuzziness) to link to the relevant article item directly.

WD:N currently references the 2013 discussion on Commons links as justification for the current text. That discussion was never really very satisfactory. The outcome selected was not the one the heaviest majority of participants supported; it never had buy-in from Commons; and it is not the path that has ultimately been followed in practice. For that, and for the arguments above, I believe it would now be better to mark that discussion as {{Historical}}, remove reference to it, and move on. We're in a far better position now that we were in 2013 to assess the reality of the options and their consequences.

Proposed. Jheald (talk) 13:57, 7 November 2017 (UTC)

  •   Question Reads good in general, but I am concerned about Commons’s notability criteria. I do not have much experience with that project, but my impression is that it is much easier to permanently have promotional media files and thus a category over there than it is in any other Wikimedia project (including Wikidata). How does the Commons community make sure that this wouldn’t be a simple path to secure a promotional item at Wikidata as well? —MisterSynergy (talk) 19:07, 7 November 2017 (UTC)
As I understand it (others may correct me), Commons’s notability criteria is intentionally weak. If private entities want to release free photographs about themselves, Commons basically sees that as a good thing. For that media to be classifiable and searchable by the Structured Data system, it would require items here to hold the relevant information. If as Wikidata we do not wish this, the best way might be to attach a rider in a footnote to the clause above, excluding a list of certain limited types of entity -- eg individual Commons users not otherwise notable -- if we want to do that. Jheald (talk) 19:17, 7 November 2017 (UTC)
  • An example of how category are created in wikimedia commons is that, I have seen many aircraft registration category, where aircraft photo files are categorized by their registration number, so most of those individual aircrafts have their own registration number. If the aircraft changed hand or changed registration number in its life there might even be more categories. Given the sheer amount of aircrafts flying on the sky now and most of these aircraft went from freshly manufactured to retired uneventfully, I believe they aren't likely to be qualified as notable in most language variants wikipedia. Will these be qualified to create an item in wikidata? C933103 (talk) 20:55, 7 November 2017 (UTC)
    Probably we should ping c:Commons:WikiProject Aviation to see if they have thoughts or comments on this question.
c:Commons:WikiProject Aviation notified, at c:Commons_talk:WikiProject_Aviation#Wikidata_items_for_individual_registrations Jheald (talk) 15:01, 8 November 2017 (UTC)
What is true is that c:Category:Aircraft by registration currently has 74,487 sub-categories, each for a particular registration.
Do these deserve entries here? I'd suggest that at least some of them do -- for example if only a small number of a particular type are preserved, it probably does make sense to have an item on each one, with details of its individual history. Similarly eg for preserved railway locomotives, and for ships (even not preserved) since they are so individual. For some historic types, eg c:Category:Avro Lancaster, it does seem that there are only very few sub-categories devoted to individual airframes. On the other hand for aircraft types in present-day or recent service, eg c:Category:Boeing_737, the number of subcategories for individual registrations is huge -- look at the counts of sub-sub-categories of c:Category:Boeing 737-300 by registration. Looking at a representative individual one of these pages, eg c:Category:B-5063_(aircraft), there is information here that might benefit from being stored in a structured way, for example: the type of the aircraft, the date it was built, its operator and dates in use with that operator, a serial number for its airframe (and hence other operators and other registration numbers it carried). I think a case might be made that this could indeed be useful to store on Wikidata, if we as the Wikidata community were prepared to permit it. At any rate it might lead to a distinct item here for each airframe. There is also the question of whether it is useful to be able to see all pictures of a particular plane, and to be able to drill down to those by registration number. If so, does the Structured Data team anticipate that that would require an item for each registration number? Or alternatively, will some detailed identifying information (like a name, for a person; or a registration number or specific identification) be held locally on the CommonsData wikibase, perhaps as a qualifier with a string value on a "depicts" statement, and do the Structured Data team anticipate that that too would be discoverable via the envisioned search/refine process? I think she's away for a few days, but @SandraF (WMF): might be able to advise.
Perhaps if we could identify examples of types of such information that it might be possible to hold with local strings on CommonsData, that might be used to create a considered list of types of entities that don't need separate items here -- the sort of exception list I was thinking about in my response to MisterSynergy above. Jheald (talk) 14:19, 8 November 2017 (UTC)
One question here is how easily we are going to want pictures of people like c:Category:Maarten_Dammers or c:Category:Lydia_Pintscher to be findable. I think there are at least a couple of issues here. One is name standardisation. An alternative to using depicts (P180) <some Wikidata item for the person> to describe a file in the Structured Data system might be to use a mechanism with a qualifier that was string-valued, something like depicts (P180) human (Q5) // author name string (P2093) "Maarten Dammers" (except it would just be "name string" not "author name string"). But if we're not using Wikidata items, we lose the advantages of a controlled vocabulary -- if in another photo you get tagged as "Multichill", that won't be recognised as an alias. Similarly for variants of American names, eg with an without middle initials. There would also need to be a special mechanism to search for name fragments, eg just "Dammers"; and one wouldn't necessarily be able to specify what was a surname or a first name. A second issue is that your very inclusion in c:Category:Wikimedians in the Netherlands is a representation of information to make you findable. With a Wikidata item, that can be translated to occupation (P106) Wikimedian (Q41546637), country of citizenship (P27) Netherlands (Q55) -- and one presumes that, starting from "Wikimedian" the search/refine system might be able to suggest (amongst other options) that you could be looking for instances of people with that as an occupation; then country as a refinement direction; then Netherlands as a country value; then name as a refinement direction; then Maarten Dammers as a possible name value. But if the name were held just as a string, then none of that would be possible. On the other hand the inclusion of regular people in Wikidata (not just celebrities) raises a whole new dimension of BLP, privacy and data protection issues - something I know that User:Lydia Pintscher (WMDE) has been quite wary of. Perhaps there could be a special ad-hoc search method for Wikipedians. But generally, to the extent that we want to make images of individuals discoverable on Structured Data by what those individuals are (currently achieved on Commons through the category hierarchy), then it seems to me that those individuals are going to need individual items. Jheald (talk) 18:04, 8 November 2017 (UTC)
  •   Comment I think that the proposed text may need to be a little more user friendly as "distinct identifiable conceptual or material entity" ... is well <shrug>. For this discussion, it might also be worthwhile producing some examples to better contextualise what is looking to be achieved where the language used may not be well-interpreted. Part of this discussion should also cover 1) is it expected that every category should be linked from Wikidata? If no, what categories are we looking to exclude? If yes, what categories do we see may be problematic or inflammatory to Wikidata? Examples of each would be useful.  — billinghurst sDrewth 00:31, 8 November 2017 (UTC)
  •   Comment Maybe a threshold should be existence of independent references. For example people, places or taxons should have some identifiers or sources which were not created by the subject. Companies should have some sources other than company website. The sources would not need to be provided at the time of item creation, but items without independent sources should be removed. --Jarekt (talk) 02:35, 8 November 2017 (UTC)
I am cautious about writing that into policy. For example, a building that people have photographed might not be listed as an official historic heritage object, yet may still be a distinct structure with eg a street address, geo-coordinates, architectural style, demolition date etc, all of which one might want to record. I suspect that, particularly for a building, if it is distinct and distinctive enough that people have photographed it and named a category for it (and regard it as an entity in its own right), then it is probably worth having an item for here -- and if people can source further statements about it, then so much the better. Jheald (talk) 15:10, 8 November 2017 (UTC)
  •   Comment There are two points to the proposal. The first is that Commons should be able to rely on Wikidata for providing its interlanguage links, and to provide information for templates, without concern that somebody will delete a "category" link from an "article" item, on the misguided idea that this is against Wikidata policy. I haven't seen this happen for quite a while, and its seems that it's not something that's argued about anymore. There was a second RFC that also led to no conclusion: Wikidata:Requests for comment/Category commons P373 and "Other sites". There shouldn't be a notability problem in these cases, because items with interlanguage links should always have site links to other projects. The second point would be whether new items can be created on Wikidata for Commons categories. This may be a possible requirement for the structured data project, since it seems structured data won't give Commons a facility to create private items of its own. If a Commons category is to be replaced by a link to multilingual Wikidata item, the item will need to be created on Wikidata. Ghouston (talk) 04:23, 8 November 2017 (UTC)
    • But then, how about if another mediawiki projects with equally weak notability requirement emerges? For example if meta:NonFreeWiki become a reality and then all the commons category are duplicated to it, there will need to be some form of link between them. Could wikidata provide such a link in cases like this?C933103 (talk) 17:29, 9 November 2017 (UTC)
  • Another apparently large category tree in commons: commons:Category:Categories by camera C933103 (talk) 12:25, 10 November 2017 (UTC)

I   Support the change. Especially with the planned integration. ChristianKl (✉) 01:41, 12 November 2017 (UTC)

Bots vs assisted editing

Do we need a policy that defines the boundary between a bot and assisted editing (sometimes involving the API or pywikibot) more clearly? Wikidata:Bots sets out the policy on the use of bots and how to get approvals for bot tasks, but doesn't actually say what we consider a "bot" in the way w:Wikipedia:Bots does. Deryck Chan (talk) 23:11, 9 November 2017 (UTC)

Why would we need a policy? What is its purpose, what should it solve and what are its risks? Only floating ideas without any justification is imho a bad thing. Thanks, GerardM (talk) 06:47, 10 November 2017 (UTC)
QuickStatements can do the same things as a bot. Normal edits are easily undone when someone disagrees. Undoing bots who make a lot of edits is harder and as such it worthwhile to discuss large batches of edits before they are made. I   Support writing a policy. ChristianKl (talk) 13:16, 10 November 2017 (UTC)
I think we indeed need such a policy (or, to be exact, likely an extension of the existing policy to such edit). We also need to decide what we want to do with bots approved on the projects making edits on the projects (likely nothing, but it should be on the policy as well).--Ymblanter (talk) 14:15, 10 November 2017 (UTC)

@GerardM, ChristianKl, Ymblanter: A boundary between assisted editing and bots will be useful because:

  1. Editing Wikidata often involves making mass changes that involve dozens or hundreds of edits each time an editor wants to make a major change. This typically involves setting up a tool for multiple pages and committing an entire batch of edits in one go. This means that a significant proportion of work on Wikidata falls in the grey area between what other wikis define as a "manual edit" (one edit per human decision, usually through the web interface) and a "bot edit" (setting off a bot to crawl pages and edit them automatically without human attention). English Wikipedia defines the middle range as "assisted editing" and such a definition will be more useful on Wikidata than Wikipedia.
  2. If an editor makes a batch of edits whose magnitude is smaller than the threshold, they will feel assured that their edits will be assessed for their merits and the editor will not be punished for "running an unauthorised bot". --Deryck Chan (talk) 09:28, 11 November 2017 (UTC)

@Deryck Chan: I do not care what other Wikis define. What I care about is Wikidata. I have no interest in more regulation creep. It serves no purpose and there is no proof that it will help us achieve a better result. When I add data to Wikidata I typically do this based on information from Wikipedia, they have their error rate and I am unapologetic that errors are introduced. There is potential in improving both Wikidata and Wikipedia when we collaborate. We don't. At Wikidata we have not reached the tipping point where we have sufficient data to make a difference. Your proposal will imho be detrimental to our success. Thanks, GerardM (talk) 10:16, 11 November 2017 (UTC)

@GerardM: I'm afraid the opposite of what you worry about is also happening. Some people are discouraged from making mass edits because they feel that it would be considered bot editing. I agree we shouldn't have regulation creep - Wikimedia project policies should be as decriptive as possible. But if we have an assurance written down (not necessarily on Wikidata:Bots) that says "batch edits involving fewer than X edits to Y items over a time period of Z don't require bot flags" that will give people confidence. Deryck Chan (talk) 11:16, 11 November 2017 (UTC)

Instead of writing policy, which would artificially restrict all active users from editing Wikidata, I would recommend to:

  • Write and publish a tool to simplify batch undoing.
  • Write a csv import special page as a part of wikibase software.
  • Rewrite RecentChanges special page to show edits grouped by some kind of batch_id (QuickStatements 2 does this).

Right now manual editing Wikidata is like writing Wikipedia articles word-by-word: go to edit mode, add exactly one word, save, go again to edit mode, add exactly one word, save, etc. If somebody wants to write 1000 words (with tools, obviously), it would require requesting flooder flag for a short period of time. Where would be Wikipedia now? --Lockal (talk) 11:02, 11 November 2017 (UTC)

@Lockal: Great ideas, particularly the one on batch undoing. Is it on the 2017 Community Wishlist Survey yet? Deryck Chan (talk) 11:16, 11 November 2017 (UTC)

  Support to improve clarity, legal security for everyone and to reduce disputes on what is ok and what isn't. (Rules can also increase liberty.) --Marsupium (talk) 11:38, 12 November 2017 (UTC)

IP adress making modification on his own beliefs

I want to point this IP author of this modification: https://www.wikidata.org/w/index.php?title=Q5482740&diff=588260041&oldid=588249764 who does several similar modification. As a computer engineer as a diploma and as a passionate about computer science and science is general, I am really very confident that this precise modification is totally wrong. (and as nearly as confident that the same kind modification they did are wrong to) For the ones who are not aware of such things and want to make their own opinion, I suggest to read the wikipedia article about the subjects where he undo everything and check that he is incorrect. for example for this precise diff https://en.wikipedia.org/wiki/Computer_scientist and https://en.wikipedia.org/wiki/Programmer . I tried to established dialog with this ip just now, but if the dialog fail I suppose I will have to ask for a bloc of the ip for lack of discussion. Xavier Combelle (talk) 02:11, 5 November 2017 (UTC)

@Xavier Combelle: Hmm... I'm kind of torn about this. Full disclosure - I am a computer scientist by training and hold two degrees in "C.S." and engineering. If the question was, "Is programming a subclass of computer science?" I would say it is clearly true. However, the moniker "computer scientist" implies a pedigree or title that may or may not apply to "programmer." So to clarify - if someone who's never taken a CS class in their life learns a lot from StackExchange and picks up PHP from an online tutorial, and builds a cool script, they can be called a programmer. Would I call them a computer scientist? No. That person now knows programming. Is that person utilizing computer science skills and concepts? Yes. Is that person necessarily a computer scientist? No. Does this distinction make sense? -- Fuzheado (talk) 12:35, 5 November 2017 (UTC)
I think calling programming a subclass of computer science is like calling truck driving a subclass of mechanical engineering. A basketball player uses concepts of physics but that doesn't make him a physicist or suggest that we should classify basketball as a subclass of physics. ChristianKl (talk) 12:46, 5 November 2017 (UTC)
I beg to differ. Ethan "Bubblegum" Tate. SharkD  Talk  19:11, 5 November 2017 (UTC)
Hello,
For your information, I also sometimes edit connected, as Hercule. That's not because I edit not connected that Xavier Combelle can start a discussion by anticipating a block if I don't agree with him...
Don't forget that Wikikada is not a database just for english speakers, but for all the Wikimedia projets. As a french native speaker I'm argue with the french definitions.
"Computer science" = "informatique". "Computer scientist" = "informaticien". There is no doubt that a "programmeur" is an "informaticien", that means someone working on the "informatique" field.
The question is not about someone leaning how to make some few programming, but about someone whose profession is to program. This person is a computer scientist, even if he can't be compared with a computer science doctor. Comparing the relation between a programmer and the computer science to a baseball player and physic is an insult to programmers. A baseball player can have no idea about physic whereas a programmer (as a professionnal) must know computer science to have a diploma to justify their professional skills.
Regards.
--90.63.244.220 08:42, 6 November 2017 (UTC)
I consider that Xavier Combelle can't impose me to argue previously to revert his changes, because it's him that change an established situation. He must establish a consensus to change it.
--90.63.244.220 08:48, 6 November 2017 (UTC)
Having a degree in programming isn't required to work as a programmer. These days people work as a programmer after taking a 3 month coding bootcamp in which they aren't trained in what's considered computer science. ChristianKl (talk) 15:12, 6 November 2017 (UTC)

There is the same problem with Q81096. Engineer is a scientific profession. So the "engineer" must be a sub-class of "scientific", even if it has nothing to do with science academic research. Otherwise there can be not structured hierarchy between professions. If this must be changed it can't be by removing the relation between these elements, but by creating more sub-elements to make the relation more precise (both are scientific professions). --90.63.244.220 08:57, 6 November 2017 (UTC)

Even if having a "Bachelor of Science" would be sufficient for being a scientist, not every engineer has a "Bachelor of Science". Plenty of engineers have a "Bachelor of Engineering". ChristianKl (talk) 15:15, 6 November 2017 (UTC)
Hi User:Hercule I am french too. In french there is no distinction between the name given to Computer science and computer engenering that doesn't mean the difference doesn't exists. The profession of a computer scientist is not to make program. It's profession is to make mathematical reasoning about programs. It's discipline is a subset of mathematics. As they usually implement their mathematical reasoning, computer scientist are de facto programmers. And as such computer scientist are basically a subset of programmer. But the vast majority of programmer are not computer scientist. For engineer the problem is pretty similar, an engineer use science, he doesn't build science and as such is not a scientist. One of the main criteria to be qualified of scientist is to have scientific activity which concretely translate by publishing in pair review journals. I actually found both of the problems by looking on the persons who had profession as a subset of scientist where I found a lot of persons who never published a scientific article but were programmers, to start with Jimbo Wales (Actually I did not check if Jimbo Wales actually published a scientific article, but he is for sure not known to be a computer scientist).
Please Hercule read the wikipedia articles on the subjects they should be clear.
Note that User:SharkD beg to differ is a reference to the humoristic comics futurama is not intended to be take seriously.
If a consensus should be drawn from this state of discussion, we are three persons, User:Fuzheado,User:ChristianKl and me to agree and User:Hercule is singled out, so the consensus result should be on our position and not Hercule one. Xavier Combelle (talk) 21:15, 6 November 2017 (UTC)
Ok, I understand your point and agree you're at least 3.
But I think that there is a problem with the removal of engeneers and programmer for the "science related professions". Do you see a way to have such a relation established on the professions hierarchy ?
--90.63.244.220 13:39, 7 November 2017 (UTC)
What exactly do you mean with the term "science related professions"? Is it something besides needing a "Bachlor of Science" to practice the profession? ChristianKl (talk) 23:39, 8 November 2017 (UTC)
"science related professions" means that this is a work that needs science skills ("Computer science"), by opposition of other professions that are "legal profession" and other that are "creation professions". In french we call it "profession scientifique", to distinguish it from "profession juridique", "profession artistique",...
--Hercule (talk) 08:45, 9 November 2017 (UTC)
@Hercule: In German we have no trouble to speak of Rechtswissenschaft (science of the law) which we model in Wikidata as jurisprudence (Q16549001) which subclasses "social science". On what ground are you arguing that people use the skills they learn in their studies of "Rechtswissenschaft" (science of the law) aren't practicing science based skills but computer science somehow implies a scientific skill? ChristianKl (✉) 20:55, 12 November 2017 (UTC)

@Hercule You can try to create a class named "science related profession" and include programmer and engineer in it if you want. Just make it different of scientist (Q901) and not a subclass of it. Xavier Combelle (talk) 20:41, 12 November 2017 (UTC)

Edit summaries displaying the change

I didn't find anything in Phabricator or the archives, so I wonder: Has it been considered that edit summaries automatically display (part of) the content change, like so "Label X was changed from Y to Z"? Currently they only show "Label X was changed". Jo-Jo Eumerus (talk, contributions) 14:44, 11 November 2017 (UTC)

There are lot of areas of Wikidata that provide opportunities for improvement and I think that edit summary descriptions haven't gotten that much work yet. ChristianKl (✉) 15:49, 11 November 2017 (UTC)
I was thinking that this may be an useful proposal for the Community Wishlist Survey, to make it easier to detect vandalism and problem edits. I wanted to make sure that this hadn't already been discussed and judged impractical here, however. Jo-Jo Eumerus (talk, contributions) 16:08, 11 November 2017 (UTC)
There is a user script which might be of interest: put mw.loader.load( '//www.wikidata.org/w/index.php?title=User:Yair_rand/DiffLists.js&action=raw&ctype=text/javascript' ); // [[User:Yair rand/DiffLists.js]] to Special:MyPage/common.js and have a look at your watchlist, item histories, or recent changes. The script also offers filter options. —MisterSynergy (talk) 16:16, 11 November 2017 (UTC)
  Comment Wikipedia:Tools/Navigation popups (Q11305696) might also be useful here. On Special:Preferences#mw-prefsection-gadgets it says "Currently has no support for previews or diffs of Wikidata items or properties (pages in the main or Property: namespaces)", but recently diffs of the internal representation work. (BTW: How to update Special:Preferences#mw-prefsection-gadgets?) --Marsupium (talk) 11:28, 12 November 2017 (UTC)
I am guessing these system messages are the place. Jo-Jo Eumerus (talk, contributions) 15:59, 13 November 2017 (UTC)

Representation of average height as a histogram

In some cases a statement is about a partial statistic about the described entity, typically bins from a histogram of areas within some height limits for a municipality. This is the case for municipalities in Norway, we have the histograms but we don't have average heights.

The average height of an area could be described as several bins of area (P2046) ranging from and to a specific height (P2048). It is not given that bins are of equal size, and it is not given that they use some specific units. For time we have start time (P580) and end time (P582), but not for height. This must then be identified as a bin, possibly with determination method (P459) set to histogram (Q185020), but it would be clearer if there was a "bin for histogram". ("Determination method" is slightly wrong, for this purpose it should be "Description method".)

This would place the bins in area (P2046) and make the interpretation of that property more difficult. An alternative could be to make a property "area for average height", but a more obvious property would be "average height". Still we might not have an average height even if we have a histogram over the bins.

How should this be represented, any ideas? I wonder if this is best represented as statistics and not as metadata as such. This is about the fine distinction whether the qualifier identifies a statement or whether it describes a (composite) selection. Jeblad (talk) 09:57, 13 November 2017 (UTC)

Can you give a concrete example? ChristianKl (✉) 11:33, 13 November 2017 (UTC)
Data from Statistics Norway for one municipality is like below, but this is just an example of histogram data. Check Wikipedia for a description of a w:histogram. Jeblad (talk) 12:39, 13 November 2017 (UTC)
Example data from Halden, Norway
Areal (km²), efter region, height above sea, and area"
Municipality Height (m) Area (km²)
Halden
0-59 66.83
60-159 315.44
160-299 258.72

Wikidata weekly summary #286

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

You can see all open tickets related to Wikidata here.


Coin og Museu Nacional d'Art de Catalunya

I find a lot of artworks placed at Museu Nacional d'Art de Catalunya who has instance of (P31) of coin (Q41207). Does anyone know if this is correct?. Seems that they are created by a bot. Example Q27519005 Pmt (talk) 23:56, 11 November 2017 (UTC)

If you want to find out it would make sense to ping the bot owner. ChristianKl (✉) 00:35, 12 November 2017 (UTC)
I can confirm that the museum indeed has a coin collection. —seav (talk) 11:43, 14 November 2017 (UTC)

Modelling Performing Arts Productions

Dear all,

In the context of the WikiProject Performing Arts, we are presently working on a pilot ingest of data about performing arts productions. We have now largely mapped the data from a first dataset and identified areas where the ontology on Wikidata needs to be complemented in our view. We have gathered the mapping information and our reflections on the data mapping page. Please have a look and comment here or on the talk page of the data mapping page. I will now start creating new classes and propose new properties, starting with the less problematic ones. Topics covered are:

  • Adopting a more nuanced approach towards modelling FRBR Group 1 classes (literary, musical, choreographic works)
  • Modelling of Performance Works
  • How to ingest data about character roles
  • Modelling theater seasons and how to point to them (pointer to time intervals)

--Beat Estermann (talk) 20:15, 12 November 2017 (UTC)   WikiProject Books has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.

Congratulations, this is really a good work and I hope this can be used later as example for data import: we have a clear analysis of 2 different classifications, and a preliminary discussion before import. For the discussion itself I propose to use the talk page of the proposition page in order to keep everything at the same place. Snipre (talk) 23:20, 12 November 2017 (UTC)
  WikiProject Cultural heritage has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead.  Notified participants of WikiProject Theatre
Pinging the two projects again; apparently, the ping template won't work if the participants pages contain bullet points.--Beat Estermann (talk) 07:33, 14 November 2017 (UTC)

Sorting of values

Currently, there's an open request for a property to store information about how to sort https://www.wikidata.org/w/index.php?title=Wikidata:Property_proposal/sort_by_qualifier . It would be great to have more examples of cases where things currently aren't ordered well and we could profit from a better sort order. If anybody of you notice interesting cases, please add them to that discussion. ChristianKl (✉) 14:30, 14 November 2017 (UTC)

Vectors are Matrices

  Notified participants of WikiProject Mathematics

scalar (Q1289248)   vector (Q13471665)   matrix (Q44337)   tensor (Q188524).

Actually it depends on the defention of the respective items. Could we anyway tag this with subclass of (P279)? --Bigbossfarin (talk) 16:33, 14 November 2017 (UTC)

To the extend that it depends on the definition of the respective items, about how we start by making that definition more clear? ChristianKl (✉) 16:47, 14 November 2017 (UTC)
There are multiple definitions to be dealt with;
  • scalar   row vector, column vector   matrix   multi-dimensional matrix ~ tensor
  • scalar   vector, covector   tensor
I don't know which are considered primary definitions here. Arthur Rubin (talk) 17:02, 14 November 2017 (UTC)
I think it is pretty mixed up, scalars and vectors are allready tagged as tensors. The two relations you mentioned do not contradict, if you consider row vectors as a representation of covectors. (as defined in German) Bigbossfarin (talk) 18:08, 14 November 2017 (UTC)

Use cases for no value

What do you think, is it appropriate to use the "no value" for statements, that could eventually be filled with custom values? E.g. for a living person, there is no value for date of death (P570), but at some time there could be one. --MB-one (talk) 20:38, 7 November 2017 (UTC)

Better use "no value" only when the statement can be referenced and will stay valid in foreseeable time. -- JakobVoss (talk)
For a living person you can use floruit (P1317) 2017 if you want to note down that they are currently alive with an appropriate source. In other cases where there's no special property and you want to state that an organisation still exists I think it's oaky to use "no value" even if the organisation could get disolved in the future. ChristianKl (talk) 21:45, 7 November 2017 (UTC)

  Comment Where it is factual. Forcing a "no value" would actually apply a date of death which is incorrect. Having no date of death is the only means to determine someone is living, there is no other system means to do that. We also use "no value" where we have done authority identifier checks and find no value exists; note that it is required to qualify that with retrieved (P813).

On a similar note, do not apply "unknown value" just because you don't know it. It is quite problematic when people systematically apply selected vague dates of birth/death as they flow through the system, AND it doesn't allow those empty values to be researched as they are already filled with what equates to junk data.  — billinghurst sDrewth 00:04, 8 November 2017 (UTC)

I am currently adding "no values" regularly for the Thai monuments. About half of the registered Thai historic site (Q16438244) have been officially announced in the Royal Gazette of Thailand (Q869928) and becoming effective by the date of publication, but the other half is also included in the database by the Fine Arts Department (Q1416884) and have a Thai cultural heritage ID (P1626) already. Thus for those I set start time (P580) to "No value". Ahoerstemeier (talk) 17:29, 8 November 2017 (UTC)
@Jura1::As of now, there's no way to determine if a person is currently living, or if the date of death (P570) isn't set yet. Hence, I would propose the use of "no value" to indicate that, there is indeed no date of death, because the person is currently alive. More generally speaking, creating information by omitting information, doesn't seem to be a sensible approach to me. --MB-one (talk) 09:31, 14 November 2017 (UTC)
How would you reference such a statement? What happens when someone uses the most recent rdf dump (that is slightly outdated)?
--- Jura 12:26, 14 November 2017 (UTC)
@MB-one: see above.
--- Jura 10:55, 15 November 2017 (UTC)
@MB-one: I respectfully disagree. Starting to push a "no value" for an open end date doesn't seem to be a good practice. If we start that on death dates, every other qualifier for end date is problematic.

No date of death usually would mean living, or data missing. There are means to manage missing death dates where obviously dead (immortals aside). At English Wikisource, where no death date, we calculate up to 120 years max. age for living, and label accordingly. Forcing a no value, gives death date a value, and makes it harder to find missing values.  — billinghurst sDrewth 21:46, 14 November 2017 (UTC)

@billinghurst: How would this make it any harder to find missing values? If anything it would make it easier, by excluding living people from the list of items, potentially missing date of death (P570). And yes, this does technically apply for all end date properties. --MB-one (talk) 22:04, 14 November 2017 (UTC)
If date of death (P570) is not known use "unknown value" . If the person is immortal (like Jack Harkness (Q561140)) use "no value". Otherwise don't add date of death (P570). --Succu (talk) 22:42, 14 November 2017 (UTC)

Random URL parts

Hello. We have a few external IDs which have a formatter URL (P1630) with a small part that can be changed randomly without breaking the link. For example:

External ID formatter URL (P1630)
NFL.com ID (former scheme) (P3539) http://www.nfl.com/player/n/$1/profile
AustralianFootball.com player ID (P3546) http://australianfootball.com/players/player/a/$1
FCI rider ID (P4278) http://www.federciclismo.it/it/person/p/$1/

As we can see, some of them use a single letter that is whether randomly picked or chosen to match the first one in the name of the database. But maybe should we try to make these things obvious and use the same wd or wikidata everywhere. This is really a detail but that would be nice, wouldn't it? Thierry Caro (talk) 14:10, 13 November 2017 (UTC)

When I've seen that I've encouraged the use of '_' as the random string in the formatter URL to indicate it's arbitrary (if that works). But I'd be fine with 'wd' too. 'wikidata' seems long. ArthurPSmith (talk) 15:49, 13 November 2017 (UTC)
The problem with using _ as random part is that's a valid character that can appear in urls. ChristianKl (✉) 23:04, 14 November 2017 (UTC)
@ArthurPSmith, ChristianKl: I guess I'll do this with wd and add as Wikidata usage instructions (P2559) on formatter URL (P1630) the following note: You may use 'wd' whenever part of the URL can be randomly modified without breaking the link.. Thierry Caro (talk) 23:15, 14 November 2017 (UTC)

Get Wikidata item given site ID and page name

Hi, given a site ID (eg. enwiki) and a page name (eg: Italy) how can I get automatically Q38? I'd like to do that work with more than just one page, so this work should be automatic and quick. Thank you, --★ → Airon 90 17:13, 14 November 2017 (UTC)

API requests such as https://www.wikidata.org/w/api.php?action=wbgetentities&sites=enwiki&titles=Italy&format=json would do the job, but you still have to parse the JSON object. See mw:Wikibase/API#wbgetentities for details. —MisterSynergy (talk) 17:26, 14 November 2017 (UTC)
That's not a problem. I would use Python to do that work ;) Thank you very much, MisterSynergy! :) --★ → Airon 90 18:12, 14 November 2017 (UTC)
There is also Special:ItemByTitle with which you can do things like https://www.wikidata.org/wiki/Special:ItemByTitle/enwiki/Italy. --Lydia Pintscher (WMDE) (talk) 19:39, 14 November 2017 (UTC)
You can also use Pywikibot. What is your use case exactly? What problem are you trying to solve? Multichill (talk) 21:02, 14 November 2017 (UTC)

I found the quickest way for me was to use a spreadsheet with the WP article name as first column, second "P31", third simply some bogus like "a" (all without the " marks), and then copy/paste the whole thing to https://tools.wmflabs.org/wikidata-todo/quick_statements.php , say "enwiki" in the first field there, and let it start to run (which will do nothing), and then copy the whole thing again from the text field there after a couple of seconds when the tool has converted everything to item numbers. Maybe there's a tool that does just this without the workaround I use, but this is what I came up with by accident. Anvilaquarius --08:25, 15 November 2017 (UTC)

Have a look at pagepile. —MisterSynergy (talk) 08:27, 15 November 2017 (UTC)
This looks like what I want, but it doesn't really work. The output is not in table form, but a list with numbers and article names (which I don't want and don't need) and the alternative output in txt is without line breaks, neither is usable by c/p in my spreadsheet as quick as I would like to. --Anvilaquarius (talk) 09:23, 15 November 2017 (UTC)

FactGrid workshop in Berlin

Hey everyone :)

We are working with researchers from the University of Erfurt on a project called FactGrid. The goal is to build a wiki for researchers to collect data related to their research. It will use Wikibase as the underlying technology, just like Wikidata. We'll start with data related to the Illuminati based on the Gothaer Illuminati Research Database. You can find blog posts with more details and musings here. I also started a page on Wikidata about it here. My hope is that this will be one of the many Wikibase installations in Wikidata's ecosystem in the future that is highly connected with Wikidata and that Wikidata can partially import data from where wanted and needed. There will be a kick-off workshop at Wikimedia Deutschland's office in Berlin on the 1st and 2nd of December. It'd be great if we have a few more people there who want to drive this project forward and can help with their experience on Wikidata. If you'd like to attend please let me know. Details about the workshop are here.

Cheers --Lydia Pintscher (WMDE) (talk) 21:05, 14 November 2017 (UTC)

It took me a while to understand the value of pushing external Wikibase installation but I came around to agreeing that it's a good idea. External projects can host original research. At WikidataCon I also understand the need of some projects to be able to simply create a new property to experiment with it, which Wikidata can't serve because we need to put more effort into creating properties that are actually working over problem domains.
The memorandum of understanding reads well and I'm happy to take part in the workshop. ChristianKl (✉) 13:24, 15 November 2017 (UTC)

Facto Post – Issue 6 – 15 November 2017

Facto Post – Issue 6 – 15 November 2017
 

WikidataCon Berlin 28–9 October 2017

 
WikidataCon 2017 group photo

Under the heading rerum causas cognescere, the first ever Wikidata conference got under way in the Tagesspiegel building with two keynotes, One was on YAGO, about a knowledge base conceived ten years ago, if you assume automatic compilation from Wikipedia. The other was from manager Lydia Pintscher, on the "state of the data". Interesting rumours flourished: the mix'n'match tool and its 600+ datasets, mostly in digital humanities, to be taken off the hands of its author Magnus Manske by the Wikimedia Foundation; a Wikibase incubator site is on its way. Announcements came in talks: structured data on Wikimedia Commons is scheduled to make substantive progress by 2019. The lexeme development on Wikidata is now not expected to make the Wiktionary sites redundant, but may facilitate automated compilation of dictionaries.

 
WD-FIST explained

And so it went, with five strands of talks and workshops, through to 11 pm on Saturday. Wikidata applies to GLAM work via metadata. It may be used in education, raises issues such as author disambiguation, and lends itself to different types of graphical display and reuse. Many millions of SPARQL queries are run on the site every day. Over the summer a large open science bibliography has come into existence there.

Wikidata's fifth birthday party on the Sunday brought matters to a close. See a dozen and more reports by other hands.

Links

Editor User:Charles Matthews. Please leave feedback for him.
If you wish to receive no further issues of Facto Post, please remove your name from our mailing list. Alternatively, to opt out of all massmessage mailings, you may add w:Category:Wikipedians who opt out of message delivery to your Wikipedia user talk page.

Charles Matthews (talk) 15:00, 15 November 2017 (UTC)

Video game genre

I have questions about science fiction video game (Q27670585) and fantasy video game (Q42409239). Both of these are listed as instances of video game genre (Q659563). However, video game genres are supposed to be all about different types of gameplay, not setting. Is there a better thing that these should be sub-classes or instances of? Could someone clear this up for us? Thanks. SharkD  Talk  09:06, 5 November 2017 (UTC)

@SharkD: Category:Video games by theme (Q6849492) contains both Category:Fantasy video games (Q7801174) and Category:Science fiction video games (Q6339113) as sub-subcategories, so I'd welcome an item for "video game theme".   Notified participants of WikiProject Video games Mahir256 (talk) 21:57, 5 November 2017 (UTC)
I think that's a good interim solution. However, what should "video game theme" be a child/instance of? I don't know how to set this up. SharkD  Talk  06:53, 9 November 2017 (UTC)
The problem with genre, in general, is that it doesn't have any specific meaning or standard rules. It's really just another word for category, and is quite ambiguous. My view is that we have too many genre items for creative works, but not enough more specific properties. I don't know why we have many different science fiction (Q24925) items, such as science fiction film (Q471839), science fiction comics (Q1782964), et al. I have never seen this done this way elsewhere, and it makes no sense to me. science fiction (Q24925) is science fiction (Q24925) regardless of the medium. Danrok (talk) 16:26, 6 November 2017 (UTC)
Thanks for the responses. I have my own ideas about genre. They will probably never get implemented here. I think there are four main types of video game genre:
  • gameplay (shooter, RPG, beat 'em up, etc.)
  • setting (historical, science fiction, fantasy, modern/contemporary, etc.)
  • dramatic (comedy, tragedy, mystery, etc.)
  • purpose (advergame, serious game, educational game, etc.)
I also think you can combine these into nearly any combination you want. (role-playing shooter, dramedy, science fantasy, etc.)
But, I will continue to monitor this discussion to see what the rest of you think. SharkD  Talk  06:37, 9 November 2017 (UTC)
@SharkD: All values of genre (P136) are required to be instances of genre (Q483394), or instances of one of its subclasses. We could have subclasses of genre (Q483394) be for gameplay/setting/dramatic/purpose categories, and each individual "genre" being an instance of one of those and used as a genre (P136) value. Perhaps there should also be a more general class for "genre of video games", as a superclass for video game genre (Q659563), video game theme (Q42907216), etc. and a subclass of genre (Q483394). --Yair rand (talk) 15:10, 15 November 2017 (UTC)
That would be a big change affecting a lot of projects, so we would need their feedback too. SharkD  Talk  23:10, 15 November 2017 (UTC)

How to add a database to Wikidata?

I want to plot the geographical area in New Spain / Mexico since ~1776 to serve as a companion and data quality check on the Wikipedia articles on the "Territorial evolution of Mexico" and the "Territorial evolution of North America since 1763" -- and to compare with a similar plot for "Territorial evolution of the United States".

I have numbers that suggest I may be missing roughly 8 percent of the land area; see en.wikipedia.org/wiki/Talk:Territorial_evolution_of_Mexico.

I'd like to port my summary to Wikidata, then write some people who might know about this and ask them for clarification.

Suggestions on how to get started?

Thanks, DavidMCEddy (talk) 23:40, 12 November 2017 (UTC)

@DavidMCEddy: I'm not sure I understand what exactly you want to import.
Thanks for the reply.
I want to be able to create a table of events and land area involved that changed the size of New Spain and Mexico, like I posted in en.wikipedia.org/wiki/Talk:Territorial_evolution_of_Mexico (without the columns for the US). I'm currently interested in what was New Spain prior to 1821 and Mexico since.
To work backwards from the current status, I'd like to start with an entry for Mexico (= P150?) being 1,972,550 km2 (per the Wikipedia article on Mexico) with end time (P582) = “current”. Then we could change “|area_km2 = 1,972,550” and “|area_sq_mi = 761,606” in the Wikipedia article on Mexico to point to this Wikidata entry. The article also includes “|percent_water = 2.5”. The article on the United States includes “|area_sq_mi=3,796,742 |percent_water=6.97 |area_label2=Total land area |area_data2={{convert|3,531,905|sqmi|km2|abbr=on}}”.
What should be done to distinguish between area in land, water, and total and percent water? Do we use area (P2046) for all of these or create new data elements? And which should be primary and which computed? Or is it easy to create four data elements with two relationships, so a user can specify any two, and the other two are computed automatically?
Then I'd like to add entries reflecting changes in external boundaries going backwards. If I read correctly the Wikipedia article on w:Territorial evolution of Mexico, the most recent change in the external size of Mexico was its transfer of 49.2 acres to the US by the “Mexico–United States International Boundary and Water Commission” in 1976 due to changes in the Rio Grand / Rio Bravo that separates the two countries. I did not include these Boundary and Water Commission changes in the table I posted to en.wikipedia.org/wiki/Talk:Territorial_evolution_of_Mexico, but I did include Mexico's loss to France on January 28, 1931 of Clipperton Island with 6 km^2 = 2.3 mi^2.
My primary interest at the moment is producing a plot of the size of Mexico / New Spain back to 1783 or so. I've already done that and found a discrepancy of just over 7 percent. When I posted this question, I thought it would be a good idea to put what I had in Wikidata and invite others to fix any deficiencies there. However, I'm beginning to suspect that doing anything with Wikidata at this point may involve more work than I can afford right now.
However, if this can be set up correctly, others could later add shape files that could ultimately make it relatively easy for people to create maps showing the evolution of virtually any geographic entity, city or state, over time AND plot the size in km^2 or acres, hectare, etc.
Comments? Thanks again, DavidMCEddy (talk) 21:32, 13 November 2017 (UTC)
@DavidMCEddy:
  • Re distinguishing land/water area: The current way of doing that is for the area (P2046) statement to have the qualifier applies to part (P518) land (Q11081619) or applies to part (P518) water (Q283). (That's literally the item for water the chemical, which is clearly a bad way to do things, but with ~4500 uses, it looks like it's become a standard. Something we'll probably have to fix at some point...)
  • For the current area, the way to indicate that something is current is to not have a end time (P582) qualifier on that statement. Mexico (Q96) already has a statement for the current area.
  • For marking the events that resulted in the change of area, I think (?) you can use end cause (P1534) as a qualifier to point to the event.
--Yair rand (talk) 13:43, 14 November 2017 (UTC)
Thanks so much.
Sadly, the time I had to put this into Wikidata has expired, and I may not find time in the near future to do this. However, I will create my own private copy of your comments, so I can refer to them the next time I see something for which Wikidata might be useful. DavidMCEddy (talk) 20:52, 14 November 2017 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ What is the recommended procedure for a conflict between a Wikidata item and what seems to be the same quantity in a Wikipedia article?

For example, the area (P2046) of Mexico (Q96) is "1,972,550±1 square kilometre", but apparently the same item "Superficie" = 1 964 375 2​ km² in w:es:México. In w:Mexico, it matches Wikidata. (I will post a question on the talk page for w:es:México, as that is probably closest to the source of the discrepancy.)

Also, I see items "|area_km2 = 1,972,550" and "|area_sq_mi = 761,606" in w:Mexico. I assume these could be changed to refer to area (P2046) of Mexico (Q96), but I have so far not found the documentation that would tell me how to do that.

I found Wikidata:How to use data on Wikimedia projects. From this I concluded that {{#statements:Mexico|from=area}} might insert 1,972,550. I tried this in w:Mexico. It sort of worked, but had a curious side effect:
  • "|area_km2 = 1,972,550" displayed "1,972,550 km2 (761,610 sq mi)".
  • "|area_km2 = " displayed "761,606 sq mi (1,972,550 km2)": The metric units were demoted to a parenthetical display, while the English units were promoted to the main display.
Since most of the world's population uses the metric system, I decided NOT to implement this change and instead report the experience here.
Comments? Wikidata is quite valuable. It would be more valuable if it didn't demote the metric system like this. DavidMCEddy (talk) 21:33, 15 November 2017 (UTC)

The expert versus the amateur contributor ... Help Needed

It's a simple thing I seek. Throughout the world are experts in the most obscure of things and much of that knowledge is in their heads. When they die, the knowledge is lost.

I have a rare English car called a Bristol (411, for those who know about them). I belong to an email forum that uses Yahoo... long in the tooth, and not the best way to record knowledge. So I have proposed to the members, mostly men, many of whom are over 60, that we should set up and use a wiki. We can't use Wikipedia because original source documentation would be deleted instantly... not encyclopedic. So I started looking around. I came across wikicars.org, but while it has a log-in, it does not have a create account option, nor can I find an email contact. It looks like it has not been used since 2010.

I then found this site, wikidata.org, but for the amateur volunteer, it's a bit daunting. Wikipedia I understand, and I can show others how to make entries on it. Not so sure about Wikidata.

Here is what we want to do:

The knowledge base needs to be permanent, not something maintained by a car club that could cease if its membership dwindles.

We wish to divide the knowledge into models from 1946 to about 2010. The 400, 401, 402... to 411 and then some of the more recent cars. Then we need subcategories just like one finds in car repair manuals... engine, brakes, body restoration, fixing instruments, etc. We need the subcategories to be simple to access... follow some instructions and post your knowledge, your photographs and drawings. We then will ask our colleagues to fill the pages with their knowledge. It's not like fixing a new car where you order parts from the dealer, and the old ones come apart easily. You have to know different ways to do it, including what parts came from other British car companies, what can be adapted, or what needs to be done better. All that will go into the Wiki, if we can get it going.

I would appreciate advice on how to accomplish this.  – The preceding unsigned comment was added by Bristol7527 (talk • contribs) at 15. 11. 2017, 07:22 (UTC).

I don't see how Wikidata could help you. Maybe Wikia is something for you? --Anvilaquarius (talk) 07:43, 15 November 2017 (UTC)
  • Wikidata could be suitable to include details about car models and cross-reference any existing manual about them. In some obscure corner of WMF sites, there is a lot of info on tractor models.
    Information about individual cars might be harder to include, especially in the level of detail you are interested in.
    --- Jura 07:53, 15 November 2017 (UTC)

--- Any idea how to gain access to Wikicars.org?Bristol7527 (talk) 10:20, 15 November 2017 (UTC)

https://www.whois.com/whois/wikicars.org gives you the contact information for the owner of Wikicars.org but I agree that setting up a new Wiki with Wikia is likley the better solution for you. ChristianKl (✉) 12:49, 15 November 2017 (UTC)

Speaking of which, is there any alternative to wikia? As in I know there are tons of wikifarm but most of them are so small that contribution toward them would be flooded within mass of webpages when searching via search engines. Wikia itself is an option but the way it's gearing toward social network and "fandom", embracing nonstandard toolkits which would hesitate users when they try to create new projects on it. I have proposed an WMF-backed project for it on meta but I don't think WMF would actually accept that proposal. What other options would be available? C933103 (talk) 22:38, 15 November 2017 (UTC)

The proposal goes against Jimmy's financial interests so, there's pushback from that direction. On the other hand, if you really believe in the vision 2030 statement on bringing all kinds of knowledge under the Wikimedia umbrella it would make a lot of sense to also have a place within the project that can host English content that's not notable by enwiki standards.
If you want to invest a lot of efforts you could setup your own foundation and have as mission to hand over the project to the WMF as soon as the WMF would accept handover. If you get traction I don't think the WMF would oppose a handover. ChristianKl (✉) 00:24, 16 November 2017 (UTC)

The Wikidata template in the English Wikipedia is up for deletion again

w:Wikipedia:Templates_for_discussion/Log/2017_November_15#Template:Authority_control, decide for yourself what to vote. --RAN (talk) 15:09, 15 November 2017 (UTC)

In the meantime, can't this be defined directly in the interface?
--- Jura 19:45, 15 November 2017 (UTC)
I left my response (diff) after the proposer tried to come up with a lame excuse.
You know what the downside is of standing up to bullies? I'll probably get bullied a lot more in the future. Multichill (talk) 22:10, 15 November 2017 (UTC)
  • Looks like the discussion was closed before people could bring up the new Wikibase development that no longer requires the template being added to articles. Somehow Wikipedia seems to have lost the capacity to discuss ..
    --- Jura 02:16, 16 November 2017 (UTC)

Main page issue

Two quote from the main page (emboldening mine):

  • Welcome to Wikidata - the free knowledge base with 38,643,300 data items that anyone can edit.
  • 2017-11-12: The forty-three-millionth item, about an article in a scientific journal, is created.

These apparently-conflicting figures are likely to confuse people who do not understand the context - and the page's audience, new users, is likely to have such people in the majority. How can we improve things? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:40, 15 November 2017 (UTC)

IMO the creation of Hepatitis B or hepatitis C co-infection in individuals infected with human immunodeficiency virus and effect of anti-tuberculosis drugs on liver function. (Q43000000) is not worth a entry on the main page. I think real achievements like average of 8 statements per item or 600M edits should be added to that section. Also other news, i.e. WikidataCon, could be added instead of the boring Q-count. --Pasleim (talk) 17:57, 15 November 2017 (UTC)
  • I think we shouldn't report "forty-three-millionth item" in the news-section. Especially a metric that includes deleted items is not a metric that's worthy to be celebrated. Wikimedia projects are supposed to measure their success by the number of active editors (people who edit 5/100 times per month). When we reach new milestones about our editorship that's worth celebrating by putting it into the news section.
I don't think that the fact that WikiCite raised the average statements per items to 8 by using a lot of statements in items about scientific papers is something to celebrate either.
Metrics are a way to decide whether we make progress. When we ponder a policy change it's more important to think about whether it raises the number of active editors than whether it increases the number of items, the number of statements per item or the number of labels per item. ChristianKl (✉) 18:00, 15 November 2017 (UTC)
To all users, you are free and encouraged to write all the news that you think they could be interesting. Pamputt (talk) 19:43, 15 November 2017 (UTC)
Are imports of new datasets to Mix'n'Match and milestones in matching datasets appropriate for Main Page news? - PKM (talk) 20:53, 15 November 2017 (UTC)
I would say yes. If we have "too more" information at some points, we may filter some of them but for now, let's go. Pamputt (talk) 21:28, 15 November 2017 (UTC)

QuickStatements and 2017 Community Wishlist Survey

I went over the limit of 3 proposals per user in 2017 Community Wishlist Survey and have to drop my proposal to Expand QuickStatements to allow wider variety of statements. If someone feels like this proposal is a good idea they can adopt it by adding their name as the Proposer and adding the page to meta:2017 Community Wishlist Survey/Wikidata. --Jarekt (talk) 20:02, 15 November 2017 (UTC)

Is the use of diplomatic relation (P530) restricted to country with formal diplomatic ties?

  • The label for the wikidata property in most western language only say "diplomatic relationship" without referring to it as formal or not.
  • However the CJK label for the property refer to "Countries with diplomatic ties" signifying it should only be applied on countries that have formal diplomatic relationships.
  • In the original application for the property, it was not made clear, although the proposed qualifier for start and end date indicate it might mean formal diplomatic relationship.
  • In actual usage, the situation is mixed. For instance, in Taiwan (Q865), you see its diplomatic relation (P530) with countries like United Kingdom (Q145) and Panama (Q804) have been given an end date, there are also the like of United States of America (Q30), Australia (Q408) and Ukraine (Q212) that are not given an end date, especially the relationship between Taiwan (Q865) and Ukraine (Q212) were never formal but they're still linked there (with main article).
  • Could its meaning be standarized? C933103 (talk) 22:29, 15 November 2017 (UTC)
  • It seems to me that the pattern in which this property is supposed to be used is along with statement is subject of (P805) to point to an item about the actual relationship. Given that there might be Wikipedia articles about diplomatic relationship that aren't formal I would suggest that this property includes nonformal ties and it might make sense to standardize it that way.
Additionally, I think it makes sense to raise such a topic on the talk page of the property and ping relevant Wikiprojects. Having discussions like this on the project chat means that interested parties might not know about it. ChristianKl (✉) 22:55, 15 November 2017 (UTC)
Copied to Property_talk:P530, not surewhat wikiproject could be pinged. C933103 (talk) 00:27, 16 November 2017 (UTC)

Wikipedia article with multiple meaning in same article.

For example, ja:専用軌道 is currently introducing 3 different things on same page. 1. tramways (as defined by Japanese law) dedicated to cargo transport, 2. tramways (as defined by Japanese law) that are laid with dedicated right of ways, 3. railways laid on dedicated land. How should they be linked in wikidata? The current linked article to ko wikipedia in the currently linked item describe the (3.) point. C933103 (talk) 00:26, 16 November 2017 (UTC)

Wikidata weekly summary #284

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • Continued working on persistent storage of edits on Lexeme pages
    • Fixed an encoding problem in the SVG download of query results in the query service (phabricator:T178564)
    • Fixed a problem with change dispatching changes to Wikipedia and co, that sometimes lead to single wikis falling way behind. (T179060)
    • Added full URIs for external identifiers to the RDF export
    • You can leave feedback on how improved fulltext search result page should look like

You can see all open tickets related to Wikidata here.



I just fixed the father's father issues, by removing both relationships. Syced (talk) 06:07, 2 November 2017 (UTC)
@Syced: Removing both seems to be bad as it also removes the correct one. It should be possible to determine the correct one by looking at the data of birth. ChristianKl (talk) 23:23, 8 November 2017 (UTC)
I tried to leave the correct information in simple cases, but in most cases the situation is just too complicated for a non-expert, for instance when reference information is no readily available or when many members of a family share the exact same name+surname. Let's think the other way: Would you add two statements while knowing that one of the two is wrong (but you don't know which one), or would you just refrain from adding them? Syced (talk) 06:59, 9 November 2017 (UTC)
Terribly unhelpful, especially as the query was sent to many people who could try to sort it out as well. The problem we had in the past with some of these properties is that when one typed "father" property:P40 came up first. The inverse might have been true for "child".
--- Jura 09:59, 9 November 2017 (UTC)
@Syced: I generally follow "When in doubt, do no harm", so when I don't know anything about whether a statement is true or false I neither add nor remove statements. ChristianKl (✉) 21:26, 15 November 2017 (UTC)
@ChristianKl:@Jura1: If a statement is probably false (50%), then I would rather not have it in Wikidata, but that's just me. As a maker of applications that use Wikidata, I hate it when the data is wrong (that happens often). I would rather have only 4 solid statement than 8 statements with a 50% probability that each is false. I highly doubt any specialist about them reads Wikidata Weekly (maybe I am just bad at finding reference information?), but anyway here are the persons whom relationship I modified, feel free to fix them and add references:
Thanks all for caring! Syced (talk) 08:15, 16 November 2017 (UTC)

Link to off site wiki?

As mentioned above, there is currently a wiki for Mongolian in Traditional Mongolian Script as uhaan wiki, a Vietnamese Han-nom Wiki at Hannom-rcv.org, a minnan hanji wiki at taigi-pahkho.wikia.com, and a Korean Hanja wiki at kore.wikia.com. Additionally there are also a large Korean general-purpose wiki at namu.wiki. And then there are also other wikis like many specific-subject wikia sites (like a proposal for a car wiki mentioned above), uncyclopedia (you know what it is), and baidu baike (largest chinese online encyclopedia on the internet although copyright and other aspects of the site are not that clean), and such. Is it possible to link all of them in the wikidata database? There are currently Encyclopædia Britannica Online ID (P1417) for Encyclopædia Britannica (Q455), can these be handled the same way?C933103 (talk) 01:05, 16 November 2017 (UTC)

Sounds like you are needing to review Wikidata:Wikidata for authority control and Wikidata:WikiProject Authority control if you are talking about linking item to item, rather than a link from the item to the official websites.  — billinghurst sDrewth 03:28, 16 November 2017 (UTC)
Not sure what do you mean and not sure if these are related to what I'm talking aboutC933103 (talk) 04:02, 16 November 2017 (UTC)
Hmm, then maybe you are looking for Help:Identifiers. It is unclear to me exactly what you are trying to achieve.  — billinghurst sDrewth 05:20, 16 November 2017 (UTC)
I mean, allow users to use wikidata to navigate toward items on other wikis in the same way that users can use wikidata to navigate toward items on wikipedia, encyclopedia britanhnica, and other WMF sites.C933103 (talk)06:50, 16 November 2017 (UTC)
@C933103: We already have some properties that link to outside wiki's; just make a property proposal for the ones you think will be useful and that will allow adding these links. Or you can always add links with the more generic property described at URL (P973). ArthurPSmith (talk) 15:14, 16 November 2017 (UTC)

One of seven equal parts of a whole

Seven Wonders of the Ancient World (Q489772) is heptad (Q29441572) (group of seven things). Great Pyramid of Giza (Q37200) is Q42879824 (one of seven equal parts of a whole)? Is there another way to indicate that "Great Pyramid of Giza" is one of seven equal parts of "Seven Wonders of the Ancient World" and such structuring of data is not necessary? --Fractaler (talk) 13:34, 12 November 2017 (UTC)

I oppose using instance of (P31) and subclass of (P279) for this use-case as those properties are about identity. Your concern is about meronym/holonym relationships so part of (P361) and has part(s) (P527) seem appropriate. If you want you can express with has part(s) of the class (P2670) and quantity (P1114) that there are exactly 7 wonders of the ancient world. ChristianKl (✉) 14:03, 12 November 2017 (UTC)
Ok, no problem. Great Pyramid of Giza (Q37200) is part (Q15989253)? --Fractaler (talk) 14:13, 12 November 2017 (UTC)
Not if you mean instance of (P31) or subclass of (P279) with is. If you want to speak about something being a part, use part of (P361) and has part(s) (P527) for it. ChristianKl (✉) 14:34, 12 November 2017 (UTC)
Sorry, I don't understand. "Great Pyramid of Giza" is a part of "Seven Wonders of the Ancient World" or is not? --Fractaler (talk) 15:14, 12 November 2017 (UTC)
The word "is" in the English language is very overloaded with many different meanings. You can't conclude from the fact that the English word "is" is being used that instance of (P31) or subclass of (P279) are appropriate. We don't add instance of (P31) "part" to everything that uses "part of". That property already communicates all the necessary information. ChristianKl (✉) 15:22, 12 November 2017 (UTC)
  • Part of what editing a Wiki together is about is to come to consensus and edit based on common understanding. It seems you interacted with the Russian Wikipedia community in a way that they banned you because you didn't integrate yourself into the local consensus. I would like to have you continue to be in our community and that would be more likely to happen if you would invest more energy into trying understand other people views and use that understanding to guide your editing. In addition to understanding fellow community members, reading about how other communities model a problem domain is also about understaind other people views. ChristianKl (✉) 17:36, 12 November 2017 (UTC)
  •   Comment Maybe a more general response: in one way or the other, it should be possible to store this information in Wikidata. Maybe if you present ways to display/extract it, it's more likely to gain some support. I find some of your things interesting, but they do need some thought to comprehend what is generally not needed for most subcategories ;) As I only looked at statements on random item, I can't really say if they are consistently applied and extractable across a series of items.
    --- Jura 18:44, 12 November 2017 (UTC)
Sorry, I don't understand. How by WD add info "A is B" to WD? --Fractaler (talk) 20:05, 12 November 2017 (UTC)
There no general way. The word "is" can imply many different properties depending on context.
The wording "A is part of B" implies "part of". "A is green" would imply using "has quality". "A is FMA1000" would imply using the external identifier for Foundational Model of Anatomy. ChristianKl (✉) 20:31, 12 November 2017 (UTC)
If 1) "A is part of B" implies "part of" (and only) then this violation of the rights of the item, non-compliance with neutrality.The right to exist also has another point of view: in "A is part of B" it is meant that "A is С" (where С is a part of a divisible object B). Then, if 2) "A is green" would imply using "has quality" (and only), this is also violation of the rights of the item, non-compliance with neutrality. Second point of view: "A is a carrier of green color".
Because the ideology of properties does not allow deduction and induction, it does not provide an opportunity to trace the transitivity (as I informed in one topic before), I think sooner or later such an instrument in Wikidata will die (become atavism, rudiment).
In Foundational Model of Anatomy everything is mixed in a heap (for example, objects of a group and a group of objects have one superset). But the benefit of it is: it gives the right to the existence of certain scientific terms --Fractaler (talk) 07:10, 13 November 2017 (UTC)
  • Items for scientific terms have a right to exist when they can be reliably sourced under (2) of our notability guidelines. You don't need to abuse the instance_of or subclass_of to justify their existence. When it comes to mathematical terms, MathWorld ID (P2812) is enough to justify the existence of an item. ChristianKl (✉) 12:14, 13 November 2017 (UTC)
We don't have a charter of rights for items. Neutrality in the sense of our project defines the term doesn't mean that you can express your own personal opinion. We are a secondary database. It means that you have to provide sources. If you want to use instance_of or subclass_of in a nonstandard way it's not enough that you want to translate "A is green" to "A is subclasses of green object" you would actually need to bring a source where someone who models the domain says "A is subclass of green object". ChristianKl (✉) 11:49, 13 November 2017 (UTC)
How a deleted item can has some right to protection (to to exist)?
"A is part of B" implies "part of". "A is green" would imply using "has quality" - this is just one point of view, one of the ways to model the world. Do you need references to the fact that if "A is part of B", then "A is part"? Or if "A is green", then "A is a carrier of green color". I understood correctly? And if I point out these sources, then the items and edits that I have created that satisfy these sources will be returned, and the administrator, the behavior of which, I think, does not correspond to the behavior of the administrator, will be responsible for lack of a team member feedback? --Fractaler (talk) 13:33, 13 November 2017 (UTC)
A reference that says "A is part" doesn't say anything about A being an instance or a subclass of part. You would actually need one that does. ChristianKl (✉) 15:51, 13 November 2017 (UTC)
For example, Great Pyramid of Giza" (A) is a part (of B, of Seven Wonders of the Ancient World), here unclear what is A (being an instance or a subclass of part)? --Fractaler (talk) 18:45, 13 November 2017 (UTC)
That sentence doesn't say anything about something being a instance or a subclass. Just because the word "is" appears doesn't mean that there's an implication that something is either. ChristianKl (✉) 21:12, 13 November 2017 (UTC)
Seven Wonders of the Ancient World (Q489772) does not allow us to determine what Great Pyramid of Giza (Q37200) is (a instance or a subclass)? --Fractaler (talk) 09:17, 14 November 2017 (UTC)
No, in this case Great Pyramid of Giza (Q37200) is a meronym of Seven Wonders of the Ancient World (Q489772). This is modeled via the "has part/part of" relationship. Your idea that "is" mean either instance or subclass is wrong. You could read literature about applied ontology if you want to understand this better. Apart from that Seven Wonders of the Ancient World (Q489772) does not hav to contain any information about the identity of another item but only about how it relates to other items.
Wikidata is about structured data. Structured data means modeling relationship between different item with well defined relationshis and not with common language. As a result there no 1-on-1 correspondence between a word like "is" and a property. You always have to look at the relationship you want to model and see whether we have a property that was created for the relationship. If we have an existing one you can use it. Otherwise you can write a property proposal. On the other hand using a property in a way that it wasn't designed to be used because the you like the way it sounds in natural language is wrong. ChristianKl (✉) 12:07, 14 November 2017 (UTC)
relation (Q930933)? So far, only subset (Q177646)<->set (Q36161)<->superset (Q15882515) provides transitivity. What I am now talking about arose not just after reading the literature on ontology, but after its practical use and obtaining negative results on transitivity (which I already talked about). The lack of transitivity in your method is a defect that it can not eliminate, which will lead to its extinction. Nothing, better than set theory (Q12482) has not yet come up.
structured data (Q26813700) - description: empty.
Seven Wonders of the Ancient World (Q489772) is heptad (Q29441572) (group of seven things) or isn't? --Fractaler (talk) 13:20, 14 November 2017 (UTC)
Set theory doesn't tell you anything about the property of items or relationships such as meronymity. "part of" does happen to be transitive. Set theory has nothing directly to say about concepts of identity such as 'instance' and 'subclass'. I see no problem with classifying Seven Wonders of the Ancient World (Q489772) as heptad (Q29441572). ChristianKl (✉) 14:21, 14 November 2017 (UTC)
Set theory defines a rule. The properties of the object are determined, for example, in an empirical way.
Ok, if a set Seven Wonders of the Ancient World (Q489772) is the a subset of the set heptad (Q29441572) ("group of seven things"), then Seven Wonders of the Ancient World (Q489772) also consists of a seven things? --Fractaler (talk) 14:44, 14 November 2017 (UTC)
The phrase "of a seven things" is not in line with the grammar of English, so let's pretend you mean "of seven things". {1, 2} is a subset of {1, 2, 3} according to the classic meaning of the term in set theory, so the fact that the {1, 2, 3} contains three elements doesn't mean that {1, 2} also contains 3 elements. ChristianKl (✉) 17:04, 14 November 2017 (UTC)
a: yes, of course, thanks (in general, I belong to set (Q36161) "object that knows English", but, unfortunately, so far in its subset (Q177646) "object that knows English very badly").
The main reason for disputes is the lack of synchronization of terms. Set theory begins with the definition (extensional definition (Q5421961), intensional definition (Q1026899)) of a set. Will not it make it difficult for you to define {1, 2, 3} and {1, 2}? --Fractaler (talk) 06:55, 15 November 2017 (UTC)
This isn't just about standard English. Even the average English speaker likely doesn't understand the difference between a subclass and a subset. This is about concepts of ontology. If you don't understand the difference between terms like subclass and subset, why argue about them instead of accepting the current practice of how relevant properties are used in Wikidata? ChristianKl (✉) 13:17, 15 November 2017 (UTC)
I in such cases do not argue, I just ask the interlocutor to show a difference between terms on an example of items of a Wikidata. Ie, where there will be a subset, but there will not be a subclass. And vice versa. For some reason, no one has dared to show the difference. Can you succeed? By the way, I remind you that at one time to believe that the Sun revolves around the Earth was a common thing. --Fractaler (talk) 14:11, 15 November 2017 (UTC)
I already gave you above an example. {1, 2} is a subset of {1, 2, 3} but no subclass. In Wikidata we could say '{1, 2, 3} "has part" 1', '{1, 2, 3} "has part" 2' and '{1, 2, 3} "has part" 3' to describe the contents of it.
If you think you have discovered a new theory about ontology go and try to publish an academic paper about it. If you categorily disagree with the ontology community, that would be the way to go about it. The problem is that you lack the relevant knowledge and it's not my role to write a book to teach you. You could go and do actual research by reading a few books about applied ontology and then come back. ChristianKl (✉) 15:03, 15 November 2017 (UTC)
Why you don't want to define {1, 2, 3} and {1, 2}? {1, 2, 3} and {1, 2} are items of Wikidata? And why is there no pluralism of opinions, there is no neutrality, only one faith is imposed, and all the others are deleted without discussion? Is it normal? Fractaler (talk) 19:00, 15 November 2017 (UTC)
To the extend that Wikidata is neutral, that doesn't mean that Wikidata is in favor of original research. If you have a revolutionary theory that's like the earth resolving around the sun, Wikidata isn't the place to publish it but there are other venues.
If you have too much pluralism and different people use a property or item to mean to many different things than the data becomes useless for structured data, because the person who uses the data has no idea what the person who entered the data wanted to express. Coming to consensus is essential to doing structured data well. ChristianKl (✉) 21:39, 15 November 2017 (UTC)
Why do not you just 1) give a definition (for {1, 2, 3} and {1, 2}), 2) show difference between terms subsets/subclass on an example of items of a Wikidata? Are you afraid to lose? --Fractaler (talk) 07:17, 16 November 2017 (UTC)
I doubt meronymy (Q837495) applies here, ChristianKl. --Succu (talk) 22:54, 15 November 2017 (UTC)
Why? As Wikipedia writes in it's article on meronymy (Q837495): 'A meronym refers to a part of a whole. A word denoting a subset of what another word denotes is a hyponym. In knowledge representation languages, meronymy is often expressed as "part-of"'. ChristianKl (✉) 23:33, 15 November 2017 (UTC)
Can you define those terms that you use, please: 1) "part"; 2) "whole"; 3) "part of a whole"? --Fractaler (talk) 07:17, 16 November 2017 (UTC)
I use them as they are used in the literature. If you want something to read "Applied Ontology: An Introduction" by Katherine Munn and Barry Smith is good. ChristianKl (✉) 12:15, 16 November 2017 (UTC)
I believe you and have no doubt that you have 100_500 references to literature. Just to make a substantive conversation, scientists first agree, they give a definition of the terms that are used. I'm just asking for a definition of the terms that you use. Can you do it and not leave the answer? --Fractaler (talk) 12:39, 16 November 2017 (UTC)
Scientists frequently point to the literature for the terms they use and add a lot of citations to their work so that people who want to understand things better can see where concepts come from. I had a long discussion with you and invested plenty of effort in explaining myself. A lot more than the other admin that deleted your items and then spoke about banning you in case you create more of them. I don't think it's my role to explain how applied ontology works to you. If you want to convince me who actually talks to you instead of just deleting your items, then you would first need to familiarize yourself with the topic. ChristianKl (✉) 18:50, 16 November 2017 (UTC)
How scientists differ from representatives of religion or pseudoscience: when the representatives of religion or pseudoscience run out of arguments, they begin to send to read literature, refer to authorities, society, etc. But you certainly are not one of those? Fractaler (talk) 20:14, 16 November 2017 (UTC)
If you discuss the nature of energy with a physicist and have no clue of modern physics they will tell you to read a textbook. I'm doing the same thing. It not worth the time of physicist to explain physics to every person who thinks they invented a new theory of everything. Keep in mind as well that Wikidata is not about doing original research and showing with arguments that your original research is right. ChristianKl (✉) 21:33, 16 November 2017 (UTC)
I have already written that questions on the terms you use have arisen not from amateurism, as you are trying to present here, but after studying the special literature and applying the theory in practice. The scientist, if he does not know the answer, says so: "I do not know the answer", but do not lead the conversation aside. When I discuss the problem with scientist, we first try to synchronize (use the common terminology space) our terminology (otherwise there will simply be a loss of time, how it happens now). Such a procedure does not work with representatives of religion or pseudoscience (they do not even want to think about the alternative, and find all sorts of excuses for this). Moreover, in the case of representatives of religion ends as always - they do not even allow thought about alternatives (otherwise it will no longer be a religion). And what aggressive representatives of religion do with other points of view? Yes, for today there are a lot of examples of consequences. On a personal example with ru-Wikipedia version 2012.01 I will say that there was completely no tolerance, neutrality, and the alternative point of view was being killed was baned. --Fractaler (talk) 06:52, 17 November 2017 (UTC)

Odd system messages

I use the British English locale and have recently been getting a few system messages in Italian instead of in English, despite the rest of the interface being in English. For example, when entering data for title (P1476) I get "Linguaggio (obbligatorio)" instead of "Language (mandatory)" (which appears in the en and en-ca locales). Is there a logical reason for this? Jc86035 (talk) 15:39, 16 November 2017 (UTC)

Someone had added a wrong translation on translatewiki.net, now fixed. It will be updated on Wikidata after few days. Stryn (talk) 19:13, 16 November 2017 (UTC)

Importing thesis data from the University of Edinburgh Research Archive

Hi, at the University of Edinburgh we have been discussing importing the thesis data from the 27,000 records in the Edinburgh Research Archive (ERA) into Wikidata following on from the good work that Martin Poulter did with importing the Oxford Research Archive thesis data. e.g. South Asian Muslim politics, 1937-1958. So far we have decided on the statements we would like to create for the born digital records and for the pre-2000 records which can be found in this Google doc. The statements in red are the 'would like to include' but may not be initially possible. If there are any comments on the model proposed then do let me know. The main questions would be:

  1. how to model the wide variety of thesis types on Wikidata. I count 24 types in the ERA collection and some do not have an item on Wikidata yet while Master of Letters and Master of Literature seem to be treated as 1 item rather two distinct items. We proposed a statement of P31 instance of thesis (Q1266946) with a qualifier added to this statement of instance of (P31) Master of Science (Q950900) depending on which of the 24 thesis types is needed. Open to other suggestions though.
  2. Do we need an identifier property created for the Edinburgh Research Archive if we are using (P4101) Dissertation submitted to – University of Edinburgh (Q160302) and reference URL (S854) – “ERA link” as well as Described at URL (P973) – abstract link / ERA link.
  3. If we use (P1922) – First line Should this be the first line of the thesis or the thesis abstract?

Any comments welcome. I can knock up an example data item for a born digital record and an example item for the pre-2000 records if this would be helpful? Cheers, Stinglehammer (talk) 16:00, 16 November 2017 (UTC)

@Nomen ad hoc, Pigsonthewing:, as the proposers of Ecole des chartes thesis abstract ID (P4465) and Wikidata:Property proposal/EThOS thesis ID. Mahir256 (talk) 18:59, 16 November 2017 (UTC)

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

Sounds good!

  1. I would create separate items "Master of Science thesis", "Master of Letters or Litterature thesis", etc.
  2. For 27K records, I would say so.
  3. Use both, qualified with "applies to part"

Please include EThOS IDs, where you can, and once the property is created. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:29, 16 November 2017 (UTC)

I left some notes on the google doc. Two things I'd perhaps disagree with Andy on - firstly I think a generic "masters thesis" and then some kind of new property to indicate the degree it was submitted for would be a better way to go; and secondly a ERA identifier property seems superfluous if we have the URL plus also an ETHOS ID. Better to use the aggregator than the local value - simplifies querying once you're working across institutions. Andrew Gray (talk) 21:15, 16 November 2017 (UTC)
Thanks Andrew, these had mostly been my thoughts. I agree that the ERA identifier is superfluous when we are already planning to provide an aggregated ID and a link back to the source record on ERA. As for the hierarchy of the different thesis details, I must profess ignorance to best practice for these sorts of things on Wikidata so grateful to be getting some feedback ChaoticReality (talk) 11:51, 17 November 2017 (UTC)
Thanks Andy, could you explain the thinking behind the seperate items for each type of thesis? This is my first time working with Wikidata (although I was on the ORA team for a couple of years pre-Martin) and so I am not so knowledgable about how categorisation works. My instinct tells me that asserting the item is a thesis first, (which, apart from the University, is the only data in common across these records) and then the more specific type gives them the broadest possible scope to be useful to wikidata. Since most of the "sub"-types already exist as WD objects, it would also mean only needing to create a couple of new objects, and reusing as many as possible of the objects already in use. Cheers ChaoticReality (talk) 11:51, 17 November 2017 (UTC)
With the exception of humans, we class items specifically ("instance of non-fiction book" or "...of novel" being prefereed to "...of book", for example), and make the classes subclasses of other classes. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:50, 17 November 2017 (UTC)
Ok, that makes sense. So we would create a data item for each thesis type that doesn't already exist and use P279 to link them to the thesis object? Or P31? ChaoticReality (talk) 12:56, 17 November 2017 (UTC)
Should we also update existing items that don't yet have links? For example Doctor of Philosophy (Q752297) is a subclass of Doctoral Degree (Q849697) but Doctor of Education (Q837184) is not. So should I include statements to make that connection? Thanks for all the assistance, ChaoticReality (talk) 13:13, 17 November 2017 (UTC)
If the author has an existing Wikidata entry, and I can definitively match it to our details, then yes. Otherwise we haven't made a final decision yet as to whether to also create records for each author as opposed to just a text property of the thesis object. ChaoticReality (talk) 14:18, 17 November 2017 (UTC)
If you can't match it with existing items, I'd suggest to create it with educated at (P69), academic degree (P512) and floruit (P1317); possibly also occupation (P106). This would be more than we have on many people.
--- Jura 14:22, 17 November 2017 (UTC)

position held (P39)

Hello all, I have a question about the qualifiers of position held (P39).

My problem is: Getúlio Vargas (Q156844), for example, was one of the Presidents of Brazil, during two periods (11-03-1930 to 10-29-1945 and 01-31-1951 to 08-24-1954). In his item, the position held (P39) property has two President of Brazil (Q5176750) values, both with start time (P580) and end time (P582). Is that correct? Another example is Geraldo Alckmin (Q964493), who is in his 4th mandate as a State Governor of Brazil in São Paulo (Q174) (from 2001 to 2003, re-elected in 2003 until resigning in 2006, elected in 2011 until 2015 and re-elected in 2015 until now (so no end time (P582) for that mandate)). What should one do? Put 4 entries of State Governor of Brazil? Propose the creation of a mandate property? Please, help. Ederporto (talk) 16:44, 16 November 2017 (UTC)

@Ederporto: You have exactly the right idea; add the position multiple times for each contiguous interval during which the person held the position. Getúlio Vargas (Q156844) is set up properly, and Geraldo Alckmin (Q964493) may be set up similarly as well. (Note that with respect to contiguous intervals, there should be two position held (P39) statements--one for his tenure from 2001 to 2006 and one from 2011 to the present.) Mahir256 (talk) 17:13, 16 November 2017 (UTC)
@Mahir256: Thanks for the quick asnwer! I'm worried about the result of having two equal values for P39, for example, in a infobox that takes P39 of Getúlio Vargas (Q156844). Will appear President of Brazil (Q5176750) twice? Does not seem to be right, and may be extremely laborious create an if command in the template code for every possible value of P39. Ederporto (talk) 18:10, 16 November 2017 (UTC)
@Ederporto: It depends how it's handled for Brazilian politicians. In general, the approach is do a single statement for continuous mandates.
--- Jura 19:14, 16 November 2017 (UTC)
@Jura1: Can you provide an actual reference for that? Previous discussions, in which you were invloved, led to the contrary (examples: 1, 2). The general consensus is to do a statement by mandate, even if they are contiguous, for the reasons explained in these discussions and below by Oravrattas. — Envlh (talk) 00:10, 17 November 2017 (UTC)
I think you are confusing France and the UK with other countries. UK MP have a very special model.
--- Jura 07:24, 17 November 2017 (UTC)
You are not answering to my question and you are not providing any reference to the point of view you are pushing (and by the way, the first link is mainly about Estonia, not UK). — Envlh (talk) 10:08, 17 November 2017 (UTC)
I would say that one position held (P39) is fine on continuing mandates where there would be no difference other than contiguous end time (P582)/start time (P580) dates. However, if we want to record different information about each mandate (e.g. a distinct elected in (P2715), or different significant event (P793) qualifiers pointing at separate inauguration items, or the like, then splitting into separate P39s would be appropriate. Some countries also use a new series ordinal (P1545) after a re-election, in which case we'd usually also want to split into separate statements. --Oravrattas (talk) 20:44, 16 November 2017 (UTC)

Parameter: Medical specialty (P1995)

(moved from Wikidata:Contact the development team There is a problem/confusion with this parameter in mental disorders. In this way, the scheme seems to indicate that the disorder is only of a medical nature, when in reality its approach is often psychological/psychotherapeutic (not only psychiatric). For example, in this disorder Q1060310, at the parameter medical specility there is: psychiatry, but psychologists/psychotherapists often have to deal with it (or Q178190). Likewise, with most personality disorders (Q270673), except for cases of severe episodes of psychosis, because the treatment will have to be primarily psychiatric-pharmacological. The solution would be: 1) to add "psychology" or "psychotherapy" in the same parameter medical specialty; 2) to create a new parameter: "other specialty", where to add "psychology or "psychotherapy". How to fix this problem? --Dapifer (talk) 15:48, 16 November 2017 (UTC)

Moved here from Wikidata:Contact the development team --ValterVB (talk) 20:20, 16 November 2017 (UTC)
I think that you can add in property health specialty (P1995) value psychoanalysis (Q41630) or psychology (Q9418) (I don't know what is correct). --ValterVB (talk) 20:25, 16 November 2017 (UTC)
Wikidata is still far from being complete. There's no problem with adding multiple values. In the best case I would recommend to add the new value with a good reference of an academic paper that talks about how psychologists treat a given problem. ChristianKl (✉) 20:30, 16 November 2017 (UTC)
Thanks for the answers. I would like to add another parameter, called "other spacilty" under "medical specialty". But I do not know how to do it.
@ValterVB:, scusami per aver sbagliato "posto", non sapevo dell'esistenza di questa pagina di servizio. Comunque, sì, se non riuscissi ad inserire il nuovo parametro, sarebbe il caso di aggiungere "psicoterapia" e/o "psicologia" al parametro "specializzazione medica". Grazie ancora. --Dapifer (talk) 10:58, 17 November 2017 (UTC)

P1532

country for sport (P1532): country a person or a team represents when playing a sport

Spain national association football team (Q42267) --> country for sport (P1532) --> Spain (Q29)

but

Andalusia autonomous football team (Q1125275) --> country for sport (P1532) --> Andalusia (Q5783)

Andalusia (Q5783) is not a country. Should we change the label?  – The preceding unsigned comment was added by Xaris333 (talk • contribs) at 13. 11. 2017, 12:36‎ (UTC).

country for sport (P1532) mean "countries/region" in my language. C933103 (talk) 22:41, 15 November 2017 (UTC)
@C933103: But the English description is "country for sport". Should we change it? Xaris333 (talk) 06:28, 18 November 2017 (UTC)

Modifier

What's the best way to create a WikiData item for modifying occupations? I'd like to distinguish assistant attorney general from attorney general, for example. Madcoverboy (talk) 22:26, 17 November 2017 (UTC)

Just create it and make it a subclass of attorney general and instance of "position". Almost all of the cabinet level positions need to have their assistant level positions created. --RAN (talk) 05:59, 18 November 2017 (UTC)
We have United States Assistant Secretary of the Treasury (Q7889546) and United States Secretary of the Treasury (Q4215834). Does anyone know of a better example showing the hierarchical structure of a cabinet level position with at least 4 layers with second assistants and third assistants? --RAN (talk) 06:01, 18 November 2017 (UTC)
One would think that they have their deputy level, and their assistant level, though those would be dynamic in time. An example.  — billinghurst sDrewth 06:35, 18 November 2017 (UTC)

date of baptism in early childhood (Property:P1636)

We have date of baptism (P1636) but do we have "place of baptism in early childhood"? RAN (talk) 00:03, 18 November 2017 (UTC)

No, but location (P276) qualifier is usually applied. (propolal failed) - Kareyac (talk) 06:18, 18 November 2017 (UTC)
Works fine as a qualifier. The place of baptism is interesting and useful, though usually not of consequential value.  — billinghurst sDrewth 06:30, 18 November 2017 (UTC)

Module

Anyone can help me create module(s) for Wikipedia, to take data from wikidata ? Xaris333 (talk) 06:10, 18 November 2017 (UTC)

Maybe existing (Module:Wikidata (Q12069631), Module:Wikidata2 (Q25936424)) are acceptable, and look at Wikidata:How to use data on Wikimedia projects. - Kareyac (talk) 06:38, 18 November 2017 (UTC)
I know that pages. I need specific module. Xaris333 (talk) 06:44, 18 November 2017 (UTC)
For what? Maybe already exist. --ValterVB (talk) 08:02, 18 November 2017 (UTC)
@ValterVB: I want a module to show the most recent value (according to point in time (P585)) of ranking (P1352) with the date in brackets. For example, Argentina national association football team (Q79800) --> ranking (P1352). It must show "4 (16 October 2017)". (I don't want to use preferred option). Consider that determination method (P459) must always be FIFA World Rankings (Q180825). Xaris333 (talk) 08:10, 18 November 2017 (UTC)
Italian template it:Template:Wikidata based on Italian module it:Modulo:Wikidata can do it, you can see the result in my sandbox on Italian wiki: it:Utente:ValterVB/Sandbox/Temp. The first row show all the results, the second shou only the more recent. --ValterVB (talk) 09:42, 18 November 2017 (UTC)
@ValterVB: Two problems. The main one is that your example is not considering that determination method (P459) must always be FIFA World Rankings (Q180825). Check your sandbox. Xaris333 (talk) 11:13, 18 November 2017 (UTC)
Yes but I used only standard template/module, I think that is easy modify the module for add a filter. --ValterVB (talk) 11:25, 18 November 2017 (UTC)
@ValterVB: Yes. But I need a new module to do what I want. Can you help with this? Xaris333 (talk) 11:32, 18 November 2017 (UTC)
Not so expert but if I found something I ping you. --ValterVB (talk) 11:51, 18 November 2017 (UTC)

Add 90.000+ IMDb items

I have 90.000+ IMDb ids with labels which are not used in WD for now. Could anybody provide them with information from IMDb like instance of (P31), publication date (P577) or possible descriptions, so that we can add valid new items for all IDs? I saw that some of you did already scraping data from IMDb.   WikiProject Movies has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. Queryzo (talk) 12:13, 18 November 2017 (UTC)

QuickStatements

Hello. I am trying to add previous FIFA World Ranking with QuickStatements. I have already add the latest ranking. I have a serious problem. If a team have not change ranking, the tool don't add the second value as a new value. It just adding the new source and points to the previous value. This is a problem because the date of publication of the ranking is different and the points are different. Please see https://www.wikidata.org/w/index.php?title=Q79800&oldid=595437481 (ranking (P1352). That the way is must be it the item [ https://www.wikidata.org/w/index.php?title=Q79800&oldid=595441497].

Is there a way to solve that? Is there other tools to use? I have the data. I am ready to add it. I need a tool :) Xaris333 (talk) 16:19, 18 November 2017 (UTC)

Drag'n'drop gadget

I find this gadget rather useful, but what I don't like about it (or maybe I don't know how to use it properly) is that when dragging statements from, let's say, one of Wikipedia's langauge versions, you still need to add the reference (imported from) manually. Is there any better way to do it? Powerek38 (talk) 16:31, 18 November 2017 (UTC)

@Powerek38: It bugs me too, I will try to rewrite this plugin in the next few days. Yarl 💭  00:13, 20 November 2017 (UTC)

If you have time today…

… you may want to improve the brand new Commonwealth Games Federation athlete ID (archived) (P4548) by matching IDs on Mix'n'match. There are thousands of entries so any help would be appreciated. Thierry Caro (talk) 07:35, 19 November 2017 (UTC)

  Notified participants of WikiProject Tennis: You may also want to help with Tennis Temple player ID (P4544) through this. Again, there are thousands of IDs waiting to be matched to our items! Thierry Caro (talk) 01:54, 20 November 2017 (UTC)

Deletion of article in tg-wiki does not always remove it from wikidata item

Article tg:Энрике Пакеко Лима was deleted on Oct 23, but was not removed from Henrique Pacheco Lima (Q197584). Wasn't this supposed to be done automatically? --Ghuron (talk) 08:03, 19 November 2017 (UTC)

Indeed, it should have happened automatically. Unless someone comes up here with a good idea, we can ask the developers what happened.--Ymblanter (talk) 09:12, 19 November 2017 (UTC)
26 October, it was mass-deletion of stubs about football players. Maybe the problem is that the system can't process such massive deletions?--Soul Train (talk) 20:58, 19 November 2017 (UTC)

Infobox medical condition: medical specialty

I talked about this problem on it.wiki, with a Wikidata administrator and in this chat. But I did not manage to solve the problem (surely, because of my poor English and my little experience on Wikidata). The problem concerns the infobox of mental illnesses and medical specialty: psychiatry (P1995).

All over the Wikipedia, in the infobox about mental disorders, it appears as a specialty, exclusively psychiatry. This, in addition to confusing the reader, can be a problem. A person suffering from one of these mental disorders, looking at the infobox, believes that the only way to cure his disorder is to seek psychiatric care.

For example, a person suffering from this personality disorder → Q509122 (avoidant personality disorder) might believe that his disorder should be treated according to the psychiatric method. When, his true specialty is "psychotherapy" or "psychology" (except for rare cases, when the patient has a serious reactive psychosis). And this is true for many mental disorders. Often, psychiatric treatment, for some disorders is not recommended.

Psychotherapy remains the strongest specialty in this field, often alone or accompanied by psychiatric/pharmacological treatments. But psychotherapy (often conducted by psychologists) never appears, as does psychology.

In the past discussion, it was decided to add "psychology" or "psychotherapy" to the parameter "medical specialty". But that did not work, as you can see here and here. And of course I understand, are not medical specialties. But instead, in this personality disorder → Q178190 (obsessive-compulsive disorder), pscicology , has been inserted correctly.

Now, and this is the problem, where or how can we add these specializations (psychotherapy or psychology)? It's possible create a new property? For example, other specialty: psychotherapy or psychology, replacing the omnipresent psychiatry specialization? --Dapifer (talk) 22:48, 19 November 2017 (UTC)

The problem seems to be that now psychotherapy (Q183257) is homonym (Q160843): 1) specialty (Q1047113); 2) medical treatment (Q179661). --Fractaler (talk) 08:43, 20 November 2017 (UTC)
Given the constraints that are set psychology (Q9418) doesn't violate them but psychiatry (Q7867) does violate the constraint. When there's conflict it's worthwhile to add references to your claims but I see no reason why the non-constraint violating psychology (Q9418) should be removed if it would be properly referenced. ChristianKl (✉) 09:58, 20 November 2017 (UTC)
I still do not understand how we can solve the problem. Although I should look for a source, every time for a single mental disorder (and it would be a long work), there will always be someone who will say, "psychology/psychotherapy is not a medical specialty." Of course this can not be a solution, as we have seen. Would not it be the case to add another property or find another less ambiguous solution? --Dapifer (talk) 10:36, 20 November 2017 (UTC)
If different people consider different things to be true and both people have references for their view Wikidata is capable of displaying both positions and use statement disputed by (P1310). It's worth noting that we do have possible treatment (P924). If you still think we need a new property, feel free to write a property proposal under https://www.wikidata.org/wiki/Wikidata:Property_proposal ChristianKl (✉) 11:52, 20 November 2017 (UTC)
The problem of homonymy (Q21701659) has only one solution - by clarification (ie, the introduction of terms that will have only 1 definition, meaning). Homonyms in Wikidata produce only chaos, as we see it now in the next concrete example. --Fractaler (talk) 12:01, 20 November 2017 (UTC)
I try, I think is the least controversial solution. I do not have much experience with Wikidata, so I do not know how the "elements" or "properties" work and I have never even created one. I just know this is a problem... although apparently small. Anyway, let's see what happens. --Dapifer (talk) 13:23, 20 November 2017 (UTC)
Done here, I hope it is clear. --Dapifer (talk) 14:33, 20 November 2017 (UTC)

A link has been changed for more than a month, but on en.wiki it still uses the old one.

  Resolved

Q16465: the link for zh.wiki has been changed to zh:Wikipedia:典范条目 since 1 October 2017‎. But on English Wikipedia w:Wikipedia:Featured_articles, it still uses the old zh:Wikipedia:特色条目.

What went wrong here, and how can I fix it? Thanks.--66.76.81.9 04:48, 20 November 2017 (UTC)

The problem was fixed. New name link is now shown for zh.wiki. --Was a bee (talk) 13:40, 20 November 2017 (UTC)
Thank you very much. -Fireattack (talk) 18:00, 20 November 2017 (UTC)

Where are wikidata to wikipedia links available?

Where can one find the mapping between Wikidata and Wikipedia. I can see the links online in both Wikidata and Wikipedia, but I cannot find the links in neither the RDF, JSON nor the XML-dumps.  – The preceding unsigned comment was added by 90.185.101.112 (talk • contribs) at 09:51, 20 November 2017‎ (UTC).

See mw:Wikibase/Indexing/RDF Dump Format#Sitelinks - the format is <wikipedia article> schema:about <wikidata item> in the RDF dump. ArthurPSmith (talk) 18:25, 20 November 2017 (UTC)
I can find this info on the Wikidata Query Service but not in the RDF dump file (ie. wikidata-20170530-truthy-BETA.nt)?

BREAKING CHANGE: wbcheckconstraints detail output

Hi all!

This is an announcement for a breaking change to the output of the WikibaseQualityConstraints constraint checking API, to go live on . It potentially affects clients that use the wbcheckconstraints API action. (We are not aware of any such clients apart from the checkConstraints gadget, which is not affected.)

Currently, the description of a constraint in the API response includes the detail and detailHTML fields, which contain the constraint parameters. The gadget has never used these fields, since the error messages for some time now contain all the information needed to understand the constraint violation (that is, the constraint parameters are part of the message where necessary). Additionally, since the move from the {{Constraint}} templates to constraint statements on properties (using property constraint (P2302)), parsing constraint parameters is no longer the complex task it once was, and consumers interested in the constraint parameters can inspect the constraint statements using the standard Wikibase APIs or the Wikidata Query Service.

Since these two fields can account for up to 40% of the wbcheckconstraints API response size, and we want to start caching those responses soon, we will remove the detail and detailHTML fields on . This is already in effect on the Wikidata constraints test system; you can test your tools or other code there.

Relevant tickets:

Relevant patches:

Please let us know if you have any comments or objections. --Lucas Werkmeister (WMDE) (talk) 11:25, 20 November 2017 (UTC)

Help change a property pleae

I have posted a question at Talk:Q6343552#Is this item correct for the he-wikipedia a few weeks go, but no one answered. I would like to go ahead and correct the label, but don't know how to. If I try to edit and save I get a message:Could not save due to an error. Item Q13244694 already has label "קטגוריה:מנכ"לים אמריקאים" associated with language code he, usi.

Can anyone help? Ottawahitech (talk) 14:18, 20 November 2017 (UTC) Please ping me

@Ottawahitech:
in fact, you could not change the link on Category:American chief executives (Q6343552), because the one you wanted to add in replacement is already used on Category:American business executives (Q13244694)
I can't read hebrew, and therefore cannot check which is which, but you've got to remove a link from the item it is linked to, to be able to add it on another item. :) --Hsarrazin (talk) 19:02, 20 November 2017 (UTC)

Website property detect language?

Is there a way to set an official website property which varies depending on the language of the wiki it's used on? For example, I'm trying to add the official website for Q650929 which is http://www.koga.com, however this URL redirects to the company's Dutch website at http://www.koga.com/nl (the English website is http://www.koga.com/en, and there are also French and German versions). If there's not a way to send a different URL to different wikis, is there a standard here for which to choose as the default? Ivanvector (talk) 15:42, 20 November 2017 (UTC)

What is wrong with http://www.koga.com? They might change the system one day to http://nl.koga.com for example. Use the most stable URL. Sjoerd de Bruin (talk) 18:57, 20 November 2017 (UTC)
Sure, if that's what needs to be done then so be it. If they change the websites then it's a matter of updating the data anyway, maybe they change their main URL to http://www.kogabicycles.com or something. No URL is permanent. Ivanvector (talk) 19:20, 20 November 2017 (UTC)
@Ivanvector: Qualify with language of work or name (P407). However, are you sure the redirect is not dependent on your browser's language setting, or your locale? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:16, 20 November 2017 (UTC)
I'm not sure that it's not dependent, but as I'm English and in Canada, I doubt that's the case. Thanks for the tips. Ivanvector (talk) 19:20, 20 November 2017 (UTC)

Discussion: We must delete genes that are deprecated by NCBI?

For those who are interested: discussion. --ValterVB (talk) 20:29, 20 November 2017 (UTC)

New print to pdf feature for mobile web readers

New print to pdf feature for mobile web readers

The Readers web team will be deploying a new feature this week to make it easier to download PDF versions of articles on the mobile website.

Providing better offline functionality was one of the highlighted areas from the research done by the New Readers team in Mexico, Nigeria, and India. The teams created a prototype for mobile PDFs which was evaluated by user research and community feedback. The prototype evaluation received positive feedback and results, so development continued.

For the initial deployment, the feature will be available to Google Chrome browsers on Android. Support for other mobile browsers to come in the future. For Chrome, the feature will use the native Android print functionality. Users can choose to download a webpage as a PDF. Mobile print styles will be used for these PDFs to ensure optimal readability for smaller screens.

The feature is available starting Wednesday, Nov 15. For more information, see the project page on MediaWiki.org.

Thank you!

CKoerner (WMF) (talk) 22:07, 20 November 2017 (UTC)

floruit (P1317)

If I have multiple dates, about which is known, that the person in question has been alive, which would be the best one to be used for floruit (P1317)? The most recent one, the oldest one, random? --Edgars2007 (talk) 05:15, 19 November 2017 (UTC)

Assuming that the references are of equal quality: If there is a date of birth or date of death, the one furthest away from that. Otherwise the first and the last. At least that gives an idea when the person lived.
--- Jura 05:31, 19 November 2017 (UTC)
@Edgars2007: Use instead work period (start) (P2031) and work period (end) (P2032). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:15, 20 November 2017 (UTC)

Red Wikipedia links and Wikidata

On Wikipedia we have red links to indicate an article doesn't exist. Often an article about the subject exists in another language. Currently we don't really have a consistent system to keep track of this. We do have Template:Red wikidata link (Q15977575) and Template:Interlanguage link (Q13606208) in some language Wikipedia's, but both seems to be a lot of manual article work that don't really seem to scale.

Would this be something we could store in the Wikidata items? So for example for Malang (Q11095) that the wanted article should land on nl:Malang (regentschap)? That would make this data available in a structured format. A gadget on the local Wikipedia could change red links to include links to other languages based on your babel option or even links to invoke the translation extension. It also becomes to include this data in queries.

We would need a way to store this. A new string property would make sense ("wanted article"?) with some qualifier to indicate the project (new or existing one?) like for example Dutch Wikipedia (Q10000).

Anyway, thinking out loud here. Is this a direction we should investigate? Input appreciated. Multichill (talk) 20:52, 20 November 2017 (UTC)

@Multichill: The Mandarin Wikipedia has an amazing system whereby many of the red links there have tooltips with links to a page in another language. (See the template in action here.) It is worth seeing if fallbacks involving sitelinks from Wikidata can be incorporated into this template or something similar. Such a development would be different from the existing {{Q}} and {{P}} templates on other wikis in that it links not to the Wikidata item itself, but to a sitelink on another wiki that is present within the Wikidata item. Mahir256 (talk) 06:55, 21 November 2017 (UTC)

Biodiversity Heritage Library DOI

Hoi, I have a list with publications from the BHL and it includes publications with a DOI, publications that are free to read on the internet. It does include the publication, the article name, the license.. Details like authors may follow. I want to upload this information to Wikidata. Based on the DOI and the title it will easy to add author information at a later date. There are more publications but they are less interesting as they are not available for reading. They can wait until the information s more complete. Thanks, GerardM (talk) 06:47, 21 November 2017 (UTC) NB they are 167439 publications.

I just learned the authors known to BHL are in a Mix N' Match, so they will all be connected .. :) Thanks, GerardM (talk) 12:59, 21 November 2017 (UTC)

Mountain passes, property

Hi. Since mountain passes are landform (Q271669)s (a relative minimum between two mountains?)... what property would be the most adequate in order to link a mountain pass with the road crossing it? strakhov (talk) 13:42, 21 November 2017 (UTC)

@Strakhov: carries (P2505) has been modified to take care of this. Thierry Caro (talk) 14:44, 21 November 2017 (UTC)
Thanks a lot! :) strakhov (talk) 14:49, 21 November 2017 (UTC)

How to link "Han-ji" articles on minnan wikipedia in wikidata?

On min nan wikipedia, a number of articles written in "han-ji" script have been created in "Talk" namespace, as oppose to latin scripts that are commonly found for articles on main namespace. Those articles in "Talk" namespace would be a duplicate of articles in main namespace but are written in different script. For example, see nan:Canada and nan:Talk:Canada. There seems to be a proposal to create a "hanji" namespace for articles written in hanji script on the wikipedia which have been passed but doesn't seems to be enacted yet. So:

  1. I have added nan:Talk:Canada to Canada (Q21287484). Is this an acceptable way to link in this way?
  2. Canada (Q21287484) is a permanent duplicated item (P2959) of Canada (Q16). On meta's community tech wishlist, I have expressed desire to enhance the system so that linked entry in Canada (Q21287484) can be added to Canada (Q16) directly without the need of creating duplicated item. However some users said it have been discussed before and the result is against such measure. Where was such discussion being made and why was that? C933103 (talk) 22:16, 15 November 2017 (UTC)
  • There's value in being able to directly navigate a person to a page in a wiki if a page exist. At the moment you have multiple pages it's not clear where the user is supposed to be sent. That means that every tool has to incorporate the case. In the case you are proposing Wikidata can't directly send the person who wants to see the hanji version to the hanji version and the person who wants to see the latin version to the latin version.
Additionally, it seems like a hack. There's no good reason why there shouldn't be a separate subdomain for a separate script so that Wikidata can link to the two different versions under different subdomains. It would be better if the mediawiki software would be configuered in a way that allows in this case a single Wikipedia community to have two subdomains that point to the relevant scripts.
I looked at IETF and currently it seems like there isn't even a IETF language tag for something like nan-han-ji. It might be worthwhile to register one. ChristianKl (✉) 23:38, 15 November 2017 (UTC)
@ChristianKl: What happened to nan-Hani, nan-Hans, or nan-Hant as appropriate? Also @C933103: and suffer from a similar problem. Mahir256 (talk) 00:44, 16 November 2017 (UTC)
It is indeed a hack. As for reason why it's done? There is a request for such a wikipedia on meta but Langcom still haven't verified the request yet (The request was 9 years old). Even if they process the request, they will likely turn it down too on the ground of the lack of ISO 639 code.
The script code for hanji should be either Hani, Hant or Hans. These combinations aren't registerede in IANA subtag registry yet. I suppose it's possible to ask wikimedia taiwan to reg relevant IETF combination but I don't think that'd help what we're talking about.
Other than nan/hak/cdo, hy/gom/mn are other wikipedias that have same situation although those articles are placed on main namespace instead. Additionally, vi and ko and mn have off-site wikis because of alternative script C933103 (talk) 00:48, 16 November 2017 (UTC)
@C933103: See also phab:T165882, they are now requesting to create a "Hàn-jī" namespace so all the talk pages will be actual talk pages. --Liuxinyu970226 (talk) 04:13, 16 November 2017 (UTC)
But that's still not done yet. And jusging byu the lack of recent activity on the ticket I don't think it's going to be done anytime soonC933103 (talk) 06:46, 16 November 2017 (UTC)
@C933103: Sorry but what means "jusging byu"? I can't find wikt:jusging on Wiktionary, and wikt:byu seems to be a Rōmaji? --Liuxinyu970226 (talk) 03:53, 17 November 2017 (UTC)
@Liuxinyu970226: sorry that is a typo of "judging by". C933103 (talk) 03:59, 17 November 2017 (UTC)
It seems like the status quo is that languages like Min Nan try to hack around decisions of the language committee. This leaves to the bad consequences that different languages use different solutions for solving the problem of having multiple scripts.
I don't think it's worthwhile for Wikidata to try to accommodate some hack. It would be better to formulate a language policy that actually works for the communities in question so that they don't have to hack around the language committee.
One way to do this would be to write a RfC on meta that declares change the policy of needing a ISO 639 code into needing a valid IETF language tag.
It might also be worthwhile to talk to the WMF who actually want to help small wikis. A policy that results in off-site Wikis getting created should be neither in the interest of the Wikimedia community nor the WMF. ChristianKl (✉) 18:02, 16 November 2017 (UTC)
There are actually talking about two situation here,
  1. First, the use of Talk namespace to host hanji article in Minnan wikipedia. Since it's probably transient so it probably don't need much attention.
  2. Second, after the phabricator ticket completed, then Minnan wiki will still have two separate article for each individual subject, and it will the same as at least half dozen other wikis being mentioned.
These would still need to be supported. And I wouldn't exactly call them "hack", as they have been used by various wikipedia including chinese wikipedia ever since their creation (for chines wikipedia, until the implementation of language converter), and probably predate any language policy. C933103 (talk) 05:07, 17 November 2017 (UTC)
@C933103: Does the current usage of the language converter in Chinese mean that have one article for one topic that can be displayed differently or do they still have in some cases multiple articles for the same topic? ChristianKl (✉) 16:48, 22 November 2017 (UTC)
@ChristianKl: On Chinese Wikipedia, after deploying language converter, Simplified and Traditional language version article have been merged and become single article for each concept. See also http://www.mediawiki.org/wiki/Writing_systems and meta:Wikipedias in multiple writing systems. C933103 (talk) 17:09, 22 November 2017 (UTC)
@C933103: Is there a good reason why the other languages with multiple scripts shouldn't be merged in a similar way? ChristianKl (✉) 17:11, 22 November 2017 (UTC)
@ChristianKl: The "Without Automatic Conversion System" section in meta:Wikipedias in multiple writing systems listed some reason why such systems are not employed. For instance, for vi/ko/nan/cdo/hak, Chinese character is one of the writing system being used to write the language alongside Latin alphabet(or Hangul for ko). Latin alphabet or Hangul reflect the reading of words, however Chinese characters reflect the etymological information of the word. Each single reading represented in each of these languages could correspond to multiple Chinese characters. Based on the performance of a Hangul-Hanja converter linked in the second proposal for Korean Hanja Wikipedia, it would probably mean that special markup is needed in wiki source code once every few lines in order to ensure the correct character will be displayed in the converted edition of the wikipedia. That would also require most editors to be fluent in both scripts in order to edit them. But usually it is not the case.C933103 (talk) 17:57, 22 November 2017 (UTC)

Rebranding of Wikidata:WikiProject Politicians by EveryPolitician

There is some discussion on Wikidata talk:WikiProject every politician with a member of EveryPolitician (see Wikidata:EveryPolitician) who wants to rename the project Wikidata:WikiProject Politicians and/or create a duplicate at Wikidata:Wikiproject every politician. It seems to me that it just matches more closely their website's name. The organization (and supposedly the contributor trying to rebrand this) is currently paid by WMF to do some import work of their data. Feel free to comment. --
--- Jura 21:15, 16 November 2017 (UTC)

I edited the above comment as the a page was redirected while the discussion is still ongoing. https://www.wikidata.org/w/index.php?title=Wikidata:EveryPolitician&oldid=565751502 is probably the better link for "Wikidata:EveryPolitician" mentioned above.
--- Jura 18:18, 22 November 2017 (UTC)
There is a subtle but important distinction between "paid by WMF to do some import work of their data" and "in receipt of a WMF project grant to support their work"; the latter being factual. Furthermore, it appears to be Jura who wants to rename the project. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:34, 16 November 2017 (UTC)
I've left a more detailed note at the EveryPolitician talkpage, but Wikidata:WikiProject Politicians doesn't really seem to exist - it's a dead page with no activity, few useful links, etc. I'm not sure what merging the project with activity into the one without any activity would achieve. Andrew Gray (talk) 21:54, 16 November 2017 (UTC)
@Jura1: Can you stop misrepresent the reality please? First, you claim that a member of EveryPolitician wants to rename WikiProject Politicians to WikiProject every politician. In fact, as noted by Andy Mabbett, you started the discussion to do the contrary. Then, you claim that EveryPolitician duplicated WikiProject Politicians with WikiProject every politician. Check the history. EveryPolitician created both projects, WikiProject every politician first, which is only a renaming of Wikidata:EveryPolitician (yes, a redirect is missing), and WikiProject Politicians, which was never active. We know that you don't like EveryPolitician and that you don't care about breaking other people's work on this topic (« it has been incredibly disruptive and destructive », « work which you have now broken », etc.), but please, stop mislead people with a reality that is shared only by yourself. — Envlh (talk) 00:54, 17 November 2017 (UTC)
According to a paid contributor of EveryPolitician: "we created Wikidata:WikiProject Politicians before deciding to call it WikiProject every politician instead.". If you want to discuss the datamodel that EveryPolitician tried to introduce to Wikidata (with WMF grant money) and finally dropped, we can do that. Obviously, feel free to participate in the discussion: Wikidata_talk:EveryPolitician#too_much_specific_items_for_legislature_members.
--- Jura 07:16, 17 November 2017 (UTC)
There seems to be a degree of misunderstanding as what "EveryPolitician" is here. EveryPolitician is not an organisation — it's a project set up around three years ago by mySociety (Q10851773), a UK-based charitable civic-tech organisation, to try to create a database of every politician in the world, starting with gathering that data for (almost) every national level legislature. In the last year we have changed the focus of that project so that Wikidata can become that database, rather than us duplicating effort. As part of that we have received a WMF grant — but that is very much not about simply importing data from one project to the other (which is neither possible nor desirable). One strand of this is being able to track the quality of existing political information in Wikidata, to help see gaps and errors etc. We are currently doing that at Wikidata:WikiProject every politician. We had previously created an initial stub at Wikidata:WikiProject Politicians, but that was never used (or even known by anyone beyond Lucy and myself), and I believe this should be simply redirected. This is not us somehow trying to hijack a pre-existing project and rename it, as the initial description here could imply. This is us making a tentative start at one name, and then deciding to switch to one much more evocative of what we're actually hoping to achieve, similar to Wikidata:WikiProject sum of all paintings. --Oravrattas (talk) 07:54, 17 November 2017 (UTC)
Just to correct, the import initially was at Wikidata:EveryPolitician (named as the website this comes).
--- Jura 08:06, 17 November 2017 (UTC)
I edited the above comment as the a page was redirected while the discussion is still ongoing. https://www.wikidata.org/w/index.php?title=Wikidata:EveryPolitician&oldid=565751502 is probably the better link for "Wikidata:EveryPolitician" mentioned above.
--- Jura 18:18, 22 November 2017 (UTC)
As has been explained earlier, that was set up in 2015 by Andrew Gray, to look at how to use the external mySociety project within Wikidata, long before the grant or the mySociety switch to making Wikidata the replacement approach (in fact it was that effort, and the things that we both learned from that, that led directly to this "reboot".) That earlier approach is largely obsolete now, as the goal is not to simply harmonise the two datasets, but to build a project and community within Wikidata to generate and maintain political data. --Oravrattas (talk) 08:26, 17 November 2017 (UTC)
Maybe it's worth mentioning that Wikidata:EveryPolitician is the page that was on the GBP 40,000 grant application (mw:Grants:Project/mySociety/EveryPolitician#Documentation).
I find it problematic that you keep deleting content from project documentation pages while it's being discussed in the forum.
To clearly differentiate WMF websites from EveryPolitician, I think we should restore the more appropriately named Wikidata:WikiProject Politicians and Wikidata:EveryPolitician.
--- Jura 08:45, 17 November 2017 (UTC)
I edited the above comment as the a page was redirected while the discussion is still ongoing. https://www.wikidata.org/w/index.php?title=Wikidata:EveryPolitician&oldid=565751502 is probably the better link for "Wikidata:EveryPolitician" mentioned above.
--- Jura 18:18, 22 November 2017 (UTC)
Well, yes, obviously when we applied for the grant we pointed at the page that already existed and led us towards making the application, rather than the project that we didn't create until after receiving the grant. If any of the links in the project proposal or elsewhere are now confusing because they point to information that existed then, rather than now, we can easily change those links to point to historic versions of URLs, but as one of the problems you're raising is that duplication is confusing, reverting changes that try to harmonise things and clean everything up doesn't help matters much. --Oravrattas (talk) 10:38, 17 November 2017 (UTC)
Links to previous versions of pages are helpful if the content evolved, they are only marginally useful when it was merged/moved and replaced by something else.
In the current situation, it seems that EveryPolitician amalgamates anything related under then name of its website. The result is that things are less structured than before the GBP 40,000 grant.
--- Jura 12:04, 17 November 2017 (UTC)
@Jura1: Why do you feel the need of a making the distinction clear. Grant funded projects like Woman-in-red also had their own Wikiprojects. ChristianKl (✉) 16:27, 22 November 2017 (UTC)
There isn't really a problem in having a project with an appropriate scope. There used to be one with an appropriate scope at Wikidata:EveryPolitician.
We also have wikisource:fr:Wikisource:Partenariats/Bibliothèque nationale de France. It would just be odd if they would use it to replace "WikiProject Books". Otherwise people might think that Wikidata was some satellite site of BNF as all book related info comes from there. Given the page redirecting done by a paid contributor of MySociety's-EveryPolitician while we discuss this, it appears even more problematic.
--- Jura 18:18, 22 November 2017 (UTC)

Birth date of a geisha

Dear community, I am not aware of the Wikidata rules that applied concerning the birth date of living people who do not want that others know it. If you know how we are supposed to manage such case, please let a message here. Pamputt (talk) 06:43, 17 November 2017 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ As a starting point for discussion, the en.Wikipedia policy is here: en:Wikipedia:DOB, and says:

With identity theft a serious ongoing concern, people increasingly regard their full names and dates of birth as private. Wikipedia includes full names and dates of birth that have been widely published by reliable sources, or by sources linked to the subject such that it may reasonably be inferred that the subject does not object. If the subject complains about the inclusion of the date of birth, or the person is borderline notable, err on the side of caution and simply list the year,

-- Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:19, 17 November 2017 (UTC)

  • The conclusion of enwiki is a bit odd, maybe US-centric. birthday (P3150) can be used if the full date shouldn't be included, but the day/month can. Some wikis routinely include that, but not the year.
    In the current case, it seems that the date would be hard to reference. Maybe we could simply add the century.
    --- Jura 11:58, 17 November 2017 (UTC)
Her year of birth was included in the bibliographic information in the book she authored, and the Library of Congress had her full birthday at LCCN. Someone pretending to be her emailed them to have it deleted, or Fiona Graham herself was impersonating someone advocating on her behalf when they contacted me and told me they were going to write LCCN and have it removed. It is possible that all the SPAs that edit her English Wikipedia entry are her, I do not know. Both choices are disturbing. --RAN (talk) 00:49, 18 November 2017 (UTC)
Wikidata:Living people (draft) does not include date of birth (P569) on its list of statements likely to violate privacy (or be challenged), but perhaps it should? This is the first example I'm aware of where data about a living person has been asked to be removed in this manner. ArthurPSmith (talk) 15:49, 17 November 2017 (UTC)
This is about w:Fiona Graham and she doesn't want people to know how old she is. A person wrote me asking me to remove the date in the English Wikipedia, because geishas do not reveal their age, and this person said they were advocating on the behalf of Graham. When I showed that the birth date was a matter of public record, they said they would get that record removed ... and they did! By pretending to be the geisha, or the geisha was pretending to be an independent advocate when they wrote me. So, identity deception does occur. That geisha's record has at least 10 SPAs editing it, they can be seen in red. Geishas may have a 100 year old rule about not revealing their age, but we need to follow the 2,000 year old rule that an encyclopedia should be complete and accurate. If you want the anonymity of a geisha life, then do not become an author and reveal your age and your real name in a book. The SPA User:Lilly1985 is deleting the information here. The argument was that geishas do not reveal their age and since that did not get enough support, the argument has been reworded as to prevent identity theft. We all get old, deal with it. --RAN (talk) 00:16, 18 November 2017 (UTC)
I don't think it leaves a big gap if we delete/oversight it. The only problem is that it's likely to be re-add unless some more general statements is added to Wikidata instead.
--- Jura 07:40, 18 November 2017 (UTC)

I'm intrigued as to what is supposedly "US centric" about the quoted en.Wikipedia guideline. Can anyone enlighten us? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:47, 18 November 2017 (UTC)

The en.Wikipedia guideline considers the birthday more worthy of protection then the year of birth. The Japanese consider the year of birth to be more worthy of protection. The Gaisha has no real problem with her birthday being publically known but she has an issue with her year of birth being publically known. ChristianKl (✉) 19:50, 18 November 2017 (UTC)
Is this about her hiding her age, or is it a custom in Japan to not reveal your birth year? Can you point me to a reference on the custom, if that is the answer. --RAN (talk) 21:22, 18 November 2017 (UTC)
That would be an explanation if everywhere that is not Japan is the USA. Clearly, that is not so. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:52, 18 November 2017 (UTC)
In the discussion that lead to the creation of our birthday property someone advocated that we need the property because Japan has many people for which the birthday is public knowledge while the birthyear isn't. I didn't introduce the term "US-centric" into this discussion but I think he differences between US (or Western) norms and Japanese were the point of contention. ChristianKl (✉) 09:58, 20 November 2017 (UTC)

I don't think we have rules. I would be okay with a rule that says: "If a subject request information about them to be removed that isn't of public interest, we will delete/oversight it". I'm also fine with adding some qualifier before "public interest". Given the importance of birthdates for disambiguating people I would however make no special rule for it. ChristianKl (✉) 19:40, 18 November 2017 (UTC)

How do we know we are in communication with her, there are at least 10 SPAs deleting her birth date from the English Wikipedia article, how do we know we are not being catfished? --RAN (talk) 21:26, 18 November 2017 (UTC)
My point of view is the following, since this information (her birth date) is easy to find on the Web, we cannot avoid, if we decide to remove this information, that someone readd it in the future. So if this geisha does not want that people know her birth date, she should start to ask to other websites (VIAF, upclosed.com, ...) to remove this information. Once the information will not be available, it could be removed here. Pamputt (talk) 22:48, 18 November 2017 (UTC)
Amazon refused to removed birth dates/ages from IMDB They decided to not tolerate forced "removal of factual information from the public domain." --RAN (talk) 03:41, 19 November 2017 (UTC)
I don't think the harm done to Wikidata by someone impersonating her to get her birth date removed is big. Having a unburocratic procedure where a person like her can ask for having data removed has the advantage that it will prevent certain conflicts over privacy for getting big heated issues.
As far as preventing data from being readded we could have "birth date: unknown" with a qualifier that suggest that the data was removed because of a request from the person in question.
As far as the Californian law goes it's worth pointing out that AB-1687 is about "commercial online entertainment employment service providers". Neither Wikipedia nor Wikidata are a "commercial online entertainment employment service provider" and thus the law doesn't apply here even through Wikimedia is registered in California. ChristianKl (✉) 13:35, 22 November 2017 (UTC)

Language codes revisited

Hoi, at this time the introduction of new codes to be used in Wikidata has all the outside qualifications of a stamp collection. There are no arguments given why particular codes are to be introduced. It is silly. It is silly particularly because with the introduction of Wiktionary data there will be an influx of codes that will have a known use; they are after all what Wiktionary offers at this time. For this reason I want us to stop adding codes unless there is a strong argument and concentrate on the needs for identifying Wiktionary content for the language, the dialect it is said to be. Thanks, GerardM (talk) 07:07, 17 November 2017 (UTC)

I do not really understand what you are talking about. Do you have any link that points to a page where the problem is explained? Pamputt (talk) 16:04, 17 November 2017 (UTC)
#Monolingual language code creation process. Matěj Suchánek (talk) 09:42, 18 November 2017 (UTC)
Thanks for the link. I get the point now. To reply to GerardM I do not see any relation between adding language code and Wikidata support for Wiktionary because from what I understood, Wiktionary lexeme would use Q-item for languages and dialects, not language code. Could you develop a bit more? Pamputt (talk) 13:56, 18 November 2017 (UTC)
So your English has no relation to English as per the standard? Thanks, GerardM (talk) 14:20, 18 November 2017 (UTC)
As soon as a person adds name in native language (P1559), native label (P1705) or inscription (P1684) they need a code. If a person wants to record 10 entries with one of those, they need to have the language code acccepted. It seems that you argue that this isn't enough and it's bad to add languages that are only seldomly used. Can you give an argument why it's bad to have many codes? ChristianKl (✉) 19:45, 18 November 2017 (UTC)
No that is not what I said. The same person has requested many codes. This is not a stamp collection so even though it might be useful, it is requested without arguments. The other reason is that there have codes requested that are questionable. The final reason is that we are looking towards a great amount of languages from Wiktionary that have to be mapped to codes. Thanks, GerardM (talk) 05:31, 20 November 2017 (UTC)
About Wiktionary, as far as I known, the language will use Q-item and not "Wikimedia code". Do you have any link that explains the opposite? Pamputt (talk) 18:37, 21 November 2017 (UTC)
If something is "questionable" then ask questions in the discussion where the code is requested. There might be a need for a name in native language (P1559)/native label (P1705)/inscription (P1684) even for languages that are not covered in Wikidictionary.
It's also my understanding that the Wikidictionary integration isn't planned to focus on languages codes but on Q-items. ChristianKl (✉) 16:48, 22 November 2017 (UTC)

Military Service numbers

While we have United States Armed Forces service number (P2028), we don't have the equivalent for other countries. I added a service number for a British sailor in this edit; is there a better method? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 19:14, 20 November 2017 (UTC)

I remember saying at the time of the property proposal that we should have a generic "military service number" rather than such a narrowly defined property (US only, 1918-1974), and we could use a qualifier to indicate the affiliation (or numbering system, etc). It's only been used a handful of times - perhaps now is a good opportunity to make that change? Andrew Gray (talk) 21:33, 20 November 2017 (UTC)
Well, I have a couple of data sources for UK & commonwealth people... But then, what does a dedicated, qualified, property offer us over catalog code (P528) with the same qualifiers? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:28, 20 November 2017 (UTC)
Well, potentially, any identifier could be catalog code (P528) with appropriate qualifiers ;-).
More seriously, I think the major benefit is that you can do a query to say "give me P2028 service number, and whatever numbering system or branch it's in", without already knowing the qualifier, but confident that you'll get something appropriate to the general topic. If you use P528 then you might get back anything that's been associated with that field, including non-military stuff; if you don't already know what the qualifier is going to be then you can't filter on it. Something like WWI UK service numbers (which were remarkably granular - at points individual battalions had their own sequences) would make it hard to predict qualifiers in advance. I think for a broad class of numbers it's fair to split it out. Andrew Gray (talk) 22:34, 20 November 2017 (UTC)
Any identifier with no formatter URL could be. But your other points is well made. If we're going to repurpose P2028, how should we proceed - is this discusison sufficent? 10:02, 21 November 2017 (UTC)Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits
It's currently a unique identifier, so generalizing it isn't a good idea.
--- Jura 13:22, 21 November 2017 (UTC)
Did you not read the preceding discussion, where Andrew wrote "It's only been used a handful of times"? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:10, 21 November 2017 (UTC)

Using ranks for false statements

For various items, multiple sources may claim different things. In taxonomy, it might be that a taxon is an accepted name, or that it might be an unacceptable synonym. For humans, it might be a birth year which may be conflicting within different sources, or that a human might not have actually existed at all and just been a fictional myth. For these statements, in my opinion, we should include all possible statements, and using Wikidata:Ranks to help show which is the most acceptable statement. This reasoning seems to be supported by Wikidata:Tours/Ranks, which states that:

Within a large knowledge base, it's still important to record the previous values of items. These records—in the form of statements with multiple values—help us to better understand the world, see patterns and relationships, and make connections and predictions based on what we already know. [...]A deprecated rank is used for a value that contains information that may not be considered reliable or is known to include errors. For example, an item of a city may feature an incorrect population figure that was published in a historic document. The statement is not wrong as the figure is accurate according to the (erroneous) historic document, however because it known to contain errors it should receive the deprecated rank.
—Wikidata:Tours/Ranks

I have found sources (for which we have external identifiers as well) that claim Phalaena citrata Linnaeus, 1761 (Q43242043) is an acceptable taxon, or at least it once was and therefore sinstance of (P31):taxon (Q16521) and parent taxon (P171) etc. should be added, since we should not do original research and determine which sources are "right", and simply add all data we can find, and add citations to them (and later rank these statements with Wikidata:Ranks. However, Brya removed these statements (which I should have added the sources for), and added editing notes in the item description "name that may not be used" (something which should be on the talk page). This name was once used, and we should add statements to show that (and have "end dates", as well as mark such statements as deprecated).

I brought up this question on Wikidata_talk:WikiProject_Taxonomy#Deprecated_status, where I thought would get some positive response that my reasoning was sound, but instead got responses such as "We don't need all the GBIF garbage" (meaning that we should not include statements based on that source), "our main goals is two fix errors we find in the Wikipeadias" (we should right wrongs, and do orignial research and exclude sourced data and statements, based on our own judgements, and not include all possible information, despite being sourced; see the incorrect population figure in the quote above from Wikidata:Tours/Ranks for a counter argument). I also got a response that including all available data is original research, and instead that excluding statements which most sources do not mention is not. And lastly, that we should not use Wikidata:Ranks because third-party tools and queries might not work properly with these.

It is my understanding that we should include as many statements as possible, right or wrong if we can properly cite them. Is that an incorrect understanding of Wikidata's goals and purpose? "There is no truth, only sources"... Thankful for insights from other users. (t) Josve05a (c) 19:46, 20 November 2017 (UTC)

  • Yes, that is how wikidata is supposed to work - even a P31 statement could be deprecated or given an end date if it is no longer considered valid; reason for deprecated rank (P2241) can be used to provide additional context for deprecation. We also have statement disputed by (P1310) to clarify where a statement may be disputed. Deprecated statements should not be returned in most queries, but they should not be deleted unless it is clear they were never supported by the original source (or were unsourced). @Brya: care to explain what's going on here? ArthurPSmith (talk) 21:41, 20 November 2017 (UTC)
Josve05a should know better. The WMF vision speaks of making "the sum of all knowledge" available; that is knowledge, based on information, based on data. The purpose is to have a place where the reader can go to when he does not want to be confused by commercial claims, propaganda, false news, drivel, etc.
        To stay with the population figure, suppose there is a database, somewhere out there, that says that publicist A states that the population of New York City in 2016 was 69, while publicist B holds that the population of New York City in 2016 was 254, then that is no reason to copy that into Wikidata, not even with a deprecated status.
        If a user cannot tell right from wrong he does not belong in the Wikimedia franchise and should not make any edit: being able to tell right from wrong is what he is here for. At the Wikipedias there are policies to help him tell right from wrong, such as NOR, NPoV and VER. And competency in the field he addresses is expected. It is only when there is doubt among the experts, in reliable literature that a user should avoid labelling something as "wrong" or "right": if the literature is divided on the question of whether cattle constitute a species or a subspecies then it is not allowed to say that the name of cattle is Bos taurus and that the name at the rank of subspecies, Bos primigenus taurus, is wrong: there is lots of support for either position.
        What Josve05a proposes to do here is to replace accurate information by junk, and apply a label "deprecated" as if that would reduce its misleadingness. In the matter of naming organisms, scientists have laboured for a century and a half now, to create order out of chaos, coming to international agreement on rules to govern names, and on rulings on what names may not be used (ever). This is one of the big successes of international cooperation (which is not quite achieved in measurements, a much simpler matter: part of the world uses inches, while the rest uses centimeters). Mostly this nomenclature draws quite black and white distinctions. The ICZN (Q1071346) has ruled (Opinion 450, 1957) to suppress the name Phalaena, which means that it cannot be the name of a taxon, and that none of the combinations with Phalaena can be the name of a taxon. There is good reason for this, as the name caused great confusion in nomenclature. The ruling is as black and white as can be: there is no confusion. The Wikidata entry (Q11887871) clearly states that this is a suppressed name and even links to the ButMoth ID page which says so. That it was once (long ago) regarded as a taxon is unmistakably included in that statement. No need to introduce confusion. - Brya (talk) 06:17, 21 November 2017 (UTC)
We generally don't remove wrong information when there are references for the information but deprecate statements. I see no reason why there should be a depricated instance of (P31) taxon statement here. There's no reason to have special rules for taxons that differ from how we deal with wrong information that can be backed up by references in other domains in Wikidata. ChristianKl (✉) 10:35, 21 November 2017 (UTC)
@Brya: It is also information to know that a widely repeated wrong statement is in fact wrong (or no longer correct). Jheald (talk) 23:56, 21 November 2017 (UTC)

Without expressing a view on any of the above, I have changed the description of Phalaena Linnaeus (1758) (Q11887871) from "suppressed name (do not use)" to "suppressed taxon name", and added the alias "Phalaena". Also, I am not convinced that the claim "instance of (P31) = Noctuidae (Q459180)" is helpful. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:26, 21 November 2017 (UTC)

instance of (P31) = Noctuidae (Q459180) seems wrong. If anything subclass of (P279) would be appropriate. ChristianKl (✉) 10:36, 21 November 2017 (UTC)
Well, "subclass of NOCTUIDAE" is guaranteed to be wrong, given that a name cannot be a class, let alone a subclass, nevermind a subclass of a class of organisms. Indeed, "instance of NOCTUIDAE" admittedly looks odd, but at least indicates a relationship. Given the confusion surrounding many of these objectively invalid names this is of eminent practical importance. - Brya (talk) 11:55, 21 November 2017 (UTC)
I see no reason why names shouldn't be able to be classes in some special cases but if the item is about a name that it surely isn't a instance of NOCTUIDAE instances of that should be physical living beings. ChristianKl (✉) 16:55, 22 November 2017 (UTC)
And Brya promptly reverted my change to the description, with the edit summary "not a taxon", even though my wording did not say that it was a taxon. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:13, 21 November 2017 (UTC)
Seems to be an abusive pattern of reverting without dicussing. (t) Josve05a (c) 15:26, 21 November 2017 (UTC)
I cannot even begin to imagine the reasoning whereby a "taxon name" refers to anything but a taxon. What is it supposed to refer to, a pretty girl in a polkadot bikini? A "suppressed taxon name" is a contradictio in terminis. If more explicitness is felt to be desirable "suppressed zoological name" is an option. - Brya (talk) 17:31, 21 November 2017 (UTC)
Never mind a girl in a bikini; if I want to name a cat, which is undoubtedly a zoological entity, "Phalaena", no-one - not you, and not even the mighty ICZN - can stop me. As a cat name, an asteroid name, a house name, a car name, or, yes, a girl-in-a-bikini name, it is not suppressed. In fact the only way in which that name is suppressed is as a taxon name. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:00, 21 November 2017 (UTC)
Ah, like addressing Bill Clinton as "President Clinton". I can more or less understand that, if I twist my mind enough. That does not mean it is not highly (and needlessly) confusing. - Brya (talk) 05:09, 22 November 2017 (UTC)
No, nothing like that at all. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:31, 22 November 2017 (UTC)

It is not the first discussion about User:Brya using their own research to determine which sourced statements to keep and which to remove. This discussion about Heterocheilidae (Q21440769), a family of insects, which had a misfortune of being given a name which was already used by a family of worms. The name is the latest used in scientific publications and databases and referring to both families (sometimes in the same publication). However the whole family was excluded from Wikidata taxonomic tree structure, as there is no item for family of insects Heterocheilidae with a parent taxon (P171) property and no items that link to Q21440769 through parent taxon (P171). To me it is an extreme case of own research used to suppress information in published sources. If an information in an external reference is no longer considered correct, than we a add depreciated rank but we do not delete information that at some point it was considered true, and in case of Q21440769 I do not think there was any publications about depreciating that name. --Jarekt (talk) 21:01, 21 November 2017 (UTC)

„The family-group name Heterocheilidae (McAlpine 1991b), which is based on the genus Heterocheila, is a homonym of the family-group name based on Heterocheilus Diesing 1839 (Railliet & Henry 1915, Nematoda), and a petition to the Commission of the ICZN will be needed to establish the proper name for these families.“ - I think Heterocheilidae McAlpine (1991) non Railliet & Henry (1915) (Q21440769) tries to express this. Suggestions modeling cases like this one are welcome. --Succu (talk) 21:41, 21 November 2017 (UTC)
Looks like a Wikispecies user came to the same conclusion in 2010. So please be more carefull with statements like „using their own research“, Jarekt. --Succu (talk) 22:01, 21 November 2017 (UTC)

automatically adding spouses

I just had an idea about Wikidata. I think it would be convenient if when you add a spouse to a person, that it automatically gets added for the spouse as well. For example, I recently made the item Eliza Maria Partridge Lyman. I added that she was a spouse of Joseph Smith. It would be nice if she were automatically added as a spouse on Joseph Smith's page too. Does a bot that does this already exist, or is there a reason it doesn't work that way? I'm trying to increase my Wikidata fluency. Thanks Rachel Helps (BYU) (talk) 21:42, 20 November 2017 (UTC)

@Rachel Helps (BYU): while it is not quite what you are requesting, the primary reference gadget will usually pick these up and offer them at a point of time as a suggested addition on the spouse. [I don't know the internal working mechanics of the gadget to be more helpful of a timeframe).  — billinghurst sDrewth 03:36, 21 November 2017 (UTC)
@billinghurst: thanks for the tip. Looks like an interesting gadget! Rachel Helps (BYU) (talk) 16:43, 22 November 2017 (UTC)

Wikidata and Wikipedia vs. Wikisource

Currently Wikisource deletes links to people and places named in the text that is archived there. Do you think identifying people and places in documents is important? Links to Wikipedia and Wikidata are deleted. The argument at Wikisource is that there is no way to determine if the links are correct. The same argument can be used on any Wiki project, but the recourse is always the same, we correct mistakes when they are identified. Here is an example and another example here. --RAN (talk) 07:34, 21 November 2017 (UTC)

en.wikisource does, there's other policies in other projects. What does it have to do with Wikidata? --Anvilaquarius (talk) 08:11, 21 November 2017 (UTC)
well, this is a matter of wikisource policy. On frwikisource, we don't accept links inside a text either, with very few exceptions, because the aim of wikisource is to provide a text, not to create encyclopedic links. Also, most of our texts are to be exported and read without connection, which renders these links useless AND gives a poor reading experience, because of the formatting of the links... Generally, the contributors who want to add links are wikipedians at heart, and occasionally contribute on wikisource, and do not understand clearly that the aims of wikisource are completely different from wikipedia's...
in any case, this is a matter that should be discussed in the concerned wikisource project. I completely fail to see how links in the text (not in the sidebar) could have anything to do with wikidata... --Hsarrazin (talk) 08:29, 21 November 2017 (UTC)
RAN: you misrepresent enWS's linking policy. Wikisources are not an encyclopaedia, a sea of blue links are not as pertinent, though we do have links in our works where relevant, and as appropriate, most definitely from headers. Correct that we do not allow interpretative links in the body. Guidance for enWS is at s:en:WS:Wikilinks; people can discuss the policy and its implications at s:en:WS:S after reading the guidance.  — billinghurst sDrewth 11:30, 21 November 2017 (UTC)
  • I only see aesthetic arguments like the "sea of blue" argument for not linking. What use is hosting a letter that talks about a Colonel Smith during the Civil War and not linking to him in Wikipedia or Wikidata? Context is everything. --RAN (talk) 02:51, 22 November 2017 (UTC)
    @Richard Arthur Norton (1958- ): That is simply incorrect, if the work is about Colonel Smith, then the header link to Colonel Smith at enWP is there for use. Re Wikidata link. When you register the work in Wikidata and add the wikilink, the work will have the link to Wikidata in two places, in the left side bar, and in the header, and if you have completed data here appropriately you will presumably have the main subject field for that work will be Colonel Smith. What more are you wanting? If you do not understand, then please ask at s:en:WS:S and the community will assist.  — billinghurst sDrewth 05:17, 22 November 2017 (UTC)
No, I understand completely. Links to people and places are removed. Imagine a Wikipedia entry with just two links. A letter may contain the names of a dozen people, and refer to a dozen place. --RAN (talk) 05:24, 22 November 2017 (UTC)
And to be specific, there is a section on linking to Wikidata for the headers of works. There would be zero requirement for a reproduced work to have links within the body of the work that point to Wikidata; there is simply no case for it.  — billinghurst sDrewth 11:36, 21 November 2017 (UTC)
"No case?" You may not like the case, but that claim is simply untenable. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:09, 21 November 2017 (UTC)
That's a great pity - it goes against what the web is for, and it ignores the technical tools to suppress or relocate (e.g. to footnotes) deemed-obtrusive link styling. But it's also not a Wikidata issue. What en.WS (and en.Wikivoyage, who have the same short-sighted view) may wish to look at, if not links, is structured tagging, using Wikidata items as the target. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:09, 21 November 2017 (UTC)
What are you talking about Andy? I am talking about not visibly tagging components within the (body of a) reproduced work to point to Wikidata, nothing else. We already allow the linking of every work or distinct sub-parts of works to Wikidata.

The Wikisources have for many years wished for structured data, we have tried to implement vcards in our headers, and we have looked forward to Wikidata-hosted data enabling better search, etc. We are truly open to guidance, about what can be done in the broader sense to interlink and interweave and have wikidata/functions available, though we probably lack the technical skills to do it well.

I would also much prefer that this discussion, if here, was to focus on how the sister wikis can better produce their works with available wikidata, and how wikidata community can help; rather than the lame argument of "add more visible wikilinks to the body of your reproduced works". At Wikisources most of us are good at presenting a reproduced work, we yearn the expertise of opening access to the world more efficiently.  — billinghurst sDrewth 05:54, 22 November 2017 (UTC)

It appears that Wikisource currently has no method to link from (to pick a random item) s:Peebles, David (DNB00) to, say, Protestant Reformation (Q12562), David Laing (Q5236288), Dublin (Q1761) or British Museum (Q6373); nor their Wikipedia equivalents; nor indeed to give any structured, semantic indication of any of the concepts discussed in that work. I've suggested at least three different means of doing so: Use wiki-links which are styled not to display (or whose styling is in the hands of the reader); put the links in footnotes; or use tags, like on blog posts. Note that none of these is an argument, lame or otherwise, to "add more visible wikilinks to the body of your reproduced works". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:04, 22 November 2017 (UTC)
I see at least one good way to do this : once the text gets a wikidata item, adding the subjects to the item, and then display them automatically, not inside the text, but in a box or menu, that links to subjects ;) --Hsarrazin (talk) 13:12, 22 November 2017 (UTC)
@Pigsonthewing: I don't think that the Wikisource community would have issues with the general slant of what you are trying to achieve. It isn't contrary to our goals of presenting works. Whether it is done here or in structured data at the WS. Those biographical articles are ripe for that sort of data picking/tagging, similarly works of the Popular Science Monthly. Ideally we want that sort of data more tightly aligned with WD items as that way we are not having to try and manage moved/deleted/disambiguated articles at WPs of the very real issues. The WSes don't have the people to curate links, and linking every vestige of metadata in articles. Have a look at something like s:Page:The Indian Biographical Dictionary.djvu/150 and imagine if that was whacked with wikilinks rather than had structured data.  — billinghurst sDrewth 14:04, 22 November 2017 (UTC)

Why Wikisource, then consider links

Please consider something first. At this time we do not consider the use of finished products of Wikisource. We do not know how much they are used and who uses it. We are not consider how to get a wider audience. What is considered at this time are details and, the consequences of what is decided has an impact on the user stories for what Wikisource provides.

Wikidata COULD play a role in using references in texts and elsewhere. At the same time these references can be hidden. The biggest issue I have with Wikisource is that it does not consider its audience. I know of Websites that package Wikisource because Wikisource does not do this. MY request to you all. Let us FIRST consider how we can make a bigger difference to our audience and then consider the user stories and only then consider if links in a text add value. Thanks, GerardM (talk) 12:12, 22 November 2017 (UTC)
Nothing there that many at the WSes would vehemently disagree. They are transcribers/proofreaders/researchers ... the detail; they are not necessarily the article writers, the data specialists, the librarians. Works can be fiction, can be poetry, can be non-fiction, can be correspondence; can be out-of-date science, can be copies of quackery, it is generally an unfiltered presentation of a time and a place. It is not meant to directly correlate with WPs, or WD, but it is a part of it like any library or archive.  — billinghurst sDrewth 14:16, 22 November 2017 (UTC)

Wikidata weekly summary #287

Wikidata weekly summary #287 Global message delivery/Targets/Wikidata

 
Here's your quick overview of what has been happening around Wikidata over the last week.
  • Development
    • Fixed Commons media suggester not loading all preview thumbnails (phabricator:T160528)
    • RDF exports can handle entity references to foreign Wikibase repositories now, needed for Structured Data on Commons (phabricator:T161592)
    • Worked on making statements on Forms editable and persistent (phabricator:T163724)
    • Worked on fine-grained usage tracking in Wikibase' Lua library (phabricator:T172914)
    • Worked on caching constraint check results (phabricator:T179849) and indicating when constraint check results are cached (phabricator:T179844)
    • Wikibase code base does not use Composer class autoloading any more (phabricator:T180067) as another step towards getting rid of our own build process and being more in line with the rest of Wikimedia
When it comes to the ChaosRadio interview, I think saying that we have more information than enWiki because we have 37 million items and they have only 5 million articles is misleading. There's often no 1-to-1 relationship. An article like https://en.wikipedia.org/wiki/Anatomical_terms_of_location#Anterior_and_posterior talks about many different "knowledge items". Wikipedia tends to mix multiple knowledge items into one longer article when the individual items aren't worthy to have their own article. An encyclopedia profits from having one article that covers multiple issues while Wikidata profits from separating issues into multiple items. In most domains enWiki contains much more "knowledge items" than Wikidata.
On of the problem of spreading this misconception is that it encourages people to merge items that we don't want to have merged. ChristianKl (✉) 12:33, 22 November 2017 (UTC)

Google Translator uses Wikidata

The sentence Grade I listed historic house museum in the United Kingdom only appears on Wikidata and is perfectly translated to German with "Denkmalgeschütztes historisches Hausmuseum im Vereinigten Königreich" which is not literally translated. --Bigbossfarin (talk) 20:21, 22 November 2017 (UTC)

And the reverse gives "Listed historic house museum in the United Kingdom", which is not exactly the same as the English phrase above, nor as precise, but is passable. Good spot. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:39, 22 November 2017 (UTC)

I want to add a sentence to this story.

en:Bazy Tankersley#Background and personal life

In paragraph 3: I was the FIRST Apprentice in Bazy Tankersley's Apprentice Program.

How can I add this:

"Joe Dennis Sysel was the very first Apprentice in this program. Bazy Tankersley eventually entrusted the ownership of her 1967 National Champion Arabian Stallion, Count Bazy, to Joe Sysel, telling him, "You have an unlimited amount of time to pay for him." He was paid for within 3 years.  – The preceding unsigned comment was added by 2001:48f8:1044:9b1:cc29:d33:fd74:fda4 (talk • contribs) at 21:04, 22 November 2017‎ (UTC).

You've arrived at a project called Wikidata; please ask for help on Wikipedia. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:40, 22 November 2017 (UTC)

Create item for each device model in Wikidata?

From the viewpoint of the Wikidata project: Is it desirable to have each device model as an item? E.g. have an item for each row of en:Comparison of Fritz!Box devices including and maybe even properties for each column. Same for other device categories e.g. (smart) meters, lamps and other sensors/actuators. --Saerdnaer (talk) 10:24, 20 November 2017 (UTC)

In some cases I might say "it depends". I think anything where's the information is notable enough to be written down in a table in EnWiki is clearly important enough to be within Wikidata. There's a goal to develop the capability to automatically generate tables from Wikidata and to do that we have to be able to store all information that's of interest to the various Wikipedia projects. ChristianKl (✉) 11:45, 20 November 2017 (UTC)
"Goal"? Listeria does it already. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:12, 20 November 2017 (UTC)
The problem with listeria is that the list isn't editable from the Wikipedia side and as a result the major Wikipedia to my knowledge aren't very welcoming of listeria tables. As far as I understand WMDE wants to develop further capabilities and I think that's needed before pushing a lot of data this way to Wikipedia. ChristianKl (✉) 17:35, 23 November 2017 (UTC)
Yes. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:12, 20 November 2017 (UTC)

New Identifier property for a dictionary of terms?

Hi everyone. I am currently working with Bangor University to import several thousand Welsh language labels for items on medicine, health and science. The translations come from an online Welsh dictionary of terms which provides standardized terminology for the field of education. Before I make a formal request for the creation of an identifier property i just wanted to check that this would be suitable as an identifier? It would help us manage the use of the correct Welsh language terms on Wicipedia (CY) and enable us to track and update the terms in the future. The creation of the identifier would also help to build up our relationship and could lead to further data donations in the future. Any thoughts or advice would be greatly appreciated! Jason.nlw (talk) 14:11, 22 November 2017 (UTC)

It's likely suitable as a identifier but without having an actual link to the online dictionary it's not a conclusive decision. ChristianKl (✉) 14:29, 22 November 2017 (UTC)
Thanks ChristianKl. Here is a link to an item in the dictionary. They need to build their unique identifiers into their URLs which they have said they will do, if needed. Cheers Jason.nlw (talk) 15:25, 22 November 2017 (UTC)
If it is a dictionary the links should probably be on the lexemes pages when the Wiktionary on Wikidata project is going to be deployed. So, if you wait a few month there is going to be a much better place to add your mapping to. Tpt (talk) 15:57, 22 November 2017 (UTC)
It also seems to me like the dictionary is more about words (-> lexemes) than it's about concepts (-> items). ChristianKl (✉) 16:13, 22 November 2017 (UTC)
A dictionary's word has the definition of a word. These words, which I created, the administrator deleted without discussion. Let's see how she will behave afterwards. --Fractaler (talk) 09:35, 23 November 2017 (UTC)
Hi ChristianKl. Thanks again. I think it might be more than a few months until lexemes is up and running! but yes i agree, this would be a much better place to have the identifier. In the mean time however we need a way to track these labels and to demonstrate that the Translation comes from a reliable source. I guess this is one reason why the Lexemes project is needed! Jason.nlw (talk) 10:00, 23 November 2017 (UTC)
Indeed that's one of the reasons why integrating lexemes is great. As far as demostrating that translation of the name of a label comes from a reliable source, we currently don't have a process where we source translations of label. There's also often a need to use a different label for an item than the one you would find in the dictionary. "animal behavior"(animal behaviour (Q2990593)) is for example called "behavior" in the Gene Ontology. The German National libary calls it "Vehalten" (the label also doesn't have information about animals) and the same thing is true of the Italian BNCF Thesaurus. On Wikidata we have behavior (Q9332) and it's worthwhile to define that broadly enough that the term can be used for bacteria, plants or robots.
The English word "arm" can mean 4 different Wikidata items without even distinguishing human arms from arms in general. When you automatically match translations you will get some of them wrong and that's okay. If someone comes along and finds a better one they can just give it a better label without providing a source for the better label. ChristianKl (✉) 10:54, 23 November 2017 (UTC)
Maybe name (P2561), but the website doesn't seem to be one that is easy to link to. Especially if you want to link to a specific concept linked to a word there.
--- Jura 11:09, 23 November 2017 (UTC)

Associating church and its churchyard

What is the accepted way, if any, of linking Nissi Church (Q14955401) with its churchyard, Nissi churchyard (Q20529638)? (notice both are separate Estonian heritage sites and should almost definitely remain as separate items). I see church with cemetery (Q29044875) is a thing but that seems to be mostly used with has use (P366), which doesn't help me here. --Reosarevok (talk) 16:24, 23 November 2017 (UTC)

Two options spring to mind: 1) Give the location of the church as being the churchyard (and that of the churchyard as being the town or village) or 2) use "has part". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:49, 23 November 2017 (UTC)

wikilinks on wikisource editions through wikidata

Tracked in Phabricator
Task T180303
Tracked in Phabricator
Task T180304
Tracked in Phabricator
Task T128173

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

Hi,

Wikisource editions of texts, are version, edition or translation (Q3331189) items that must be linked to the work item through edition or translation of (P629) (see Wikidata:WikiProject Books).

Unfortunately, this means that wikisource texts are not directly linked anymore to each other, thus breaking the link between the different languages.

I remember once reading here about a template that could restore these links, using the parent/work item to automatically link all editions, thus allowing to not only have access to other languages, but also other editions in same language.

Do you remember who talked about it ? It would be very important to have this, or even an automatic feature in wikisource, to automatically link texts to all other texts linked through edition or translation of (P629) to the same work item...

Thanks for your help !  – The preceding unsigned comment was added by Hsarrazin (talk • contribs) at 7. 10. 2017, 15:53‎ (UTC).

User:Hsarrazin: That sounds like s:sv:Modul:Sandlåda/Innocent bystander 3. You can see it in action at for example s:sv:Bibeln (Karl XII), and User:Innocent bystander is probably the one who best understands how it works. Peter Alberti (talk) 19:37, 16 November 2017 (UTC)
that's exactly what I was remembering... thanks a lot Peter Alberti :)
is it automatic or does it require a template or adding something on a page ?
@Tpt: - do you know it ? --Hsarrazin (talk) 13:09, 22 November 2017 (UTC)
It needs a template to be added to the page and a bit of site javascript. They use the standard header as the template, so it feels almost automatic, though. Peter Alberti (talk) 18:36, 23 November 2017 (UTC)
It's a very nice tool! Thank you for having mentioned it here. In order to be nicely implemented and don't require templates and JavaScript I believe we should implement such features as part of a MediaWiki extension. I am planning to work on it soon. Tpt (talk) 20:53, 23 November 2017 (UTC)
Adding a @Samwilson: to note.  — billinghurst sDrewth 05:16, 24 November 2017 (UTC)

Item default languages

Hello, how come when creating an item here, you get Bavarian (!) as one of the four standard languages (after English, German and French)? How does this order come about?--Herfrid (talk) 15:51, 23 November 2017 (UTC)

The languages that you get presented by default depend on your IP address. If you add the bable template you can specify which languages are shown to you. ChristianKl (✉) 16:05, 23 November 2017 (UTC)
@Herfrid: For "bable" [sic], see User:Pigsonthewing/Setup#Babel. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 16:51, 23 November 2017 (UTC)
Thank you – I didn't know that yet! I live in Germany – however, after inserting the bable template, I only get English now…! How come?--Herfrid (talk) 16:58, 23 November 2017 (UTC)
Maybe you need to clear you cache? Or perhaps the item only has a label in English? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:38, 23 November 2017 (UTC)
It seems to me that if you have the UI in English but tell the Wikidata with the bable template that you don't speak English but only German that's a case that the developers didn't forsee. If you add both German and English, the result should make more sense. ChristianKl (✉) 17:44, 23 November 2017 (UTC) S
@ChristianKl: What is UI?--Herfrid (talk) 22:16, 23 November 2017 (UTC)
UI=User Interface. The UI language is the language in which you see all the buttons, the header, and the left navigation bar. To change the UI language, click on   which you see next to your name in the header. --Pasleim (talk) 10:16, 24 November 2017 (UTC)

Please Merge Q925164 into Q15551906

moved from Wikidata:互助客棧

Anthemis tinctoria (Q925164); Cota tinctoria (Q15551906)

The name Anthemis tinctoria has been seen as synonym of Cota tinctoria.---Koala0090 (talk) 16:07, 21 November 2017 (UTC)

  WikiProject Taxonomy has more than 50 participants and couldn't be pinged. Please post on the WikiProject's talk page instead. as they can have more smart solutions rather than non-touchable merges. --Liuxinyu970226 (talk) 10:07, 24 November 2017 (UTC)

Finding authority control IDs from ISSNs

I'm adding an item for the French newspaper Le Courrier de l'Atlas (Q43534671). I've found an ISSN for it, and what appears to be some sort of BnF entry for it (see http://catalogue.bnf.fr/ark:/12148/cb42240793p ), but cannot as yet find a BnF ID for the newspaper. There are also various other ISSNs, each with its own record at the BnF website.

Is there a BnF ID for this paper, and if so, how can I find it?

In general, is there any other way of finding any kind of authority control record from an ISSN? -- The Anome (talk) 11:58, 24 November 2017 (UTC)

hello, the best way is to search directly in the Bnf catalog, using the issn. Here, the Bnf record would be "http://catalogue.bnf.fr/ark:/12148/cb409859173", then the id is "409859173" (last part of the ark, minus "cb").
by the way, beware the ISSN, the last 2 are respectively for "Marseille edition" and Ile-de-France edition (as stated in the bottom of the catalog record). --Hsarrazin (talk) 12:49, 24 November 2017 (UTC)

Are Tøyen (Q887532) and Tøyen (Q13668298) the same and can be merged?

Hi! Are Tøyen (Q887532) and Tøyen (Q13668298) the same and can be merged? //Mippzon (talk) 23:13, 25 November 2017 (UTC)

Thanks! //Mippzon (talk) 18:37, 26 November 2017 (UTC)
This section was archived on a request by: --
--- Jura 11:48, 30 November 2017 (UTC)

Pasleim - PLbot nonsense edits - municipality of Venezuela

Once more Mr. Pasleim ... https://www.wikidata.org/w/index.php?title=Q6063601&diff=210151732&oldid=210151723  – The preceding unsigned comment was added by 92.226.219.66 (talk • contribs) at 28. 11. 2017, 17:43‎ (UTC).

I don't think the project chat is the right place for such exposures. In case you want to have me blocked, start a discussion at WD:AN, in case you want to have me globally banned (like Tamawashi) follow the steps at m:Global bans. --Pasleim (talk) 17:54, 28 November 2017 (UTC)
User:Pasleim - "I don't think the project chat is the right place for such exposures" - you are a mass (mis) contributor and maybe items from other areas of interest have become victim of your mis-edits. 78.55.243.85 18:41, 28 November 2017 (UTC)
It isn't obvious to me why that is nonsense. Please change your tone otherwise it will be you who gets blocked. You might feel being important but you're just annoying. Matěj Suchánek (talk) 19:10, 28 November 2017 (UTC)
It's User:Tamawashi... --Succu (talk) 19:42, 28 November 2017 (UTC)
It is Tobias Conradi.--Ymblanter (talk) 21:22, 29 November 2017 (UTC)

166 duplicates - municipality of Venezuela

At

the 166 items from Q31961198 to Q32491773 probably all duplicates from cebwiki. The Swedish botmaster seems to not have linked the cebitems in WD 92.226.219.66 16:49, 28 November 2017 (UTC)

User:GerardM - can you help with merging? Spanish name is "Municipio X" , English name is "X Municipality". The correct state must be selected, example for ambiguous name : en:Arismendi Municipality. 77.180.4.196 09:53, 29 November 2017 (UTC)

I have another project tat is a time sinc. thanks, GerardM (talk) 12:05, 29 November 2017 (UTC)
This section was archived on a request by:
--- Jura 07:01, 30 November 2017 (UTC)

Fusionproblems

Can someone fusion en:Category:I'm a Celebrity...Get Me Out of Here! (Q8529833) with de:Kategorie:Ich bin ein Star – Holt mich hier raus! 92.76.113.106 20:26, 28 November 2017 (UTC)

They aren't the same category: the first is the category of I'm a Celebrity...Get Me Out of Here! (Q3113852) and the second is the category of Ich bin ein Star – Holt mich hier raus! (Q885167) --ValterVB (talk) 20:51, 28 November 2017 (UTC)
This section was archived on a request by: --
--- Jura 07:02, 30 November 2017 (UTC)

Merge please

"The link enwiki:Stanley Pargellis is already used by item Stanley Pargellis (Q44113800). You may remove it from Stanley Pargellis (Q44113800) if it does not belong there or merge the items if they are about the exact same topic." - "Stanley McCrory Pargellis (Q1268262): American librarian and military historian" - same person --Gerda Arendt (talk) 23:25, 29 November 2017 (UTC)

See Help:Merge. --
--- Jura 07:04, 30 November 2017 (UTC)
Thank you! --Gerda Arendt (talk) 09:13, 30 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 16:07, 30 November 2017 (UTC)

Merge request

Hi, can someone merge Q2061744 and Q19564394? Dunno how to do it. --Tolanor (talk) 23:30, 29 November 2017 (UTC)

  Done Special:Diff/600928968 --Liuxinyu970226 (talk) 02:33, 30 November 2017 (UTC)
This section was archived on a request by: --
--- Jura 07:02, 30 November 2017 (UTC)

Looking for a property

Hello, I'm working on en:Template:Ordination and am trying to make it fully Wikidata-enabled. However, I am looking or a Wikidata property that would be able to be used for the three parameters of "ordained deacon by," "ordained priest by," and "made a cardinal by." The single property would express who performed each of those actions. I would use the property as a sub-property of others. Does any such property currently exist? Thanks. Ergo Sum (talk) 06:00, 22 November 2017 (UTC)

How about appointed by (P748) as a qualifier for "hold position" deacon/priest/cardinal? ChristianKl (✉) 09:28, 22 November 2017 (UTC)
I've recently searched for such a property as well. There is consecrator (P1598) for bishops, but apparently no equivalent for other ranks. --Kam Solusar (talk) 09:44, 22 November 2017 (UTC)
@ChristianKl: That's great, I wasn't aware of that property. That works really well for cardinals. However, I might propose later an "ordained by" property for priests and deacons, though seeing how past religion properties have fared in the past, I don't know whether it will pass. Ergo Sum (talk) 03:14, 23 November 2017 (UTC)
We do have significant event (P793) -> ordination (Q7519600) with a range of qualifiers. Someone doing a recommendations page for clerics would be useful for CofE, RC and non-conf. Stepping through the stages of positions, with preferenced items would be much appreciated. ordination -> curacy -> priests (or whatever it is).  — billinghurst sDrewth 06:55, 23 November 2017 (UTC)
One of the core reasons why the properties haven't fared well in the past was because we try to create properties that are generally applicable. That means it's not enough to focus on the need for modeling Christian priests. If you create a proposal I would recommend to look at whether the properties can also be used in other religions. If ordaination in the Christian church is something qualitatively different, then it would be good to explain that in the property proposal and in the description of the property. ChristianKl (✉) 11:22, 23 November 2017 (UTC)
Ordination would work for rabbis, I think -- see en:Semikhah. The person that a rabbi was ordained by would be information worth including. Can you store/get everything you need for the template via the significant event (P793) structure? Jheald (talk) 23:16, 24 November 2017 (UTC)

Item about the concept of the QID

I was surprised to find that we had no item about our concept of a QID, so I created Wikidata Q identifier (Q43649390) (with "QID" as an alias). Please translate into your own languages.

I have already been able to ad one value for third-party formatter URL (P3303) - are there others? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:55, 25 November 2017 (UTC)

Not sure about the inclusion of several properties that are mostly used on properties and now creating constraint violations. Actions like these seems like playing the constraints system. Sjoerd de Bruin (talk) 16:44, 25 November 2017 (UTC)
That's why we have the ability to add exceptions to constraints. Describing the use of such a feature as "playing the system" is unnecessarily pejorative. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:05, 25 November 2017 (UTC)
It is playing the system. Where do we stop adding exceptions? Who would use this information this way anyway? People expect them on properties, the usage case atm is zero. Sjoerd de Bruin (talk) 18:13, 25 November 2017 (UTC)

Are Tøyen (Q887532) and Tøyen (Q13668298) the same and can be merged?

Hi! Are Tøyen (Q887532) and Tøyen (Q13668298) the same and can be merged? //Mippzon (talk) 23:13, 25 November 2017 (UTC)

Thanks! //Mippzon (talk) 18:37, 26 November 2017 (UTC)
This section was archived on a request by: --
--- Jura 11:48, 30 November 2017 (UTC)

Fusionproblems

Can someone fusion en:Category:I'm a Celebrity...Get Me Out of Here! (Q8529833) with de:Kategorie:Ich bin ein Star – Holt mich hier raus! 92.76.113.106 20:26, 28 November 2017 (UTC)

They aren't the same category: the first is the category of I'm a Celebrity...Get Me Out of Here! (Q3113852) and the second is the category of Ich bin ein Star – Holt mich hier raus! (Q885167) --ValterVB (talk) 20:51, 28 November 2017 (UTC)
This section was archived on a request by: --
--- Jura 07:02, 30 November 2017 (UTC)

Merge please

"The link enwiki:Stanley Pargellis is already used by item Stanley Pargellis (Q44113800). You may remove it from Stanley Pargellis (Q44113800) if it does not belong there or merge the items if they are about the exact same topic." - "Stanley McCrory Pargellis (Q1268262): American librarian and military historian" - same person --Gerda Arendt (talk) 23:25, 29 November 2017 (UTC)

See Help:Merge. --
--- Jura 07:04, 30 November 2017 (UTC)
Thank you! --Gerda Arendt (talk) 09:13, 30 November 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 16:07, 30 November 2017 (UTC)

Merge request

Hi, can someone merge Q2061744 and Q19564394? Dunno how to do it. --Tolanor (talk) 23:30, 29 November 2017 (UTC)

  Done Special:Diff/600928968 --Liuxinyu970226 (talk) 02:33, 30 November 2017 (UTC)
This section was archived on a request by: --
--- Jura 07:02, 30 November 2017 (UTC)

Merging problem

Hello, I wanted to merge ordinance (Q2703920) to implementing regulation (Q3704604) as the entries within these items seem to cover more or less the same thing. However, there is one single conflict that causes blocking: Q3704604 with ja:行政委任立法 vs. Q2703920 with ja:命令 (法規). Does anybody have an idea how to resolve that conflict?--Herfrid (talk) 17:25, 23 November 2017 (UTC)

Merging them in ja.wikipedia. If these concepts are not "exactly" the same, probably they shoudn't be merged. strakhov (talk) 17:35, 23 November 2017 (UTC)
@Strakhov: Thank you for commenting. However, I'm afraid I'm not the right person to do that. On the other hand, it would be a shame if the merge would fail really just because of this single conflict, wouldn't it? I really don't know what to do…--Herfrid (talk) 17:42, 23 November 2017 (UTC)
You can do nothing. It often works. Somehow. :) strakhov (talk) 17:45, 23 November 2017 (UTC)
No, it's no shame if the conflict prevents you from merging two distinct concepts. Two concepts shouldn't be merged because they are more or less the same thing but only when they actually are the same thing. ChristianKl (✉) 17:49, 23 November 2017 (UTC)
But there only appears to be a problem with ja-Wiki – the other articles I managed to move to the apporopriate items. And I would still say that the entries within the two said items strongly match.--Herfrid (talk) 22:59, 23 November 2017 (UTC)
Let's see. Even if both articles in Japanese could be merged (I don't know that, I don't know Japanese), the items may need to remain independent and 'unmerged' (if they were not exactly the same), unless one of them is deemed as non notable (in that case it should be deleted). And that's not a problem for Wikidata. Our mission is far wider than providing interwiki links to wikipedias. strakhov (talk) 00:08, 24 November 2017 (UTC)
A strong overlap is not the same thing as two topics being the same. Whether or not the two things are the same is a question that needs expertise in international law to answer. I don't have that. If you have that expertise, feel free to make the argument but if you don't have the expertise to judge in a domain whether two things are the same, don't merge.
I outsourced the question to stackexchange, which is generally a good place to get expert opinions. ChristianKl (✉) 12:47, 24 November 2017 (UTC)
It looks like ja:命令_(法規) does mention the case in Japan, but ja:行政委任立法 does not, hence I   Oppose merging. --Liuxinyu970226 (talk) 10:04, 24 November 2017 (UTC)
@Liuxinyu970226: Thank you very much for forwarding the matter, though in this case, I actually do not think it is chiefly a question of labelling (as the commenter wrote, it had to read "ordinance" instead of "ordonnance", by the way), but whether the entries within these items really deal with different things or not, and as I can only assess by using Google translator this does not seem to be the case – not even with regard to the two Japanese articles, as ja:行政委任立法 is simply a one-to-one equivalent, i. e. a translation, of w:Statutory instrument, which simply focuses on the anglophone world (which could change, however, as soon as information about non-anglophone countries is included!). Hence, I still do not see a convincing rationale for having two separate items here.--Herfrid (talk) 22:32, 25 November 2017 (UTC)

Merging Q18200543 with Q142804

Dear users,

I suggest the merging of tanker aircraft (Q18200543) currently associated to 3 languages (el, fr & nl) with aerial refueling (Q142804) that has 30 associated languages but which does not includes el & fr (nl is present in both of them). If you compare the (fr) page of Q18200543, with the (en) page of Q142804, you will see that the subject is in reality the same.

Best regards --Ndiver (talk) 17:15, 25 November 2017 (UTC)

  Oppose Whether or not the subject of the Wikipedia pages overlaps is irrelevant to the question of whether the two subjects are the same. "Tanker aircraft" is a type of aircraft while arieal refueling is the process where a tanker aircraft might be used the refuel another aircraft. ChristianKl (✉) 17:18, 25 November 2017 (UTC)
  Oppose I agree with ChristianKI. Regardless of wikilinks' final destination and how or where they end up, items should not be merged because they do not represent by any means the same concept. strakhov (talk) 18:07, 25 November 2017 (UTC)
I'm trying to solve a problem noticed on the French page. "Aerial refueling" (ravitaillement en vol) redirects to "tanker aircraft" (avion ravitailleur), but you can't go from it to the pages in other languages referring to aerial refueling, because it is associated to the 2 pages (in Greek and Dutch) specifically speaking about tanker aircraft. See : https://fr.wikipedia.org/w/index.php?title=Ravitaillement_En_Vol&redirect=no What are the possibilities to solve this problem if you do not want to merge? Should we edit the Page Information of the French page to associate it with Q142804 rather than Q18200543 ? --Ndiver (talk) 23:21, 25 November 2017 (UTC)
There is also tanker aircraft (Q3261888). This one seems, a priori (Q17315578), easier to merge with tanker aircraft (Q18200543), but if I were you, I'd wait for a more nuanced opinion. strakhov (talk) 23:47, 25 November 2017 (UTC)
I've done this, but I'm not aware of "manual" interwikis' status in fr.wikipedia, so it may not be a long term solution. strakhov (talk) 23:55, 25 November 2017 (UTC)
It was a 8-minute-solution. Il est en charge, good luck with that. strakhov (talk) 00:02, 26 November 2017 (UTC)
I merged tanker aircraft (Q3261888) with tanker aircraft (Q18200543) as both are about tanker aircraft. As far as the interwiki links go, I personally think that redirect links are the way to solve the issue. Currently it's not easy to set them with the Wikidata UI but with https://www.wikidata.org/wiki/Wikidata:Requests_for_comment/Allow_the_creation_of_links_to_redirects_in_Wikidata it hopefully will get easier in the future. ChristianKl (✉) 00:05, 26 November 2017 (UTC)

Calling Special:Nearby with coordinates

Can Special:Nearby be called with coordinate parameters? I think I have seen this somewhere but cannot find any documentation about it. --Daniel Mietchen (talk) 19:16, 30 November 2017 (UTC)

It was mentioned in the birthday talk by @Aude:, and from the video I could grasp the URL for the nearby of an item: Special:Nearby#/page/Q9188. Ahoerstemeier (talk) 21:07, 30 November 2017 (UTC)
Neat — thanks! --Daniel Mietchen (talk) 01:44, 1 December 2017 (UTC)
This section was archived on a request by: Matěj Suchánek (talk) 19:44, 2 December 2017 (UTC)

Historical humans who also appear in fiction

Bear with me. This is long. I am struggling with modeling some historical figures who are best known in their fictional or legendary forms. Yesterday I ran across an item new to me, Personnage historique in frwiki, to which I have added an EN label and description "historical character (Q3375731): character in works of fiction inspired by an actual person in history, often heavily romanticized". The linked frwiki article specifically calls out Roland (Q207535) and Cyrano de Bergerac (Q213614) as falling into this category.

Question: Can we model such persons as both human (Q5) and historical character (Q3375731)? ("Fictionalized human" currently has no instances.)

Some observations:

  • Most Wikipedias have a single article for the historical and fictional character, so for interwiki purposes a single item in Wikidata is best.
  • In a few cases Wikipedias have separate articles, especially when the person is better known by a variant name in one context: the Arthurian character Ywain (Q20751922) son of Urien <inspired by>Owain mab Urien (Q1486268).
  • My inner ontologist, who is very strict, says the legendary/fictional character who is based on a historical person should always have a separate item, but that causes problems in linking to Wikipedias which typically lead off with the historical figure and go on discuss the legends which may or may not be more notable than the actual figure.

My immediate example is Yvain's father, Urien Rheged (Q1263208) king of Rheged. In fiction, Urien is married to Morgan le Fay. In history, Urien was married to someone else. If Urien is both "human" and "fictionalized human", then we might say <spouse> "Morgan le Fay" <applies to part> "fictionalized human". (Is there a better way to qualify this?)

This solution would also allow us to say Charlemagne (Q3044) <present in work> "Matter of France (Q2487521): body of literature and legendary material associated with the history of France, in particular involving Charlemagne and his associates", since "fictionalized human" now that it has surfaced should be a valid subject for present in work (P1441). We could also then say Matter of France (Q2487521) has <characters> Charlemagne, Roland, etc., which would cause constraint violations now since they are <instance of> only human (Q5) today.

I see a spectrum of characters in Wikidata from "definitely real" to "definitely fictional" as follows:

I believe we should maintain a distinction between historical character (Q3375731) and legendary figure (Q13002315). A "fictionalized human" is definitely historical, but there is some question about "legendary" humans.

I'd love to know how others see this problem. Thanks, all. - PKM (talk) 21:06, 20 November 2017 (UTC)

I think that you have to separate the human and the fictionalised human, with something like "based on" the (real) human as the relationship. You have to separate the real and fiction. And I not sure that you would need the relationship expressed in the H -> FH direction. That said, then is each different authors' impression of that fictionalised character the same version or a different version of the FH? Then how would you represent Thomas Cromwell (Q294435) in Hilary Mantel (Q465700)'s Wolf Hall (Q2657795).  — billinghurst sDrewth 03:48, 21 November 2017 (UTC)
And how to link the same wikipedia-article to both items? The main purpose of WD is to host interwikilinking, how to achieve this? Sänger (talk) 05:33, 21 November 2017 (UTC)
@Sänger: No, that's not the main purpose of WikiData. Wikidata items are suosed to map to concepts that are out there. If that doesn't produce enough Interwiki links, you can set them manually or create redirects. ChristianKl (✉) 11:04, 27 November 2017 (UTC)
Interesting question. In my opinion items being an instance of human (Q5) should ideally contain only informations backed up by historical evidences. I think historical character (Q3375731) can be used as a class for humans where scholars widely agree that they existed but where the available information is mainly from unreliable sources (legends or unreliable historical accounts). For those characters it should actually suffice to just make them an instance of historical character (Q3375731) (without making them additionally an instance of human (Q5)). I think that a second item being an instance of human (Q5) (like with historical Jesus (Q51666)) is only needed if there are informations backed up by historical evidence and one would want to model the historical person "behind" the legendary one explicitly.
I would not use historical character (Q3375731) for characters in works of fiction based on a historical person. So Queenie from Blackadder should not be an instance of historical character (Q3375731), but of fictional human (Q15632617). - Valentina.Anitnelav (talk) 08:27, 21 November 2017 (UTC)
As to the use of historical character (Q3375731) to deal with constraint violations: I don't think that this is necessary. I see no advantage of adding historical character (Q3375731) to humans over just allowing non-fictional entities as subjects of present in work (P1441). characters (P674) already allows humans as values. - Valentina.Anitnelav (talk) 11:05, 21 November 2017 (UTC)
historical Jesus (Q51666): Jesus of Nazareth as a historical person. Ok, what is historical person? --Fractaler (talk) 12:46, 22 November 2017 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ Thanks all for the comments. I have added human (Q5) as a subject type for present in work (P1441) based on supportive discussions here and at that property's talk page back to 2014, and for harmony with characters (P674). In the specific case of Urien and his fictional wife, I have created a separate item for the fictional Uriens (Q43401974). - PKM (talk) 20:51, 27 November 2017 (UTC)

A new constraint

Hello. I think we don't have yet − and need − a new constraint establishing that another property should not have the same value as the one we're dealing with on the items where they might be both used. This would be useful for some properties linked by subproperty of (P1647) since we often don't want the main property to have a given value if its subproperty – more specific – is already telling us the same thing. Can someone set this up? Thierry Caro (talk) 09:50, 27 November 2017 (UTC)

In the meantime you can do it with {{Complex constraint}} --Edgars2007 (talk) 16:05, 27 November 2017 (UTC)

Datatype of software version identifier (P348)

Datatype of software version identifier (P348) is currently string. Shouldn't we use 'Tabular data type' instead, since it changes rapidly and a lot of values may be generated (in short spans of time), especially for commonly used softwares? Take for example, software version identifier (P348) values of MediaWiki (Q83) are already occupying a major portion of the page. Population, another similar example (some cities do release yearly population numbers) can be represented by population (P1082) and tabular population (P4179), the latter to handle vast amount of data changes. John Samuel 20:14, 27 November 2017 (UTC)

we wouldn't change the datatype of the existing property, but I think it would be fine to propose a new property for this, it sounds like a good idea. ArthurPSmith (talk) 21:24, 27 November 2017 (UTC)
I agree to create a new property for that purpose. Pamputt (talk) 22:01, 27 November 2017 (UTC)

Nepali

May we know why we can not see "Nepali" Language while selecting the language?  – The preceding unsigned comment was added by 103.10.28.52 (talk • contribs) at 16:00, 16 November 2017‎ (UTC).

This seems works for me, by typing Nepali I actually got नेपाली, can you please tell me which OS are you using? --Liuxinyu970226 (talk) 12:33, 28 November 2017 (UTC)

Wikinews article sitelinks

Recently I have often seen Wikinews article sitelinks connected to items about Wikipedia articles. However, according to Wikidata:Wikinews/Development#Interproject links it is pretty clear that Wikinews categories correspond to Wikipedia articles (somewhat similar to the situation with Commons), and Wikinews article sitelinks are always grouped into a separate item (with instance of (P31) Wikinews article (Q17633526)).

Since there were a couple of reverts after I separated Wikinews article sitelinks from Wikipedia article sitelinks, I would like to ask whether Wikidata:Wikinews/Development#Interproject links is still up to date or not. Intuitiveley it describes a reasonable solution, but maybe in practice there are problems which I am not aware of, and which have not been codified in the linked page.

@De Wikischim, Ymnes: FYI. Do we have a Wikinews taskforce? —MisterSynergy (talk) 09:19, 20 November 2017 (UTC)

This is relevant if a separate category on a particular subjects exists on Wikinews. In this specific case however, there is only a Wikinews article. So this is automatically the only corresponding page on Wikinews. De Wikischim (talk) 11:13, 20 November 2017 (UTC)
I personally think wikinews article (news) should group themselves only, and Wikinews categories are more for concept (the real article for WP) themselves. — regards, Revi 08:46, 21 November 2017 (UTC)
I see problem with wikinews categories. This linking WN-cat vs. other-article was made because there was no arbitrary acces i that time. But now we have schizophrenic situation: some wikinews categories are connected with artilces and some with ather categories. And I suppose, that majority of wn-cat which are now connected with articles should be linked with smae-name categories on other projects and then linked via category's main topic (P301). JAn Dudík (talk) 13:47, 28 November 2017 (UTC)
Are you sure that the lack of arbitrary access was responsible for the current solution? In my opinion the problem is that you can only assign one Wikinews sitelink (per Wikinews project) to a Wikidata item, although there can easily be more than one Wikinews article related to this item. This is somewhat similar to the situation with Commons: you can have many files per Wikidata entity, thus we link categories which contain the files as sitelinks, not the files themselves. —MisterSynergy (talk) 13:52, 28 November 2017 (UTC)

Instance of AND subclass of

I just approached the use of Wikidata, so sorry for unclear points: working with Linux related data I found "Debian" (Q7593). Debian (a Linux Distribuition) is a real object and so I was surprised to find "Instance of" (P31) AND "subclass of" (P279) together among the declarations. Being Debian a real object I don't think "subclass of" should be part of the element. Thanks for your comments and clarifications --FabC (talk) 12:23, 24 November 2017 (UTC)

The instance of Debian that runs on Alice computer is not the same instance that runs on Bob's computer. subclass of (P279) seems more appropriate to me. ChristianKl (✉) 12:54, 24 November 2017 (UTC)
@FabC: Debian is not a "real object" in the sense of something that has a physical manifestation at a particular time and place. It represents many physical instances, and that collection can be thought of as a class, which can be subdivided into further subclasses for example by version or underlying hardware or some other aspect of implementation. So these properties look fine to me on Debian (Q7593). ArthurPSmith (talk) 14:11, 24 November 2017 (UTC)
Thanks for the clarifications, now I understand and agree on the use on the subclass of (P279). It is still confusing me the presence of instance of (P31) among the declarations: Debian (Q7593) is an instance of (P31) of operating system (Q9135), ... . --FabC (talk) 18:41, 24 November 2017 (UTC)
operating system (Q9135) perhaps should be declared as a metaclass (class whose instances are themselves classes); hopefully it's being used consistently in this fashion. We have some areas of Wikidata where there's considerable confusion on this, but these statements seem to be justified to me here. ArthurPSmith (talk) 22:57, 24 November 2017 (UTC)
Wouldn't that same argument apply to nearly every work (book, movie, game) then, which all can have diffêrent versions, editions and physical instances? --Kam Solusar (talk) 06:39, 25 November 2017 (UTC)
+as temporal entity (Q26907166): if we are dealing with temporal entity (Q26907166), then it means that we are dealing with a group of objects. For example, we have object1, then: object1 in time 0, object1 in time 1, object1 in time 2 ..., object1 in time now, object1 in time now+1, object1 in time now+2, etc. (beginning (Q529711), initial version of object1 -> current version of object1 -> prediction version of object1). --Fractaler (talk) 12:23, 27 November 2017 (UTC)
@Kam Solusar: Yes; though there are more direct properties for some of those relations, like edition or translation of (P629). book (Q571) is also essentially a metaclass in wikidata, as specific published works are generally stated to be instance of (P31) book (Q571). ArthurPSmith (talk) 15:32, 27 November 2017 (UTC)

@ChristianKl, ArthurPSmith, FabC: - what about ZIP code (Q136208)? The user that did it (User:D1gggg) is blocked. 77.180.160.124 16:11, 28 November 2017 (UTC)

How to remove wrong sitelinks

In plwiki I go to https://pl.wikipedia.org/wiki/Kolokwium which is **not** a disambig

I click the link to EnglishWP and I get a disambig: to https://en.wikipedia.org/wiki/Colloquium

**None of the meanings listed in this disambig translate to Kolokwium**

So I reason that the link is wrong and I go to https://www.wikidata.org/wiki/Q1115343#sitelinks-wikipedia to remove it

I try to remove it, but I get a warning that unless the destination page was moved or deleted I may not remove sitelinks

As a result, may I not remove wrong sitelinks, or do I misunderstand something??

What should I do in this case?

EDIT: I suppose that if anything, https://pl.wikipedia.org/wiki/Kolokwium should link to https://en.wikipedia.org/wiki/Test_(assessment) but I'm not 100% sure; also Kolokwium has a narrower meaning than Test (assessment); as Kolokwium is an academic test, but lesser than an exam – The preceding unsigned comment was added by 89.77.194.225 (talk • contribs) at 14:35, 27 November 2017‎ (UTC).

Not an easy task, as not only disambigs, but also articles about multiple different concepts were linked to one item. I move "exam" links to exam concept item (colloquium (Q4228029)), but still some articles with another concept are linked to the Colloquium (Q1115343). Feel free to clean-up them and to create another items, sill there is a problems with articles covering multiple concepts.
You can remove links as you wish, I am not able to understand what kind of warning you got.--Jklamo (talk) 15:53, 27 November 2017 (UTC)

@Jklamo: "Warning: The action you are about to take will remove a sitelink from this item. Sitelinks should only be removed if the page in question has been deleted, or if that link is being moved into another item. If you are trying to do neither of these, please do not submit this edit again."  – The preceding unsigned comment was added by 89.77.194.225 (talk • contribs) at 17:27, 27 November 2017‎ (UTC).

Simply register account (or login to existing one) to remove this limitation.--Jklamo (talk) 08:41, 28 November 2017 (UTC)

Problem in (Add links)

Tracked in Phabricator
Task T181501

Hello all, one of ar.wikipedia users try many times to create a links between articles but she faced the same Error every time. She tried in many ways, e.g by wikidata itself, from en.wiki/arwiki..etc. I asked her to take a picture for the error message when she try to make a link. So what's your opinion? --Alaa :)..! 21:32, 27 November 2017 (UTC)

I think we had this problem around two weeks ago. Matěj Suchánek (talk) 13:40, 28 November 2017 (UTC)
@Matěj Suchánek: can you share what you know about these problem on T181501 ticket on Phabricator. Thanks --Alaa :)..! 16:05, 28 November 2017 (UTC)

described by source (P1343) and use of stated in (P248) as qualifier rather than reference

There has been a broad use of stated in (P248) to qualify a listed article (biographical, newspaper, ...) of a described source, (see example Augustus Hare (Q4821445)). This is flagging as a constraint violation (Wikidata:Database reports/Constraint violations/P248) which states that it should be used as a reference only. Is this a case of trying to modify the constraint violation, or are we missing a more appropriate property to point to the listed article? Thoughts for resolution?

As a note, the use of P248 replicates the practice of the qualification with section, verse, paragraph, or clause (P958) — rather than addition as reference — to directly specify the article component. As the use is to drill-down within a work, rather than a typical citation/reference, changing these to be in the reference section seems incorrect.  — billinghurst sDrewth 00:41, 14 November 2017 (UTC)

Well that's certainly a mess.
Perhaps described by source (P1343) could just point to the article directly? Or, if the article must be separate in a qualifier, a different property should be used. Maybe applies to part (P518) or perhaps a new property should be created?
Mangling the scope of such a widely-used property to include something not at all similar seems like a really bad idea. --Yair rand (talk) 10:09, 14 November 2017 (UTC)
@Yair rand:, okay, so we are needing to change it. The only other close 'active-link' property that I saw was statement is subject of (P805) as it is also called article, etc. and is to be as a qualifier. Applies to part is possible, though it doesn't have a natural ring as something to use.  — billinghurst sDrewth 21:34, 14 November 2017 (UTC)
@Charles Matthews, Jheald: do you have opinions here as you have been involved with some of the "described by source" work.  — billinghurst sDrewth 22:58, 19 November 2017 (UTC)

William Inglis (Q43529218), birth and death dates, shows (roughly speaking) the style I'd like. But there is a problem, namely that section, verse, paragraph, or clause (P958) is of string type. So, I'd like to see a variant in which the entry could be a link. Charles Matthews (talk) 11:33, 24 November 2017 (UTC)

In my opinion, statement is subject of (P805) should be used instead of stated in (P248). Note also that P805 has dedicated article as alias. --Pasleim (talk) 10:02, 29 November 2017 (UTC)

Question, notability, inclusion

La Novela Teatral (Q32860631) (article) was a theatrical collection published between 1916 and 1925. One issue -> one play. Does anyone foresee a problem with storing this data (currently stored as a non-structured table in es.wikipedia) here in Wikidata (as items for each edition/issue and probably its associated work, authors, ...) ...instead of using Wikipedia (wrt notability issues and such)? If needed there is a book indexing the collection, but I'd say the content, as bibliographic data, is mostly auto-source-able. strakhov (talk) 13:15, 27 November 2017 (UTC)

I see no problem with having it inside Wikidata. ChristianKl (✉) 15:31, 27 November 2017 (UTC)
Thanks! strakhov (talk) 20:32, 29 November 2017 (UTC)

Birds described as reptiles

I have discovered that we have a number of items, being protonym (Q14192851) of items about birds, but with descriptions saying they are reptiles, and with the same common name, which does not aid disambiguation. These include:

but I suspect that that list is far from complete. How should these be processed? And how can we find other such cases? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:36, 15 November 2017 (UTC)

I prefer   Support merging the second pair, as Q25652651 doesn't give me full things to be independent. --Liuxinyu970226 (talk) 13:25, 15 November 2017 (UTC)
Both represent different concepts, Liuxinyu970226. --Succu (talk) 21:56, 15 November 2017 (UTC)
The concept of the Elegant Tern being a reptile is a new one to me; doubly so as the parent taxon is described as "genus of birds". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:19, 20 November 2017 (UTC)
Maybe you'll should ask User:Mr. Ibrahem why his bot did this edit?! --Succu (talk) 22:04, 20 November 2017 (UTC)
Apart from the error by Mr.Ibrahembot, this is a good example why common names should not be used as labels: common names are not unique identifiers. - Brya (talk) 11:41, 27 November 2017 (UTC)

Does anyone have a proposed solution for this? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:27, 30 November 2017 (UTC)

Identfier types are subclasses, not instances, of identifers

unique identifier (Q6545185) is defined as "identifier which is the only one identifying an object within a system", and is subclass of (P279) of identifier (Q853614).

"286", for example, is an instance of a unique identifier, in the ISO 15924 numeric code (Q19842566) system. Why, then do we say that Q4928186 is an "instance" and not "subclass" of Q6545185? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:44, 26 November 2017 (UTC)

We had this discussion with an IP user (blocked?) on Wikidata:Identifiers if I remember the early history right. My view is "286" is not an identifier at all, it is a number. In a given context it can identify something, but in itself it does not. identifier (Q853614) really refers to a system for identifying things; some of these systems provide unique identification and so qualify as instances of unique identifier (Q6545185). The individual values assigned to identify items are not generally of interest to wikidata I would think, other than as values for the associated properties. ArthurPSmith (talk) 15:25, 27 November 2017 (UTC)
Financial Instrument Global Identifier (Q4928186) is an individual identifier system, that's why it is an instance of unique identifier (Q6545185). As explained at Wikidata:Identifiers#Identifier_items we have some items about individual identifiers or identifier values, like "286" with context, e.g. 10048 (Q4546087) and .me (Q43673). The context of each identifier value is an identifier system, e.g. ZIP code (Q136208) and country code top-level domain (Q42032). Identifier systems can further be organized in classes such as top-level domain (Q14296), domain name (Q32635), and unique identifier (Q6545185). The problem is: clean separation of items about identifier systems and items about the class of individual identifiers of each of the systems would create an unmanageable parallel structure of items. -- JakobVoss (talk) 20:55, 28 November 2017 (UTC)
unique identifier (Q6545185) != "identifier system" (albeit you added that as an alias, in August). Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:35, 30 November 2017 (UTC)
The string "286" is not an identifier, but "286" as a value in the ISO 15924 numeric code (Q19842566) system is. A system that provide unique identification is not a instances of unique identifier (Q6545185), it is an instance of an "identifier system", for which we don't as yet, have an item. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 11:35, 30 November 2017 (UTC)
@Pigsonthewing: But we don't have a wikidata item for "286" as a value in the ISO 15924 numeric code (Q19842566) system. Precisely what problem are you trying to address here? ArthurPSmith (talk) 15:14, 30 November 2017 (UTC)
We don't have items about lots of things; that's not a reason to mis-describe the items we do have; and that we do so is the problem described in my OP. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:29, 30 November 2017 (UTC)
Is it a misdescription? I think it's just an ambiguity in English. What would you say ORCID is? Or ISO 3166-1 alpha-3 or any of our other favorite "identifiers"? "Unique" has no meaning outside of a collection, so "unique identifier" surely refers to a collection of item-id relationships, not to individual id's. The individual id's certainly have no guarantee of uniqueness without the context of the collection/system. Although we haven't been entirely consistent about this, MOST of the items in wikidata that are unique identifiers (or "identifier systems" if you prefer) have been stated as instance of (P31) unique identifier (Q6545185). Also if you think of "subclass" as subsetting, which is part of the meaning, then if you insist on doing things your way the superclass (unique identifier (Q6545185)) contains all the identifiers from all the different identifier systems, and therefore is surely no longer "unique"! ArthurPSmith (talk) 16:59, 30 November 2017 (UTC)

Pages moved over linked pages

When a page is moved over a page that is already linked on Wikidata it doesn't appear to update so the new page is at the old page's item. See Q22190297 which contains the DAB page (at Ceb) and the article moved over the DAB (at Sv). Same with Q5681778. Is there a way of fixing this so that the link is automatically deleted from the old item and moved to the new one? Lucywood (talk) 18:54, 27 November 2017 (UTC)

It just randomly fails at moments. Sjoerd de Bruin (talk) 20:07, 27 November 2017 (UTC)
I think it is more like every time, presumably because before a few years ago admins had to delete a page then move, now they can just move over. Lucywood (talk) 11:47, 28 November 2017 (UTC)

could we remove the constrain of "target" (P533) to be "violent crime" (Q491040)?

From P533 discussion page, it appears that we are several to feel the need to express that something has something else for target, without it involving a violent crime (Q491040). Example: foodsaving (Q43888439) targets food waste (Q18919), could we extend target (P533) so that foodsaving (Q43888439) → target (P533) → food waste (Q18919) isn't a constrain violation? Any opposition? Or should we create an alternative property? -- Maxlath (talk) 19:55, 27 November 2017 (UTC)

I don't think that foodsaving targets foodwaste in a similar way as target (P533), so I'm against removing the constraint here. It makes more sense to think hard about what the relationship between the two actually happens to be and try to create a property that matches it. ChristianKl (✉) 23:08, 27 November 2017 (UTC)
I would agree with ChristianKl that changing it would be inconsistent with the property proposal. One of those same word, different meanings scenarios; and I would think that in other languages that the label may actually be different. The target for this property is specific and physical; the target that you are looking to achieve, is a goal. How does has goal (P3712) suit your needs?  — billinghurst sDrewth 22:03, 28 November 2017 (UTC)
@billinghurst: has goal (P3712) is pretty close, but it explicitly mentions that the statement object should be desired, so maybe we could have an equivalent negative property? -- Maxlath (talk) 15:27, 29 November 2017 (UTC)
@Maxlath: I don't see that a negative direction for a target is contrary to the property. Objectives can always be lower or higher than the present. Though what we may need to have is the expression of the starting point, or a level at a point in time on the way to the target/objective.  — billinghurst sDrewth 00:36, 30 November 2017 (UTC)
Maxlath point is that the objective of "food saving" isn't "food waste" but the prevention of "food waste". I'm open for discussing a new property for this purpose. ChristianKl (✉) 12:12, 30 November 2017 (UTC)

Item merge problem

Please see the problem described at Talk:Q26783785. I was unable to proceed from that point. -- Deborahjay (talk) 11:26, 28 November 2017 (UTC)

Can not be merged because of existing two articles in SVWP. - Kareyac (talk) 11:49, 28 November 2017 (UTC)
It seems the two items refer to two different GeoNames IDs http://www.geonames.org/1264230/majuli-island.html and http://www.geonames.org/1264231/majuli.html. One is about that actual island as an island while the other talks about the populated place. sv.wiki has bot created versions for both and would need to merge their articles before we can merge ours. ChristianKl (✉) 11:54, 28 November 2017 (UTC)

User:ChristianKl - that's nonsense. Why would one merge two different real world objects based on what svwiki does? 78.55.243.85 18:43, 28 November 2017 (UTC)

What exactly do you mean with the phrase "different real world object"? Would you say that the current US president and the person of Donald Trump are two different real world objects? ChristianKl (✉) 18:51, 28 November 2017 (UTC)
I've requested the two svWP articles to be merged. –Tommy Kronkvist (talk), 18:58, 28 November 2017 (UTC).
Even if SV wiki merges their articles, I oppose merging a geographical object (island) with the town or city which sits on it. The town is located on the island, and perhaps the island is inside the town, but they are not the same class of thing. - PKM (talk) 02:00, 29 November 2017 (UTC)
@PKM: There's a lot of nonsense duplications on cebwiki and svwiki which are created by Lsjbot (Q17430942). --Liuxinyu970226 (talk) 11:46, 29 November 2017 (UTC)
I am aware of that problem - I've been cleaning up at lot of Cebuano GEONAMES items while working on the NARA locations dataset. If there are two items for the same geographical object then of course they should be merged. But what I am saying is that a geographical object should not be merged with municipal authority even if they are discussed in one article in Wikipedia (which happens in EWN wiki as well. - PKM (talk) 19:51, 29 November 2017 (UTC)

┌────────────────────────────────────────────────────────────────────────────────────────────────────┘ As for svWP both of the two pages regarded the island (not the town) and are now merged into one. It is linked to Q277552 here at Wikidata. –Tommy Kronkvist (talk), 01:30, 30 November 2017 (UTC). "Items are for" can be read different ways. On Wikidata both were tagged as island (or river island) and not as towns. It's just that one of the GeoNames entries refers to a populated area. ChristianKl (✉) 12:16, 30 November 2017 (UTC)

Compiling some pages for general information on open data, data literacy, database rights and data reusability

Dear all

I am compiling resources for a Wikidata page that would provide general background information on open data, database rights and best practices for data producers to make their data more accessible and reusable.

I have found when talking to both organisations and individuals about Wikimedia that there is a lower degree of general data literacy than for instance copyright for images. My aim with this resource is to improve data producers understanding to where they can make informed decisions about data licensing, publishing formats etc.

My guess is that a lot of this information is already exists, produced by several organisations and a lot of the work will be to simply link out to existing resources. I'm also aware that there is some ambiguity around data rights and also some missunderstandings and missinformation.

If you know of any good resources or think something is important to include please do add it in, please feel free to braindump, I have time to do copy editing etc

https://www.wikidata.org/wiki/User:John_Cummings/Database_rights_and_best_practices

Thanks very much

--John Cummings (talk) 15:04, 28 November 2017 (UTC)

One of the main problems is that the law around databases is a lot less tested. There are plenty of court cases about image rights but no court case about European suis genesis databases rights that would tell you their limits. ChristianKl (✉) 11:06, 30 November 2017 (UTC)
Hi ChristianKl, yes very much so, its the main unknown I think for writing this. Do you know of any external resources that explain this well? I would like to include something about this ambiguity but link out to something much more in depth. --John Cummings (talk) 13:22, 30 November 2017 (UTC)

What properties to use for fictional relations between fictional objects?

Hi,
I want to to add data about star citizen (which is a game).
What property do i choose when e.g. there is an spaceship which is made by a manufacturer?
Should I use P176 "manufacturer" even though it is just a fictional product and manufacturer?
Thanks for your help.

Yes, P176 is right. The spaceship and the manufacturer should be marked as fictional via P31 <fictional ...>- Valentina.Anitnelav (talk) 14:42, 30 November 2017 (UTC)
Thanks for your answer,
could you check Q44180664 if I understood you correctly? The P31 is under "Star Citizen ship manufacturer" Q44180928 which is also linked with P31.
I checked and answered at your talk page. - Valentina.Anitnelav (talk) 15:48, 30 November 2017 (UTC)

Creation of this project and its effects on sister projects

After reading negative feedback about and drama involving Wikidata, I have wondered whether the communities agreed on creating another sister project Wikidata before it debuted in 2012. I searched for consensus/votes within revisions of meta:Wikidata but found nothing. What are pros and cons of this project? How have Wikidata and its centralization affected other existing projects, like Commons and Wikipedia? George Ho (talk) 10:23, 28 November 2017 (UTC); partially struck, 03:44, 30 November 2017 (UTC)

(Edit) Newer questions: (If necessary) Before Wikidata debuted with the Board's authorization, were analyses and reviews from the communities on the Wikidata proposal needed and necessary; why or why not? Why are certain communities, like some from English Wikinews, negative toward Wikidata and its integration with other projects? George Ho (talk) 03:44, 30 November 2017 (UTC); modified, 03:59, 30 November 2017 (UTC)
Nobody is forcing the English Wikinews to integrate with Wikidata. If they want they can make an RFC remove any integration with Wikidata from their project. It makes sense to give the English Wikinews the ability to decide whether or not they want Wikidata integration but beyond that point I see no reason why that community should be consulted on matters of Wikidata that don't affect them. I don't see why the English Wikinews should be consulted over the Wikidata integration with the English WikiVoyage.
With 133 active editors the English Wikinews community is a tiny spots in the Wikimedia universe and Wikidata wasn't designed for the needs of the Wikinews community. The Wikinews community would likely profit from getting development support but as it stands it has the decision of whether life is better with or without the existing Wikidata integration. Both decisions are fine.
If you look at the more recent projects of Wikidictionary and WikiCommons integrations that are currently developed there was some community outreach to the respective communities beforehand. ChristianKl (✉) 11:46, 30 November 2017 (UTC)
It seems to me that the decision to start new projects generally isn't done by community votes. If you look at the process of starting a new Wikipedia language version it's done by the Language Committee which decides itself about which people are allowed into the committee and which got their authority from the WMF board and not from an RFC.
As far as speaking about projects, Wikipedia isn't a single project. Each language has it's own Wikipedia project (and English has a second one with Simple Wikipedia). If a given Wikipedia doesn't want to use Wikidata they don't have to import any data from Wikidata. They are also free to continue maintain their own interwiki links. ChristianKl (✉) 12:14, 28 November 2017 (UTC)
Since the Board of Trustees authorized the project, is contacting one of current or former Trustees necessary? --George Ho (talk) 20:31, 28 November 2017 (UTC)
Personally, I am very curious about by whom and when exactly it was decided to start Wikidata. I'm sure, though, it was the projects that demanded to have a knowledge base for them. Matěj Suchánek (talk) 13:55, 28 November 2017 (UTC)
I just read de:Wikidata. According to it, the idea for a “Wikidata” project is much older than the project itself. In 2004/2005 it was already described by Erik Möller (Q577729) (e.g. here, German language), who seemingly also coined the project name back then. As far as I know there was at least one demo implementation by Magnus Manske (Q13520818) at that time as well. I have no idea why it took until 2012 that Wikidata actually went live – lack of money, technology not developed far enough, other priorities, whatever. Maybe the veterans of that time can tell the full story? —MisterSynergy (talk) 14:39, 28 November 2017 (UTC)
The story that was told at the recent Chaos Radio interview was that at before 2012 the attempts at making MediaWiki semantic produced https://www.semantic-mediawiki.org/. There's also DBpedia which already existed before Wikidata. ChristianKl (✉) 16:24, 28 November 2017 (UTC)
A few clicks away from that PDF, I found this fascinating page, which I'd never seen before. Other efforts happened in parallel, such as my work from 2007 onwards, to add microformats to Wikipedia (and sister project) pages. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:30, 28 November 2017 (UTC)
I am not sure why you said that. I looked at Wikidata:Community portal, and I figured that this is the best place out of all within Wikidata. However, if it's not the right place, then should I continue at Meta-wiki instead? –George Ho (talk) 21:20, 28 November 2017 (UTC)
Re projects. Have you read m:New project process?

Re the rest, go back and look at what you asked and how you asked it. Somewhat provocative.  — billinghurst sDrewth 01:26, 30 November 2017 (UTC)

The "New project process" was created on 12 April 2012 as a "draft"; it's still a draft. How can the draft be followed? I struck two question in favor of newer ones. Better? George Ho (talk) 03:44, 30 November 2017 (UTC)
To answer both questions, no to both. No motives or anything like that. I wanted to started a discussion out of my curiosity and concerns. For what it's worth, when Wikidata debuted, I got curious and haven't thought that much about Wikidata. I admire its interlanguage linking system and all that. I even also admire a system that encourages merging two duplicate queue items into. However, until now, I have been too busy to be curious more about Wikidata. George Ho (talk) 20:24, 30 November 2017 (UTC)
Retrieved from "https://www.wikidata.org/w/index.php?title=Wikidata:Project_chat/Archive/2017/11&oldid=657866530"