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

# Wikidata:Project chat

 Wikidata project chatPlace used to discuss any and all aspects of Wikidata: the project itself, policy and proposals, individual data items, technical issues, etc.Please take a look at the frequently asked questions to see if your question has already been answered.Please use {{Q}} or {{P}}, the first time you mention an item, or property, respectively.Requests for deletions can be made here. Merging instructions can be found here. IRC channel: #wikidata connect Start a new discussion
 On this page, old discussions are archived. An overview of all archives can be found at this page's archive index. The current archive is located at 2018/06.

## Announcing derivedstatements.js

For many items Wikidata has more information than can be seen on the item page. For example on Whitcomb L. Judson (Q731876) you see that he is an inventor but you don't see his invention. That's because discoverer or inventor (P61) only links from the invention to the inventor and not the other way round. To solve this general problem that statements are only displayed in one way I've written the script derivedstatements.js. The script adds at the end of all item pages a new button to load inverse statements.

You can install the script by adding to Special:MyPage/common.js the following line:
importScript( 'User:Pasleim/derivedstatements.js' ); // [[User:Pasleim/derivedstatements.js]]
--Pasleim (talk) 19:33, 5 June 2018 (UTC)

Wow, simply superb. This is a feature that should have been in native WD for a long time. You even thought of getting labels from inverse properties! Thanks! -- LaddΩ chat ;) 00:21, 6 June 2018 (UTC)
Yes, really nice, thanks! ArthurPSmith (talk) 13:23, 6 June 2018 (UTC)
Incredibly useful, thank you! - PKM (talk) 01:31, 7 June 2018 (UTC)
Thanks Pasleim! Would it be possible to display properties in user language please? — Ayack (talk) 07:27, 7 June 2018 (UTC)
Really nice script. My wishlist
• get local language
• a link to the query in the query editor so that I can easy and fast complement the search with more fields
• paintings with depict show a thumbnail of the painting
- Salgo60 (talk) 11:02, 7 June 2018 (UTC)
• Nice! It would be good if it was possible to then click on a link to add that derived value to the item (where it doesn't exist already). Thanks. Mike Peel (talk) 12:35, 8 June 2018 (UTC)
@Mike Peel: This is already done by User:Frettie/consistency check add.js, I've recently discovered it. --Epìdosis 14:58, 8 June 2018 (UTC)
Excellent. Very useful feature. Thanks a lot! --Beat Estermann (talk) 10:14, 9 June 2018 (UTC)

Thanks for the feedback. I will work on implementing your suggestions. --Pasleim (talk) 17:12, 10 June 2018 (UTC)

Thank you for this great script and keep on implementing on it. Nortix08 (talk) 04:59, 12 June 2018 (UTC)
Hi again. Below a few thoughts after having seen the script in action for a few days:
First, it would be good if the tool had a talk page where we can continue this discussion.
Second, it is not clear to me from where you fetch the labels of the inverse properties. Example: Rise and Fall of the City of Mahagonny (Q760054) has a series of characters <inverse of: present in work (P1441)>. The derivedstatements feature presently uses "shows" as the label of the inverse property, while present in work (P1441) gives characters (P674) as its inverse property. – Should we not be declaring an inverse property for each property. And if so, how should we go about it?
Third, how about duplicate information based on inverse statements on the items? – Example: Schauspielhaus Zürich (Q675022) has a statement <has part (P527)> Main Stage (Schauspielhaus Zurich) (Q39918282), and the derivedstatements feature again lists the same information, because Main Stage (Schauspielhaus Zurich) (Q39918282) has an inverse part of (P361) statement. – Is this duplication of information by the script to be considered a bug or a feature?
--Beat Estermann (talk) 07:31, 13 June 2018 (UTC)
@Beat Estermann: Why don't use User talk:Pasleim/derivedstatements.js? --Marsupium (talk) 11:50, 13 June 2018 (UTC)

Thanks! Could you make this a Gadget please, so we can activate it on user settings? -- JakobVoss (talk) 09:33, 16 June 2018 (UTC)

## Losing property creator rights (part 2)

(Pinging current property creators: @Ayack, Paperoastro, MichaelSchoenitzer, Emw, Izno: @Mattflaschen, JakobVoss, 99of9, ArthurPSmith, Kolja21: @Tobias1984, Viscontino, Jonathan Groß, PinkAmpersand: @Yellowcard, Thierry Caro, Ivan A. Krestinin, Nightwish62, Fralambert: @Danrok, MSGJ, GZWDer, Joshbaumgartner, Srittau: @Almondega, Jura1, Pintoch:)

After the discussion that we had here. This is the text that I would like to add to Wikidata:Property creators

Losing this right:

After a property creator has been inactive for a year, not creating properties and not participating in property proposal discussions, they might be asked by any member of the community if they want to continue holding the right. At this point they will have the opportunity to be active again, or lose temporarily the property creator rights until they decide to apply again to the right, either through the standard process or through the fast process if it is within 6 months of their last logged action (see below).

Property creator rights can also be removed after a community discussion at Wikidata:Requests for permissions/Removal, upon consensus supporting the removal.

Property creator access can also be removed due to voluntary resignation of the property creator. Property creators wishing to resign their access may request removal at Wikidata:Administrators' noticeboard. If the Property creator wishes to regain their access after a voluntary resign, it can be requested at Wikidata:Administrators' noticeboard within 6 months of their last logged action in line with the inactivity policy.

Any comment?--Micru (talk) 07:06, 7 June 2018 (UTC)

Looks good to me! − Pintoch (talk) 08:23, 7 June 2018 (UTC)
No need to ask stewards to remove rights, because Wikidata admins can remove property creator user rights. Stryn (talk) 12:45, 7 June 2018 (UTC)
Ok, I replaced it by the Wikidata:Administrators' noticeboard.--Micru (talk) 14:20, 7 June 2018 (UTC)
I'm alright with the first and last paragraphs, but oppose the vote language in the second. Rights should only be removed in cases where there is consensus, not merely on the basis of a 50/50 'vote'. I would also add the following language: "In any case in which the holder has expressed their desire to retain the property creator right, inactivity alone shall not be cause for removal of said right." There is zero harm in someone only using their PC right sparingly. Josh Baumgartner (talk) 19:59, 7 June 2018 (UTC)
@Joshbaumgartner: That vote paragraph is taken from Wikidata:Administrators#Losing_adminship. If you are not ok with it, we should change it for admins too. I agree that consensus should be more important than a vote. Any specific idea about how we should put that into words? I also agree that the PC right might be used sparingly, but I feel that a year without creating any property and not commenting on property proposals should be enough for the community to consider that this person might have lost interest or maybe doesn't have the time to work with properties. Also bear in mind that if someone loses their PC right they are neither being kicked out of the project, nor prevented of applying for the right again at a later date.--Micru (talk) 11:52, 8 June 2018 (UTC)
@Micru: PC rights are different from the far more ranging administrators rights. There is no reason adminship rules should have to follow property creator guidelines. As for how to word it, simply replace the word "vote" with "discussion", and the "with at least 50%..." part with "upon consensus". My basic point is this, if the person says let me keep it, then done, they get to keep it (unless they have abused it somehow). If they really have gone inactive and have no desire to retain that role, then sure of course they can be removed from the rolls. What no one has been able to explain is what the harm is of someone having the PC right but only using it sparingly (including once every couple of years). Just because a property regarding their area of expertise only comes up infrequently, what value is there in having them have to re-apply each time? If the stated goal is to increase participation in property creation discussions, removing people's PC rights while they aren't looking and forcing them to re-apply won't help. Is there any actual case where someone having the PC right but not using it has caused a problem and thus justifies this effort in the first place, or are we just writing law to solve a hypothetical? Josh Baumgartner (talk) 16:12, 8 June 2018 (UTC)
@Micru: The change looks good. I actually see where you are going at and I agree with you, it gets overwhelming and I am myself guilty of giving up at times in the face of the volume. I do try and respond if I am pinged on something, though if I don't have constructive input, I usually prefer to remain silent versus a generic support or oppose 'vote'. I occasionally browse the list of proposals in full, but as you mentioned there a lot for which I don't have any real input. I do think that a good ecosystem of properties created and curated by an engaged community would be ideal. I am all for anything that brings this about. One of the problems with a smaller number of people participating is that proposals need to remain longer in order to get sufficient discussion and ensure that input has been received. This exacerbates the problem of folks proposing properties, then giving up on them ever getting approved, and when they finally are created, having moved on to other things, leaving the property underutilized. I would like to see a more streamlined process for proposals that are more-or-less clear. Especially for authority control, an experience PC can look at an authority control property and pretty quickly gauge whether or not it is appropriate, and if so, how to set up the property with proper constraints and formatter, etc. As far as I am concerned, this should be something that can be done immediately by a PC if it meets certain criteria, which would both reduce backlog and the imposing volume of proposals as well as improve the use of the property by rapidly putting it in the hands of the community to use. To that end, I think we should perhaps have a select group of PCs that are even more actively engaged--a sort of property swat team. This is getting afield of the original issue, some of this is probably better discussed elsewhere. As for this proposal, I'm fine with the rewording, though I would still like to see something stating that a PC can not have their rights removed over their objections so long as there is no case of them abusing that right (my proposed wording is above). Add that simple protection and I can support this fully. As for other ways to improve the property ecosystem, consider me an ally and let's discuss some ideas. Josh Baumgartner (talk) 17:45, 8 June 2018 (UTC)
@Joshbaumgartner: I don't understand why you want that PCs can keep they right even if they do not participate. Can you please explain your reasoning? --Micru (talk) 12:48, 11 June 2018 (UTC)
@Micru: I've explained it multiple times in these threads. Let me add that just because one does not meet your narrow definition of 'participation' (as written in your proposed language) does not mean that they are not participating. It went unanswered before, but is there a real world case where we need this new rule, and in particular is there a real world case where we need the ability to remove the property creator right from a user who has not abused the right and wishes to retain it? Josh Baumgartner (talk) 17:02, 11 June 2018 (UTC)
@Joshbaumgartner: What is 'participation' in your definition? Your question came also on the previous discussion (at the end, by ChristianKl), do you find my answer satisfactory enough?--Micru (talk) 08:17, 13 June 2018 (UTC)
@Micru: I'm not the one seeking to create a limited definition of participation, that's what this proposal does above. Your answer from that previous thread seems to indicate that you want the list of property creators to be something it is not currently. So my suggestion is that instead of creating a lot of strife taking peoples' rights away, create a new list that meets your more stringent requirements. You are welcome to create a swat team of go-to folks if you want, but there is no need to punish the rest of the people who don't want to dance to that tune. Josh Baumgartner (talk) 15:39, 13 June 2018 (UTC)
@Joshbaumgartner: You say that my definition of participation is limited, but what is your definition of participation?--Micru (talk) 20:33, 13 June 2018 (UTC)
Is there a good reason to have a discussion about such a policy change in the Project Chat instead of creating a RfC? ChristianKl❫ 12:19, 12 June 2018 (UTC)
@ChristianKl: Yes, there is one, I'm quite disappointed with the RFC process here on Wikidata where the discussions stay open for too long and no effort is done to reach a reasonable conclusion before the conversation stales. Besides from my side I consider that this thread satisfies my limited time constraints, but of course if you feel that you need to talk more about this, you are free to start a RfC.--Micru (talk) 08:17, 13 June 2018 (UTC)
I  Oppose the proposal in the current form on the project chat, as I don't think this is the appropriate venue for policy changes that take away the rights of people, especially with the amount of participation that this thread got. ChristianKl❫ 22:18, 13 June 2018 (UTC)
@ChristianKl: I am fine with moving it to a more appropriate venue. Josh Baumgartner (talk) 15:41, 13 June 2018 (UTC)
Oppose There is no need for this proposal. Taking away user rights will not increase participation. There is no harm in allowing users to retain the property creator right even if inactive. I also agree with ChristianK1 that this is probably not the right place for this proposal. Josh Baumgartner (talk) 20:11, 14 June 2018 (UTC)
@Joshbaumgartner, ChristianKl: I do feel the need to clarify the rights and responsibilities of Property Creators. I have opened a new RFC per your request: Wikidata:Requests for comment/Clarifying rights and responsibilities of Property Creators --Micru (talk) 14:16, 15 June 2018 (UTC)

## Some towns don't have "instance of" Q3957 while description shows they are towns

Examples:

That makes extraction of towns from the database difficult to me. Any chance that someone might write a robot to fix these issues? Thanks  – The preceding unsigned comment was added by Mladen.adamovic (talk • contribs) at 13:30, 8 June 2018‎ (UTC).

I'm not familiar with how towns/cities/settlements are handled on Wikidata, but terms like "town" can be a bit problematic, because their exact definition can be vague and varies between languages and countries. See en:Town for all the different definitions. So a settlement (to use a broad term) might be described as a "town" (town (Q3957)) in English, but might not necessarily fall under the definitions of the respective labels in other languages on town (Q3957). --Kam Solusar (talk) 06:55, 9 June 2018 (UTC)
@Mladen.adamovic: French description of Alibunar (Q648863) says it is a "ville", should we add city (Q515)? For correction, can you say if all opština (Q572975) are town (Q3957) (or city (Q515)) ? --Infovarius (talk) 13:46, 14 June 2018 (UTC)

## Coordinates "Imported from Russian Wikipedia"

Coordinates "Imported from Russian Wikipedia" = "Coordinates are likely to be incorrect". Is there any way to put a moratorium on importing data from the Russian Wikipedia? Abductive (talk) 23:11, 9 June 2018 (UTC)

It would be possible to have a bot that removes dta imported from Russian Wikipedia periotically. Before doing that it would however make sense to speak with the people who imported the relevant data. ChristianKl❫ 11:20, 11 June 2018 (UTC)
The problem is that ru.wiki users that are doing so think that P143-sourced statement is better than no statement at all and in order to use WD data in their infoboxes... they imported their data here. Wostr (talk) 11:54, 11 June 2018 (UTC)
@Abductive: Sorry, but earlier "imported from Russian Wikipedia" was a sign of quality (your opinion agains mine, uh?). But better consider mass import from Cebuan Wikipedia. If you have some arguments (statistics, research) for imho, you should give it, otherwise your post is abusive. --Infovarius (talk) 15:21, 14 June 2018 (UTC)
Just check my contributions for when I changed coordinates. You'll see that coordinates imported from the Russian Wikipedia are frequently incorrect. Or do your own study, using the random item feature. Actually checking the coordinates against internet maps reveals glaring errors. Abductive (talk) 16:27, 14 June 2018 (UTC)

## QuickStatementsBot - mass insertion of statements without source

(cur | prev) 12:06, 27 February 2018‎ QuickStatementsBot (talk | contribs)‎ . . (28,959 bytes) (+428)‎ . . (‎Created claim: main subject (P921): Alzheimer's disease (Q11081), #quickstatements; batch #2060 by User:Daniel Mietchen) (undo) (restore)

The batch page at https://tools.wmflabs.org/quickstatements/#mode=batch&batch=2060

Batch #2060
Alzheimer papers (iv):	P921	Q11081
User: Daniel Mietchen (batches by this user)
Status: DONE
Created: 2018-02-27 02:16:26
Last change: 2018-02-27 13:32:53
Status	Count
DONE	10000


How to check the other 9999? What rule was used for all of them? 85.180.29.96 14:23, 10 June 2018 (UTC)

In what way was Q11081 not the main subject of this paper? --Tagishsimon (talk) 14:34, 10 June 2018 (UTC)
Are you trolling? 85.180.29.96 14:42, 10 June 2018 (UTC)
No. Explain why Alzheimer's disease is an inappropriate main subject for a paper entitled "ADO: a disease ontology representing the domain knowledge specific to Alzheimer's disease." --Tagishsimon (talk) 14:48, 10 June 2018 (UTC)
'cos at first glance, you'd think a paper on Alzheimer's Disease Ontology might have something to do with Alzheimer's Disease. --Tagishsimon (talk) 14:50, 10 June 2018 (UTC)
Why that? The topic is ADO not AD. 85.180.29.96 15:34, 10 June 2018 (UTC)
And meanwhile - but only because you asked so nicely and are clearly a splendid person, Special:Contributions/QuickStatementsBot is where you go to "check" the other 9999. --Tagishsimon (talk) 14:52, 10 June 2018 (UTC)
Certainly not. 85.180.29.96 15:40, 10 June 2018 (UTC)

Ah, I see. You want the main subject to be ADO so you've created Alzheimer's Disease Ontology (Q54887745) in such a way as to sever all links between the paper's item, and the disease. That seems completely misguided and counterproductive. --Tagishsimon (talk) 15:13, 10 June 2018 (UTC)

In the conversation here only to you. FTR, the ADO item was created well ahead before the statement of the article item was changed. You have no proof anywhere for your claim.
(cur | prev) 14:16, 10 June 2018‎ 85.180.29.96 (talk)‎ . . (676 bytes) (+430)‎ . . (‎Created claim: subclass of (P279): ontology (Q324254)) (undo) (restore)
(cur | prev) 14:15, 10 June 2018‎ 85.180.29.96 (talk)‎ . . (246 bytes) (+246)‎ . . (‎Created a new item: Alzheimer's Disease Ontology) (restore)

85.180.29.96 15:34, 10 June 2018 (UTC)
(cur | prev) 14:18, 10 June 2018‎ 85.180.29.96 (talk)‎ . . (28,965 bytes) (+6)‎ . . (‎Changed claim: main subject (P921): Alzheimer's Disease Ontology (Q54887745)) (undo) (Tag: new editor changing statement)
--Tagishsimon (talk) 20:27, 10 June 2018 (UTC)

Yes. One tends to create items before adding them in statements to other items. It does not work well the other way around. Meanwhile, you linked the main subject of the paper to an article which had no links to the disease. Explain, please, how someone interested in a report on all papers the main subject of which relate to Alzheimer's disease will find the paper with Q54887745 as you designed it? This is what I mean by misguided and counterproductive. --Tagishsimon (talk) 15:44, 10 June 2018 (UTC)
Not true. In most cases one can add items before creating them first, because others have created them. But still, only because an item was created before it was added, which is what must happen, does not mean that the item creator wants the created item to be the main subject of anything. So, maybe you have seen that ("Ah, I see") but your see-tools must have fooled you, since my brain is still in my head and no one else is in the room here. 85.180.29.96 16:15, 10 June 2018 (UTC)
You did not explain how the article could be found, based on your confection of main subject. That structure/mapping question is the substantive issue right now. A second issue - given the 40,000:1 ratio of artcles with main subject of AD versus main subject of ADO - is whether you have any grasp of the relationship between granularity of subject indexing, and discoverability of records; your absolutist approach suggests not. --Tagishsimon (talk) 20:24, 10 June 2018 (UTC)
Dear Tagishsimon, thank you very much for your feedback. You are very eloquent. You also seem to like to control the discussion and not only that also lead it to something useful. I have one question - why do you talk so much about users and don't spend all your time talking about what you call substantive issue? 92.230.136.177 01:21, 11 June 2018 (UTC)
You seem to be off topic. Please explain how the article can be found by someone interested in Alzheimer's Disease under your arrangement of dissociating it entirely from Alzheimer's Disease. --Tagishsimon (talk) 07:54, 11 June 2018 (UTC)
"You seem to be off topic." - why do you talk so much about users? And what is "Alzheimer's Disease"? 92.228.157.51 12:44, 11 June 2018 (UTC)
You're a dishonest person and you're arguing in bad faith. There's no point in continuing the discussion. --Tagishsimon (talk) 17:28, 11 June 2018 (UTC)
Tagishsimon, Dear Tagishsimon, you wrote "You're a dishonest person" - why do you talk so much about users? And you wrote "you're arguing in bad faith" - why do you talk so much about users? 92.228.157.51 17:33, 11 June 2018 (UTC)

### QuickStatementsBot - AD - data facts

#### QuickStatementsBot - AD - batch facts

1. https://tools.wmflabs.org/quickstatements/#mode=batch&batch=2057 - 837 // Created: 2018-02-27 02:14:53 Alzheimer papers (i): P921 Q11081
2. https://tools.wmflabs.org/quickstatements/#mode=batch&batch=2058 - 10000 // Created: 2018-02-27 02:15:14 Alzheimer papers (ii): P921 Q11081
3. https://tools.wmflabs.org/quickstatements/#mode=batch&batch=2059 - 10000 // Created: 2018-02-27 02:15:46 Alzheimer papers (iii): P921 Q11081
4. https://tools.wmflabs.org/quickstatements/#mode=batch&batch=2060 - 10000 // Created: 2018-02-27 02:16:26 Alzheimer papers (iv): P921 Q11081
5. https://tools.wmflabs.org/quickstatements/#mode=batch&batch=2061 - 10000 // Created: 2018-02-27 02:16:52 Alzheimer papers (iv): P921 Q11081

Total : 40837.

#### QuickStatementsBot - AD - item facts

SELECT ?item ?itemLabel
WHERE {
?item wdt:P921 wd:Q11081.
SERVICE wikibase:label { bd:serviceParam wikibase:language "[AUTO_LANGUAGE],en". }
}


As of 2018-06-11 : 40851 92.228.157.51 12:56, 11 June 2018 (UTC)

### QuickStatementsBot - AD - discussion part 2

Please note, main subject (P921), while its English label suggests should be unique, has no such constraint defined, and in typical use applied to scientific articles I would expect several "main subjects" to be relevant for search and retrieval purposes. So there's nothing wrong with adding "AD" and "ADO" and whatever else people think is relevant. P921 values that are to any degree helpful for searching should NOT be removed! ArthurPSmith (talk) 15:02, 11 June 2018 (UTC)

ArthurPSmith, it was replaced with something more specific. 92.228.157.51 16:29, 11 June 2018 (UTC)

I think it would be much better to add ontology (Q324254) and Alzheimer's disease (Q11081) (or maybe with of (P642)) in ADO: a disease ontology representing the domain knowledge specific to Alzheimer's disease. (Q38448813) than creating something like Alzheimer's Disease Ontology (Q54887745) that would be probably used only in ADO: a disease ontology representing the domain knowledge specific to Alzheimer's disease. (Q38448813). Wostr (talk) 15:27, 11 June 2018 (UTC)

"that would be probably used only" - what is your source? 92.228.157.51 16:27, 11 June 2018 (UTC)

## topic's main category for executive body

Why executive body (P208) has property constraint (P2302) -> value requires statement constraint (Q21510864) -> property (P2306) -> topic's main category (P910) ? I can't understand what to do... Xaris333 (talk) 14:53, 10 June 2018 (UTC)

It seems to be expecting that there is a language wikipedia category associated with each instance of an executive body, and that there is a wikidata item for the wikipedia category, and that you'll use P910 to point from the instance of executive body item to the category item, as for example Cabinet of Israel (Q2578249) does to no label (Q10092812). I'm not convinced that the expectations of the constraint will always be delivered by language wikipedias. --Tagishsimon (talk) 15:02, 10 June 2018 (UTC)
I am using executive body (P208) to Communal Council items (villages). We will never have a category to any Wikipedia.
Xaris333 (talk) 15:06, 10 June 2018 (UTC)
Then I wouldn't worry too much about the constraint. Not sure if another more appropriate property exists than P208, or if P208 is just really badly defined & constrained. --Tagishsimon (talk) 15:17, 10 June 2018 (UTC)
I removed the constraint as it doesn't make sense anymore (it might have made sense in the early days of Wikidata). ChristianKl❫ 20:09, 15 June 2018 (UTC)

## Homo sapiens as the source of human breast milk

User:Brya has twice removed from Homo sapiens (Q15978631) the statement this taxon is source of (P1672) - human breast milk (Q22728), asking in his second edit summary "where are the herds of Homo sapiens kept for milking, where are the stores selling breast milk?". I've reverted, but a that's my second revert, am raising the matter here for wider community input. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 20:12, 10 June 2018 (UTC)

Sounds like language issues. "Product" has multiple meanings in English, and I'm strongly inclined to believe that in this case it refers to the "result" sense, not the "commercial good" sense. However, sometimes ambiguities like this result in descriptions with different meanings. @Brya: What language do you use as your interface setting? --Yair rand (talk) 02:37, 11 June 2018 (UTC)
Obviously, "natural product of taxon" here refers to something that is traded, often enough internationally. The property often is misused, for example for when "found in taxon" should be used. If "result" is allowed in, a huge inflation would result: almost anything would qualify, like "human kidney", "nail clippings", and way beyond. If a property for that is felt to be necessary, create a new property for that. - Brya (talk) 04:05, 11 June 2018 (UTC)
I agree with Brya here. Otherwise it would be possible to list thousands of products. It would make more sense to link in some way from human breast milk to human (or Homo Sapiens) then the other way around. ChristianKl❫ 11:18, 11 June 2018 (UTC)
The argument - Brya's argument - for creating this taxon is source of (P1672) was "We now have the property "natural product of taxon", so it would make sense to also have the reciprocal relationship.". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:10, 11 June 2018 (UTC)
Yes, and the examples in the proposals are clear on that this is for tradeable natural products. It might be argued that Homo sapiens 'produces' sweat and urine, but it would be singularly pointless to record this with this property. And there are several ways to link "human breast milk" to Homo sapiens without using the property "natural product of taxon". - Brya (talk) 16:38, 11 June 2018 (UTC)
Since you insist on such a pedantic interpretation (wholly unjustified by either the aforesaid examples or by the terms of the property proposal): [2] [3] [4]. And human urine has been used for centuries in agriculture, tanning, gunpowder production, and other trades; including being collected commercially: en:Urine#Uses. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:11, 11 June 2018 (UTC)
The question is not if human breast milk is sometimes sold (human kidneys are sometimes sold), but if it is widely accepted as a tradable product. There are taboos and laws against selling it. More importantly, human breast milk is very commonly present, but its role in human society is not that of a tradeable product; only a very minute portion of the whole is sold. - Brya (talk) 05:25, 12 June 2018 (UTC)
That is a question of your own invention. It need not trouble us here. Levels of trade, taboos, and laws are all red herrings. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:09, 12 June 2018 (UTC)
Reality is out there, no matter how hard some people try to ignore it. - Brya (talk) 16:34, 12 June 2018 (UTC)
Reality is that this is a quite popular (but still niche) product: [5], [6], [7], [8]... --Infovarius (talk) 15:30, 14 June 2018 (UTC)

## Item for cities

Hello. We have:

city with millions of inhabitants (Q1637706) city with a population of more than 1,000,000

city with tens of thousands of inhabitants (Q896881) cities with populations from 20,000 to 100,000

Is there an item for cities with population from 100,000 to 1,000,000? Or something else between the first two items. Xaris333 (talk) 11:56, 11 June 2018 (UTC)

These are very ancient items. It is much better to use population (P1082) only. In fact it may be a good idea to deprecate usage of these items.--Jklamo (talk) 15:14, 11 June 2018 (UTC)
Yes, I agree with that. Xaris333 (talk) 15:48, 11 June 2018 (UTC)
There is big city (Q1549591) for population from 100,000 to 1,000,000 but I also recommend to not use these items in statements but instead population (P1082). --Pasleim (talk) 21:08, 11 June 2018 (UTC)
Checked items with city with millions of inhabitants (Q1637706) or city with tens of thousands of inhabitants (Q896881) and without population (P1082) (10 items), added population (P1082) in these items, city with millions of inhabitants (Q1637706) and city with tens of thousands of inhabitants (Q896881) removed.--Jklamo (talk) 19:22, 12 June 2018 (UTC)
They must be remove also from items that already have population (P1082). Xaris333 (talk) 19:43, 12 June 2018 (UTC)
I have removed most of them yesterday, but it seems some of them were missed by my query. Removed rest of them today.--Jklamo (talk) 13:59, 13 June 2018 (UTC)
Very good. I think we must remove also second largest city (Q50330360) and big city (Q1549591). I am not sure about port city (Q2264924). Xaris333 (talk) 17:18, 13 June 2018 (UTC)
@Jklamo: I object. Please don't be destructive and restore all of them. In most cities (at least where I added these statements) this values are original (in the sense that they cannot be covered by existed P1082 claims) and useful. Moreover, often deletion loses the information at all (as there are no other information about reaching such milestones, at least in P1082 claims). And about second largest city (Q50330360), let's discuss it, when you will be able to represent the same information without this item. --Infovarius (talk) 12:55, 14 June 2018 (UTC)
Try for example to answer questions: "Which millioner cities were there at 1910?", "What is the first 100,000 city?" with aid of only P1082. --Infovarius (talk) 15:55, 14 June 2018 (UTC)
Please restore all removed claims! Or at least replace them with similar ones without population. With your removal of e.g. Birmingham (Q2256) it is no longer instance of city (Q515) and it breaks a lot of queries that rely on that kind of city hierarchy. Broken query example: SELECT ?city { ?city wdt:P31/wdt:P279* wd:Q515. }

## Rank of secondary identifier

Some items have more than one ChEBI ID (P683). In ChEBI there is only one primary ID for each entry, but sometimes entries have also secondary IDs (and sometimes these sec IDs are also present in WD). ChEBI ID (P683) has single value constraint (Q19474404), so should I:

? Wostr (talk) 15:47, 11 June 2018 (UTC)

• It's a question that comes up once in a while. Most of the time, we just ignore it, unless the constraint is set to mandatory. Maybe the single value constraint should simply be removed.
--- Jura 16:29, 11 June 2018 (UTC)
• I think I'll try to ignore it then. Wostr (talk) 17:30, 11 June 2018 (UTC)

Wostr - why would secondary be WD:deprecated? Content follows reality or content follows WD's constraint violation lists? 92.228.157.51 17:22, 11 June 2018 (UTC)

Because it's one of possible options. I won't answer the second question as it's irrelevant. Wostr (talk) 17:30, 11 June 2018 (UTC)
I think it depends on why there are two identifiers. If they are a duplicate in the other database then the right thing to do is to mark one as prefered and one as normal. --LydiaPintscher (talk) 18:11, 11 June 2018 (UTC)
I think it's a result of merging different entries in this database, but I couldn't find any confirmation. Wostr (talk) 21:15, 13 June 2018 (UTC)
that however won't remove item from constraint violations list – I’d just like to point out that the single best value constraint type exists :) --Lucas Werkmeister (WMDE) (talk) 15:16, 12 June 2018 (UTC)
Okay, thanks, that may be the solution, I'll discuss this in WikiProject Chemistry. Wostr (talk) 21:15, 13 June 2018 (UTC)

## Basic Formal Ontology

Can someone create items for each ontology listed at

http://ifomis.uni-saarland.de/bfo/users


they all are claimed to use Basic Formal Ontology (Q4866972) 92.228.157.51 16:44, 11 June 2018 (UTC)

Example:

Animals in Context Ontology (Q54913761)
subclass of : ontology
uses : Basic Formal Ontology
official website : http://vtsl.vetmed.vt.edu/aco/


92.228.157.51 16:47, 11 June 2018 (UTC)

I am willing to do this, but won't have capacity until next week. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:12, 12 June 2018 (UTC)
User:Pigsonthewing - thanks a lot! Hopefully step by step real world classification - as found in reliable ontologies - can be added to Wikidata. 85.181.248.20 09:38, 12 June 2018 (UTC)

@Pigsonthewing: please don't add "subclass of" statements to particular ontologies, but use instance of (P31) instead! -- JakobVoss (talk) 13:56, 16 June 2018 (UTC)

@Pigsonthewing:, take care with JakobVoss as this user classified dozens, maybe hundreds, of ID systems as instanceOf ID. [9] BFO has been mentioned here to exactly educate people against these unilateral talk-ignoring misclassifications. 2.245.10.35 23:50, 17 June 2018 (UTC)

As there is disagreement on how to model this, I'll hold off until others have commented, and there is a clear consensus on the prefered model. FWIW, I would use "instance of ontology". Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:50, 20 June 2018 (UTC)

Support <instance of> ontology. These are clearly individually named ontologies, not types or categpries of ontologies. - PKM (talk) 19:03, 20 June 2018 (UTC)

## City as an administrative territorial entity and sister city

Is a city "a territorial entity for administration purposes, with or without its own local government" I think this is wrong. Municipality is subclass of political territorial entity (Q1048835), not the city. Sometimes are the same, so we have one item. Sometimes are not the same:

For example,

Thessaloniki (Q17151) is a city which includes many municipalities. Municipality of Thessaloniki (Q6627746) is the main municipality, but only the one municpality of the city. It is controlling only a part of the city.

Second example: Cyprus is divided to 6 district (1st administrative level). Each district have municipalites and communal councils (villages) (2nd administrative level). Cities are not administrative territorial entity. Each city have many municipalities.

I have these thoughts because of twinned administrative body (P190). Each municipality or community (the authorities of them) make an agreement with a twin municipality or a twin community. Not the city each self. But this, is not for all cases. In some cases, in some countries, "all city=one municipality".

Maybe I am wrong. I am confused. Xaris333 (talk) 20:34, 11 June 2018 (UTC)

You are right, the term "city" has two meanings - one is the human settlement in a purely geographic sense, the other is the political administrative unit. In Wikipedia and thus also here, whenever possible the two meanings are clustered together - simply because in most cases they are practically identical. This is one of problems with the import from geonames via the Cebuano Wikipedia - in geonames the "human settlement" and "political unit" are always separated, which created a lot of duplicate items here. And thus its right that for the sister city property it should be limited to political units, but not only local governments - in Thailand also provinces have done such partnerships. Ahoerstemeier (talk) 21:29, 11 June 2018 (UTC)
This is one of several instances where ceb.wp is causing problems WMF-wide. What they are doing is a nuisance. —Justin (koavf)TCM 22:49, 11 June 2018 (UTC)
• It can also be tricky when there's a distinction between an administrative body and the territory administered by that body. Sometimes there are two entities in Wikidata, sometimes only one (usually for the territory). A territory itself doesn't really have administrative powers, it's the thing that's administered. Ghouston (talk) 02:59, 12 June 2018 (UTC)

The fact is that there is always a distinction between an administrative body and the territory administered by that body, according to the point of view of each discipline. Are the boundaries of the city up to the last buildings, or at the boundaries of the administrative body? (which may include surrounding farmlands, forests, lakes...) Is the population living in the city, or living at the area administered by the local administrative body (so, a similar question would be "where in which we put the population, since the censuses reference the population by administrative bodies and not by cities, towns, and villages"). Going a level up, are the wars or treaties between countries or between the governments of the countries?

The point is that municipalities or any administrative entity are not self-sufficient. The municipalities represent their city, but the boundaries of a municipality and a city are never identical if we see it positively (for example, if we think the city extends to the last houses). There are different types of settlements from a human geography point of view, there are various relations between the city center and the suburbs (or if one wants to see it administratively, the central municipality and the regional ones), the surrounding settlements even when they are in a different administrative unit have a direct connection with the city, etc... In essence, the institution is referred to as such; between "cities" and not municipalities. Sometimes linguistic or formal conventions create ambiguities, but it is natural (self-evident) that in order for twinning to take place that would need action from their representatives (the municipalities) -like in treaties between countries, signed by their governments.

If it is considered correct that eg. if it is not Kavala (Q187352) who was twinned with Nuremberg (Q2090), then it should also apply to the second part: then it would be Municipality of Kavala (Q12282294) and the "Municipality of Nuremberg" (no item). Are not Thessaloniki (Q17151) and Melbourne (Q3141), but the Municipality of Thessaloniki (Q6627746) and the City of Melbourne (Q1919098)? (only the center, with 3% of the city's population) According to formalities, yes. These are the institutions that have signed the co-operation agreements. On the other hand, however, we see that the Municipalities themselves report twinning with other cities, not with other municipalities. That is humanistically correct, as the structure and culture of cities are generally dependent on its center, substantially or even as symbols. Thessaloniki and Melbourne where twinned because more than 152000 Greeks live in Melbourne. A number much larger than the population of the whole "City of Melbourne". It is obvious that in many cases the political administrative unit of the city centre or the greater area acts on behalf of the city. These actions may be symbolic or substantial but the do represent the city. Another example would be the twinning of New York City (Q60) and London (Q84). London is not the administrative body, Greater London (Q23306) is. Sticking with unecessary formalism, would mean that we would turn "twinned cities" to "twinned administrative bodies". -Geraki (talk) 07:34, 12 June 2018 (UTC)

Any suggestions to solve the problem? Xaris333 (talk) 12:15, 12 June 2018 (UTC)

The claim isn't there at the moment. It seems to come and go. I'd suggest pointing out on its Talk page why it shouldn't be there. Ghouston (talk) 06:56, 13 June 2018 (UTC)

My suggestion is to use city (Q515) with instance of (P31) for all cities, plus to use subclass of (P279) with administrative territorial entity (Q56061) (or a subclass of it) if the city and the municipality are the same thing (item). If we have separated items, then we must use only city (Q515) with instance of (P31). For the municipality we will use subclass of (P279) with administrative territorial entity (Q56061). This will affect some properties like twinned administrative body (P190) (I rename it to "twinned administrative body"), we have to move the statements from the city to the municipality item. Xaris333 (talk) 17:14, 13 June 2018 (UTC)

• For what it's worth, for the U.S. we have city (Q1093829) for a "city" in the legal/administrative sense; in many states, this can apply to some small settled places that are by no means a city (Q515). I suspect something parallel applies in other countries. - Jmabel (talk) 04:12, 14 June 2018 (UTC)

Indeed, the solution identified by Xaris333 makes sense:

--FocalPoint (talk) 16:02, 16 June 2018 (UTC)

## Ordre de la Glorie

How can I Express for the item Ordre de la Gloire (Q2085100) that this item must have the mandatory qualifier point in time (P585) when used for the statement award received (P166) Pmt (talk) 21:14, 11 June 2018 (UTC).

AFAIK mandatory qualifier requirements are in the property domain, and so we find P166 has a property constraint in the form of a mandatory qualifier constraint requirement for P585. Presumably the thinking is that all awards must have a temporal property. Is there anything especial about Q2085100 over and above this normal expectation? --Tagishsimon (talk) 22:49, 11 June 2018 (UTC)
No nothing special With Q2085100, just chosen as an example. The answer that in general the requirement are in the property is fine with me. Maybe I just can use Wikidata usage instructions (P2559) saying This item must have P585 as qualifier. Pmt (talk) 23:59, 11 June 2018 (UTC)
Sounds reasonable. --Tagishsimon (talk) 00:13, 12 June 2018 (UTC)
We don't have item based mandatory qualifiers. In this case I also don't see why one would be warrented as it would effectively say: "There are no serious sources that mention that a person received this award but that don't say when they received it" ChristianKl❫ 11:17, 14 June 2018 (UTC)

## Formatter URL for XML

As discussed at User talk:Pigsonthewing#ORCID iD (P496) formatter URI for RDF resource (P1921), User:Tpt suggested that we should not use formatter URI for RDF resource (P1921) for URIs like https://pub.orcid.org/0000-0001-5882-6823/person which return an XML file. Do we need a separate property for these; if not, how should such URLs be represented? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 09:55, 12 June 2018 (UTC)

To add my 2cents, formatter URI for RDF resource (P1921) is specific for patterns for RDF resources URIs and so imho could not be use to link to XML documents. It would be very interesting to create a new URI formatter that would be named something "URI pattern for machine-readable representation" that would link to JSON/XML/turtle... description of the connected entities. It would be also useful in the RDF use case when the resource URIs do not properly redirect to an RDf document describing the resource when an RDF content type is requested. Tpt (talk) 18:55, 12 June 2018 (UTC)

Does anyone have a suggestion? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:27, 16 June 2018 (UTC)

The URI form of ORCID identifiers is https://orcid.org/\$1, try for instance curl -LH 'Accept: text/turtle' http://orcid.org/0000-0002-0880-9125. The URL https://pub.orcid.org/0000-0001-5882-6823/person seems to be no URI but a request to the current ORCID API. We should not put these URLs into Wikidata but refer to the API endpoint and documentation instead. -- JakobVoss (talk) 20:44, 16 June 2018 (UTC)

## Merge pages

Can HNLMS Groningen (Q2157087) and no label (Q25712361) be merged? same ship

and

Can Belgian frigate Louise-Marie (Q339869) and Willem van der Zaan (Q2661233)? also same ship

thx in advance  – The preceding unsigned comment was added by 145.129.41.136 (talk • contribs) at 18:33, June 12, 2018‎ (UTC).

I merged the first. The second could be merged in theory, but not in practice since there's an article for each name on nlwiki. Ghouston (talk) 06:48, 13 June 2018 (UTC)
The second is not the same ship. If it really is, you need to hunt the nl-wiki community to merge the two articles there first. But I'm pretty sure it is not the same ship. Edoderoo (talk) 12:17, 13 June 2018 (UTC)
It's the same ship, it was sold by the Netherlands to Belgium and renamed. Ghouston (talk) 12:29, 13 June 2018 (UTC)
@Stunteltje: any ideas? --Liuxinyu970226 (talk) 11:00, 18 June 2018 (UTC)
Definately the same ship. I don't have any experience in Wikidata, so I cannot give advice how to handle here. But I don't think it is wise to merge the articles in NL Wikipedia. --Stunteltje (talk) 18:45, 18 June 2018 (UTC)
It's the Bonny and Clyde problem: enwiki has a single article for the ship under either name, en:Belgian_frigate_Louise-Marie_(F931), so we need 3 items for it. Ghouston (talk) 09:12, 19 June 2018 (UTC)

## Fixing a regular expression for a format constraint

Irish National Monument ID (P4059) has currently a format constraint with the following regular expression “[1-9]\d*(,\d+)?” which was in the original proposal. This is, however, not correct. Irish national monument numbers consist of multiple digits, optionally followed by a letter. Hence “[1-9]\d*[A-Z]?” would be more appropriate. Example: St. Dairbhile's Church (Q17276540) has monument number 99A per the official list by the Irish Government. Where should format fixes like this be discussed? Is it possible to run a test to see what would be affected by such a change? Thanks, AFBorchert (talk) 05:10, 13 June 2018 (UTC)

• if it's just a matter of correcting the regex, I'd go ahead and do it (it's at Property:P4059#P2302). The template on the talk page of the property includes a sparql query that generally works to check how the current values match it.
--- Jura 05:17, 13 June 2018 (UTC)
• According to this older list the identifier for that item should be 99,01. Anyway, as Jura said, on Property talk:P4059 you find this query in the constraints boxes, where you can test whether things break if you change the regex. It is okay to just fix it, but you can leave a notice on the property talk page too, if you like. —MisterSynergy (talk) 05:24, 13 June 2018 (UTC)

Thank you for your help! I've opened a discussion at the talk page of the property and will wait some days to see if there are any objections. --AFBorchert (talk) 22:02, 13 June 2018 (UTC)

## Changes by PreferentialBot

Never fun to find that a bot does something wrong, but the assumption PreferentialBot does creates problems. It assumes current state to be preferred, but this neglects the history (the past) of the claim. That is it enforces interpretation of the claim in a specific context. This is not a problem for Wikipedia-projects that focuses on infoboxes with current content (what it is now), but it poses a problem for those projects that has infoboxes with complete histories (how it were back then). Rephrased you may say that what's preferred depends on expectations about the infobox, but this is a slight simplification.

To reiterate; the error is the assumption that current is the same as preferred, which is not a general truth. In fact, preferred can be harder to get right than deprecated. The solution is either to stop the bot or to stop reusing the preferred rank while building the infoboxes for the client articles. Jeblad (talk) 09:36, 13 June 2018 (UTC)

One question on this. Does the bot create the problem, or is it only showing the problem in more items? To me it feels the latter is the real issue, and then the bot isn't the solution. Edoderoo (talk) 09:41, 13 June 2018 (UTC)
The problem is general, and as far as I know it can't be properly solved. You can rephrase it as preferred to be some extreme value in some dimension, but there can always be some other use case with other extreme values in some other dimensions. Jeblad (talk) 09:52, 13 June 2018 (UTC)
• The bot doesn't do that for all properties, only for the ones where this is thought to be useful. For infoboxes that should display any non-deprecated statements, one just needs to select differently.
--- Jura 09:43, 13 June 2018 (UTC)
The infoboxes are just an example, and the problem is even worse for content templates. I have just posted a warning at nowiki about using "best" statements, as they are now randomly broken. Jeblad (talk) 09:56, 13 June 2018 (UTC)
Do you have samples of statements that you consider broken?
--- Jura 09:59, 13 June 2018 (UTC)
In Hôtel du Gouverneur militaire de Lyon (Q20980610) [10] it can be said to be correct, as the property is the status. In Japanese Communist Party (Q641600) [11] it is wrong as it is nothing preferred with the current status. You may say that claims you can order somehow should not be marked as preferred (temporal ordering, Q2098061), or claims where the other claims does not carry any useful information (Q641600). Jeblad (talk) 11:19, 13 June 2018 (UTC)
Q641600 still exists, so the ranked statement gives the answer to the question "how many members does it have". Once it's desolved, I'd argue that the most recent value shouldn't have that rank. There are few cases were I don't think preferred rank is helpful, but here it seems helpful.
--- Jura 11:51, 13 June 2018 (UTC)
You give yourself a problem, how to formulate a question to get a specific answer, and find a question, but that isn't the only valid question and thus the solution isn't universally valid. The proper question is given by the context on the client, not carefully chosen by the answer on the repo. The only time preferred makes sense is when all ordering (instantiations) over all valid domains leads to preferred being the valid choice, or at least for the interesting domains. Note that unordered is also an ordering, as is ascending and descending order, so the only case where preferred makes sense is for a truly unordered set. Jeblad (talk) 12:09, 13 June 2018 (UTC)
Which question(s) would it answer incorrectly?
--- Jura 12:20, 13 June 2018 (UTC)
Any question involving for example the history will be wrong. (Second sentence in the opening post: It assumes current state to be preferred, but this neglects the history (the past) of the claim.) Jeblad (talk) 14:26, 13 June 2018 (UTC)
I don't think so. Can you formulate one?
--- Jura 05:02, 15 June 2018 (UTC)
• Isn't this exactly how "preferred" is supposed to be used? From Help:Ranking: "The preferred rank is assigned to the most current statement or statements that best represent consensus (be it scientific consensus or the Wikidata community consensus)." Ghouston (talk) 10:43, 13 June 2018 (UTC)
This is the core of the problem, the current value isn't necessarily preferred, and it isn't necessarily any consensus between projects on what is best. As I said back in 2012, preferred should be abandoned, it creates a lot of problems as it implies non-formal rules that isn't shared by the clients. Jeblad (talk) 11:24, 13 June 2018 (UTC)
Why don't you see Help:Sources as a consensus? ChristianKl❫ 11:47, 13 June 2018 (UTC)
Lua data access allows to retrieve either best or valid statements, the API and SPARQL allow retrieval of ranks. Whatever data consumer wants to get not only current, but also historical information has all the means in their hands to do that. --Marsupium (talk) 10:46, 13 June 2018 (UTC)
• I don't think the bot is your problem. It uses the concept as it's supposed to be used according to Help:Sources.
The problem comes from the "preferred" rank being used to mark two distinct things. On the one hand it marks that a certain datum is more trustworthy than others and on the other hand it marks with datum is current. One way to solve this would be to add an additional rank of "current&preferred". ChristianKl❫ 11:46, 13 June 2018 (UTC)
The bot does not use the sources at all, and even if it did use them it can't say anyhing about the trustworthyness of individual sources unless it has a lot more information than what is available on this project. It is also quite easy to verify that it changes unsourced claims. The present rank-solution is really a two-dimensional thingy with six valid positions, where two normal values are merged, and two other positions are neglected, leaving three positions. Adding a third dimension isn't going to simplify the problem at all. (Note that current is temporally close on a time scale, with near as spatial scales as one other alternative. What about similar in shape?) Jeblad (talk) 12:27, 13 June 2018 (UTC)

## Usefulnes of full statistical history of a country

How useful would it be to have a full history of statistics of a country? For example we would like to add a large amount of the Central Bureau of Statistics(Netherlands) data as properties with values per year for the "Netherlands" item (Q55). As an example we have added a small amount of this data to the test.wikidata page for the "Netherlands" item. https://test.wikidata.org/wiki/Q165131. Historical data for the "Population" item is the example, we would replicate this for hundreds of properties of The Netherlands. It would probably add 100,000+ data items of current and historical data of the country.

Our concern is that it might be too much data to add to a single item and wouldn't be useful to anyone since the same data is readily available at the CBS website (https://opendata.cbs.nl).  – The preceding unsigned comment was added by CBSBot (talk • contribs).

I would rather have this data available as JSONstat, with sufficient information in the item to load it as necessary. Most statistics isn't about a specific item, but about an intersection between several items. Netherlands, children and kindergarten for example. Jeblad (talk) 14:32, 13 June 2018 (UTC)
@Jeblad: So you would say that this type of data is not suitable to be placed on wikidata? CBSBot
@CBSBot: I've been wondering about statistics, and when it is sufficient important to an subject to include it as a set of claims. A nation consists of people, and I guess it would be prudent to include statistics about those people in the item about the nation. It could also be rephrased so that if you have statistics between an instance and a class, then the statistics goes in the instance. What if you turn it around, and you have statistics between a class and a set of instances, then it isn't obvious where the statistics goes. Jeblad (talk) 13:56, 14 June 2018 (UTC)
@CBSBot, Jeblad: Use tabular population (P4179), or propose another tabular property if that's not sufficient. ArthurPSmith (talk) 15:03, 13 June 2018 (UTC)
@ArthurPSmith: I have previously criticized tabular data at Commons, because JSONstat isn't implemented. Jeblad (talk) 14:01, 14 June 2018 (UTC)
@CBSBot, Jeblad: well, whether the Commons tabular format is sufficient or not, it's the only alternative option we have at the moment. The wikidata software cannot handle more than a few thousand statements on an item at present (the user interface becomes essentially unusable at that point) so no, it would not be a good idea to have 100,000+ statements on Netherlands (Q55). ArthurPSmith (talk) 14:06, 14 June 2018 (UTC)

## How would I link this from Wikisource

https://en.wikisource.org/wiki/Page:The_Biographical_Dictionary_of_America,_vol._01.djvu/100 linked to Thomas Gold Alvord (Q2424076)? Do I add a category at Wikisource or just add the url to Wikidata? --RAN (talk) 13:52, 13 June 2018 (UTC)

You would first need to create the article on Wikisource, and transclude the page content. Once there is a page on Wikisource for the specific article, then you create a Wikidata item for that article. Then you can link the WD item for the person to the WD item for the article. We already do this with the Dictionary of National Biography entries, the 1911 Encyclopædia Britannica, and a few similar works. --EncycloPetey (talk) 21:17, 13 June 2018 (UTC)
Can you show an example, I only see links to authors, and portals for subjects of articles tend to be deleted. --RAN (talk) 04:14, 14 June 2018 (UTC)

## Information On Donating Data/Information

Hello everyone, i am brand new to Wiki data and i work for a health organization where we fund research in various areas of health, and we feel strongly about creating an open access policy where information and data that has been collected through our grantees research can be used for all to share and help contribute in the spreading of sound knowledge and science. i have spoken to several other Wiki editors and what not and i am getting great feedback and i want to continue in gathering as much information as i can before moving forward in a respectful manner. my email is enabled so please feel free to email me. Again i just want information on how to donate data and what data would work the best here and i hope our services at my foundation can help. thanks again DaP87 (talk) 19:09, 13 June 2018 (UTC)

@DaP87: You might want to start with Wikidata:Data donation. Best wishes! ArthurPSmith (talk) 19:39, 13 June 2018 (UTC)
@ArthurPSmith: i just started doing exactly that, thank you for responding! DaP87 (talk) 12:19, 14 June 2018 (UTC)

## Property for designs or patterns?

Do we have a property to use to indicate the pattern or design that is typically intrinsic to or applied to an item? For example, madras (Q3276218) <?has pattern> plaid (Q7200585); Paisley shawl (Q24969701) <?has pattern> paisley (Q937704). - PKM (talk) 20:27, 13 June 2018 (UTC)

@PKM: ' don't know. We have the not so well-defined manifestation of (P1557). But I'm not sure/haven't investigated how it is (supposed to be) used. --Marsupium (talk) 22:36, 16 June 2018 (UTC)

## Registration required

Links using the property Deezer artist ID (P2722) only work if a person has an account and is logged in. Do we have some means to mark such properties and sites that require users to have an account to use their links? --EncycloPetey (talk) 21:13, 13 June 2018 (UTC)

I don't think that we have such a property. But we should get it! I'd  support a proposal. Would also be useful for AKL Online Artist ID (P4432) and also for references where there the Cite template family on enwiki for instance has the system described at en:Template:Citation#Subscription or registration required. --Marsupium (talk) 22:48, 16 June 2018 (UTC)
We don't need a property; create (or use) an item or items ("subscription-only service"; "freemium service", etc) and apply them using a suitable existing property ("instance of", or whatever) as a qualifier. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:17, 17 June 2018 (UTC)
• I think it should at least be mentioned in the property description.
--- Jura 09:00, 17 June 2018 (UTC)

## Capital punishment

Do we have a way to mark deaths from capital punishment? Would it be in manner of death? --RAN (talk) 04:07, 14 June 2018 (UTC)

Yes, manner of death capital punishment (Q8454). Ghouston (talk) 04:19, 14 June 2018 (UTC)
And then one of
cause of death (P509) object or value >
if known. Pmt (talk) 07:27, 14 June 2018 (UTC)

## Typography and alphabets

Do we have any standards about representing typography and alphabets and/or any WikiProject that would have that in its scope? - Jmabel (talk) 05:44, 14 June 2018 (UTC)

@Jmabel: There is almost inactive Wikidata:WikiProject Alphabet and ongoing edits about adding part of (P361) to letters (see e.g. A (Q9659)). Also I've done a classification of graphemes some time ago. --Infovarius (talk) 15:51, 14 June 2018 (UTC)
@Infovarius: Thanks! - Jmabel (talk) 19:47, 14 June 2018 (UTC)

## Swedish verb

mula (Q10590035) is a Swedish verb. It seemed natural to me to create a sitelink to https://sv.wiktionary.org/wiki/mula (via wiki=sv, value="mula" under Wiktionary) but I got a message telling me basically not to do that. Seems odd to me; what is correct policy? - Jmabel (talk) 05:57, 14 June 2018 (UTC)

The main issue is that items about words are not supposed in the Q namespace in Wikidata but in the newly created Lexeme/Form namespaces. Maybe the solution is to create a proper item for 'mula' in those namespaces and then add the link from sv-wiki to the newly created item and delete mula (Q10590035). ChristianKl❫ 11:35, 14 June 2018 (UTC)
@ChristianKl: So does that mean that Wikipedia articles about words shouldn't have corresponding Wikidata items, the way all other articles (even disambiguations) should? That seems a bit awkward. - Jmabel (talk) 19:51, 14 June 2018 (UTC)
@Jmabel: Why do you think it's awkward to to link such articles to a lexeme or sense instead of an item in the Q-namespace? ChristianKl❫ 10:55, 16 June 2018 (UTC)
@ChristianKl: I think it's awkward either to ask Wikipedians who may know almost nothing about Commons to have to do things differently for a class of articles that have no special distinction within the site they are familiar with, or to ask a bot to be able to make a distinction such as that the article "African American" is about an ethnic group, but the article "Negro" is about a word. - Jmabel (talk) 16:57, 16 June 2018 (UTC)

## is part of a particular cuisine

How do I indicate that a dish or an ingredient is part of a particular cuisine? Thanks, GerardM (talk) 12:10, 14 June 2018 (UTC)

It's sometimes done with part of (P361) like on Valtellina Casera (Q782709). Ghouston (talk) 12:33, 14 June 2018 (UTC)
There's also indigenous to (P2341) if you're associating the dish to a specific people, location or culture. - PKM (talk) 19:33, 14 June 2018 (UTC)
part of (P361) seems to me to be a little awkward for things like this, because it gives a warning if you don't have the reciprocal has part (P527). It seems to me like we should have an alternative for part of (P361) that doesn't expect that sort of explicit reciprocation. - Jmabel (talk) 19:53, 14 June 2018 (UTC)
In this case, the inverse can just be added to the cuisine item. But there has been a discussion about whether the inverse constraint should be required, and it was even deleted for a while but reinstated. Property_talk:P361#Removing_inverse_constraint. Ghouston (talk) 01:56, 15 June 2018 (UTC)

Maybe "facet of"? --Anvilaquarius (talk) 07:42, 20 June 2018 (UTC)

## Use of instance of (P31)

90.191.81.65 argues that mineral (Q7946) should be only used for mineral samples.
Entities using the chemical formula property should be instances of one of the following classes (or of one of their subclasses): chemical compound.
I think that a mineral sample is an instance of a mineral sample.
The Science tree is rocks > facies > inorganic non-mineral compounds > minerals > mineral varities and species polytypes. Any comments? --Chris.urs-o (talk) 15:04, 14 June 2018 (UTC)
@Chris.urs-o: This sounds like a comment from the blocked user Tobias Conradi (who is very hard to block by IP address). Nevertheless you (or the Mineralogy wikiproject) need to make a decision on whether mineral (Q7946) is a metaclass like chemical compound, or not. Chemicals are generally a bit of a mess - particularly proteins and other biological chemicals. But if you look at mineral (Q7946) right now, it's a subclass of chemical compound (Q11173) which is a metaclass but also of solid (Q11438) which (through the subclass tree) seems to be a subclass of concrete object (Q4406616), whose instances should be real physical objects and not classes of objects. So one of those two seems wrong, and which of them is removed would determine whether P31 is the right relation or not for specific types of minerals to mineral (Q7946). I would tend to favor using subclass of (P279) in this cases, with perhaps a "variety of mineral" class for the P31 relations, but I'm not an expert in this area. ArthurPSmith (talk) 15:34, 14 June 2018 (UTC)
Currently there are over 4k items. I am not able to change over 4k items. --Chris.urs-o (talk) 15:47, 14 June 2018 (UTC)
A query and some quickstatements will do 4k edits ... would take ~5 minutes or so to set it up, after which quickstatements does the grunt work ... if there is a decision on & specification of the proposed pattern of change. There are any number of people on this board with the abiity to do this, and/or to take you through the process, Chris.urs-o. --Tagishsimon (talk) 15:55, 14 June 2018 (UTC)
I posted my examples and a suggestion here: Wikidata talk:WikiProject Mineralogy/Properties#Use of instance of (P31).
To be clear, I think that mineral (Q7946) should be used for mineral species (items with chemical formula property), but the relation should be subclass of (P279). To me it seems that Help:Basic membership properties is quite clear on that. Mineral sample is "an individual or a single thing" (here real physical objects), mineral species is not. 90.191.81.65 16:10, 14 June 2018 (UTC)
Does not seem to work. A valid mineral species is a subclass of a mineral group. Create item 'valid mineral species'?
Earth Sciences:
Nickel-Strunz Mineral Classification:
--Chris.urs-o (talk) 19:59, 14 June 2018 (UTC)

## Different versions of source document

I have created item NRHP nomination: WCAU Studios (Q54987601) describing a particular document for use as a source. Two online, official outlets (the U.S. National Archives and the Pennsylvania state authorities) host PDFs purporting to be that document. However the two PDFs are significantly different: one of the PDFs (at the Archives) appears to be the final version of the document, while the other (PA) appears to be an earlier draft.

So, how should I present this situation in the full work available at (P953) statement(s)? As of right now, I've listed both links with the PA link deprecated. Is there a qualifier I can attach to the PA link to describe the situation? Or should I just use the Archives link and remove the PA link? Or is there a different way to approach the situation? — Ipoellet (talk) 20:22, 14 June 2018 (UTC)

Thank you. I did that, using a value of draft document (Q560361). — Ipoellet (talk) 18:11, 17 June 2018 (UTC)

## Taxonomy in Biology

The French article fr:Taxonomie seems to be about biological classiciation, so isn't it correctly at taxonomy (Q8269924) rather than taxonomy (Q7211)? If the first two items you mention are really duplicates, it would perhaps be possible to merge fr:Taxonomie with fr:Classification scientifique des espèces. Ghouston (talk) 02:39, 15 June 2018 (UTC)
As far as I am concerned, the purpose of Wikidata in linking is to connect pages on the same topic, not to force Wikipedias to re-compose their pages on Wikidata-selected topics. There are some users who try to place sitelinks so as to give Wikipedia pages the exposure they desire rather than in the item where the concepts is dealt with, but fortunately they are the exception. This means that if there is a single Wikipedia that has separate pages on related topics, Wikidata needs to have separate items.
"Taxonomy" and "biological classification" are indeed related topics, but they are not necessarily the same concept. There is no reason why there should not be separate items, although exact placement of sitelinks may not be easy.
The idea of moving sitelinks from "d:Q8269924 (taxonomy, science of finding, describing, defining and naming groups of biological organisms)" to the unrelated "d:Q7211 (taxonomy, classification of things or concepts)" just to make things look neater is really abhorrent. These are completely different concepts and should be kept separate, in any event. - Brya (talk) 04:23, 15 June 2018 (UTC)
@Martinogk: Do not delete your comments once others have replied to them. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 12:13, 15 June 2018 (UTC)

## Wikimania program & Wikidata-related sessions

Hello all,

I drafted a summary of all the Wikidata-related sessions on this page. Feel free to improve it, or add things if I forgot some.

You can also sign-up in the "I'm attending" section so we have a better overview of who from the community is going :)

Cheers, Lea Lacroix (WMDE) (talk) 09:20, 15 June 2018 (UTC)

## Domain and Range of Properties

Dear all,

I have looked at the Wikidata download page. However, I could not find the domain and range of properties.

By looking at a random property I understood that the domain and range definitions are handled via Contraints (https://www.wikidata.org/wiki/Property:P2302). Am I correct?

Thanks,

Giuseppe  – The preceding unsigned comment was added by 150.145.63.151 (talk • contribs) at 09:23, 15 June 2018‎ (UTC).

Yes. However, it isn't mandatory for a property to have either definition and the system does not enforce the constraints. Matěj Suchánek (talk) 13:52, 15 June 2018 (UTC)

## Soup Kitchen as Peruvian cultural monument?

There was a bunch of stuff in soup kitchen (Q2142654) about a Peruvian cultural monument. I removed it just now, but perhaps it would be better split into a new item. I couldn't figure out what was going on with it, but others more familiar with the cultural databases might. Daask (talk) 10:39, 15 June 2018 (UTC)

The Peruvian stuff was added in 2017-11 by User:André Costa (WMSE)'s bot. I suppose it got the item confused with something, the claims should probably go somewhere else. Ghouston (talk) 03:30, 16 June 2018 (UTC)
@Daask, Ghouston: This was due to an error on es.wiki, now corrected. We added some logic to spot these during the import but sadly a few snuck through. Thanks for spotting it and fixing it here. /André Costa (WMSE) (talk) 14:44, 18 June 2018 (UTC)

## New RFC: Clarifying rights and responsibilities of Property Creators

Wikidata:Requests for comment/Clarifying rights and responsibilities of Property Creators --Micru (talk) 14:14, 15 June 2018 (UTC)

## Same item with a property

Is there a constrain for using, with a property, only an item just one time? I mean, for a property, you can add two times the same item. Xaris333 (talk) 20:40, 15 June 2018 (UTC)

If I understand your question correctly ... single value constraint (Q19474404) --Tagishsimon (talk) 21:35, 15 June 2018 (UTC)
Not that. A property can has many values, sometimes the values are the same. I am asking if there is a constrain for using the same value more than one time. Xaris333 (talk) 21:46, 15 June 2018 (UTC)
distinct values constraint (Q21502410)? - PKM (talk) 22:03, 15 June 2018 (UTC)
Not that. A property of one item can has many values, sometimes the values are the same. I am asking if there is a constrain for using the same value more than one time, in the same item. Xaris333 (talk) 22:39, 15 June 2018 (UTC)
Your description is ambiguous and bold characters are not likely to help. Can you give examples of item states where the constraint would trigger a violation, and examples of item states where the constraint would be satisfied? − Pintoch (talk) 07:34, 16 June 2018 (UTC)

Ι can add this statement more than one time. Q8682#P1344. If I have 100 or 200 values to a property, It would be helpful to have a constrain showing that the value is already added. Xaris333 (talk) 09:57, 16 June 2018 (UTC)

## Argo

Argo (Q647931) has a coordinate location (P625) referenced from the Danish Wikipedia. I don't see how this could have an associated geographic location, but I don't read Danish well enough to assure myself to the contrary. Looks to me like User:Steenthbot added it 4 years ago. Anyone understand what's going on here? - Jmabel (talk)

I can't read Danish either, but I do see a coordinate template on the Danish article. I've read the Dutch article and that makes me wonder too why there is a coordinate location (P625) statement on the item and the template on the Danish article. Maybe Steenth can explain it (or another user that can read Danish, like Fnielsen). Mbch331 (talk) 08:55, 16 June 2018 (UTC)
It is come from en:Special:Diff/238250797, where a bot has inserted coordinates to en:Argo (oceanography). And this coordinates was coming to dawiki after the article was translate to danish. Maybe it's an old bot error. --Steenth (talk) 09:34, 16 June 2018 (UTC)
@Mbch331, Steenth:So can we agree that, although cited for, this information is not valid? - Jmabel (talk) 17:01, 16 June 2018 (UTC)
Lacking further response here after nearly 48 hours, I am taking the liberty of deleting the statement. - Jmabel (talk) 15:24, 18 June 2018 (UTC)

## Wikidata as an authority control, in web page metadata

I'd like to document good examples of third parties referencing Wikidata to identify a subject, in webpage markup (e.g. microdata, schema.org) or or page headers, as opposed to simple plaintext or web links in page content. Can anyone suggest some, please? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:17, 16 June 2018 (UTC)

See source of https://www.obi.de, https://www.apple.com and quite a few other large websites. --Lydia Pintscher (WMDE) (talk) 14:59, 16 June 2018 (UTC)
@Lydia Pintscher (WMDE): Thank you. That's interesting; I hadn't envisaged use for the subject of the site as a whole. What about examples for individual page subjects? I also note that OBI uses https://www.wikidata.org/wiki/Q300518 ("https"; "wiki") while Apple uses http://www.wikidata.org/entity/Q312 ("http"; "entity"). Am I correct in assuming that "https" and "entity" are preferred? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:54, 16 June 2018 (UTC)
The Apple one is more "correct" because that is the URI of the concept that is also linked in the sidebar of any item.
I have not come across sites using it for individual page subjects but that doesn't mean much. --Lydia Pintscher (WMDE) (talk) 18:38, 16 June 2018 (UTC)

I'm still seeking examples, at page-level specificity, if anyone has them - someone must be doing this, surely? Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 10:53, 20 June 2018 (UTC)

Like data.bnf.fr (see "Pages équivalentes" section) and SUDOC ("Autres identifiants"), for instance? Nomen ad hoc (talk) 17:00, 20 June 2018 (UTC).
Thank you, but the markup in the former is (for an English speaker): <a href="http://wikidata.org/entity/Q535">Equivalent record in Wikidata :</a>, with no semantic component. I did find one example, which I documented at Wikidata:Wikidata for authority control#MusicBrainz. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 21:34, 20 June 2018 (UTC)

## Pronunciation of letter in language

Hi, I would like to add the fact that a certain letter can be pronounced as a specific phoneme in a specific language; for instance, that letter A (Q9659) can be pronounced as phoneme near-open front unrounded vowel (Q740768) in language English (Q1860). What is the correct way to state this? I wasn't able to find existing statements for this. Thanks! --A3nm (talk) 12:05, 16 June 2018 (UTC)

@A3nm: There is IPA transcription (P898), although that might not be exactly what you're looking for. There doesn't appear to be a property for assigning phonemes to letters yet; maybe you could propose one for creation. Jc86035 (talk) 16:02, 19 June 2018 (UTC)
@Jc86035: Thanks! Yeah, IPA transcription (P898) isn't really it. Not sure I'll have the time to propose the creation of a property right now, but thanks for suggesting! --A3nm (talk) 21:16, 19 June 2018 (UTC)

## Multiple identifier types from the same source

A question of how to represent multiple identifier types from the same source has arisen at Wikidata:Property proposal/IPTC subject code - more comments, and arguments for or against either model, would be welcome. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 13:40, 16 June 2018 (UTC)

## Language agnostic templates for politicians

Hoi, I need help for the creation of a template to be used in multiple Wikipedias in African languages. There are lists of many national politicians, the languages where the same lists exist are en sw ts yo and zu Wikipedia.

As you can see on the example lists, do not expect there to be much (technical) local support; things like columns are often not supported. My objective is to show that once support exists for showing basic information and links do exists showing basis related information it becomes easier to convince high schools to flesh out the lists and templates and make them into more complete articles. Thanks, GerardM (talk) 07:02, 17 June 2018 (UTC)

## number of works (P3740) vs. number of accessible works

Currently, about 8000 items link to folders of the 20th century press archives (Q36948990), which comprise scanned press clippings, sometimes only a few, sometimes thousands of them. However, due to intellectual property restrictions, parts - or even all - of the clippings are not accessible on the web, but only on the premises of ZBW. I would like to update all the PM20 folder ID (P4293) links with qualifiers providing the amount of available documents, both online and in total. number of works (P3740) could be used for one of these numbers, but for the other one I found no fitting property. I suppose the situation that of a total number of works only parts are accessible online, is not so uncommon. Has somebody come across that elsewhere and solved it, or should we consider creating a new property? Jneubert (talk) 13:49, 18 June 2018 (UTC)

## Location of monuments that are in the deposit

Hello people,

I'm wondering about what should we do with monuments (statues, sculputures etc) that are in deposit? For example, the no label (Q51661216) monument, was taken by municipal prefecture of São Paulo (Q10351100) into the deposit because reasons (risk of being stolen etc), how should we put that information in the Wikidata item. Thanks, Ederporto (talk) 14:15, 18 June 2018 (UTC)

It could have an end date on the street location. The location could also be set to the storage location, if known. Otherwise, it's just located (presumably) in the administrative district, somewhere. Ghouston (talk) 00:53, 19 June 2018 (UTC)

## A proposed course of action for dealing with cebwiki/svwiki geographic duplicates

A series of deletion requests by @Joseagush:, my merging of Wikidata items for French communes, and seeing @Exec8: performing full-content replacements recently on cebwiki made me wonder if we needed a codified procedure for dealing with geographic duplicates from cebwiki/svwiki. As a result I proposed the following in response to the aforementioned deletion requests (and I hope @Lsj: and other Wikidata admins can opine on this as well):

• Visit the GeoNames pages for the two Wikidata items (linked from the GeoNames ID on both Wikidata items). Verify that the two GeoNames items actually do refer to the exact same place down to the same level of administrative division1.
• Open a GeoNames account if you do not have one already. Move any appropriate information from the less accurate2 GeoNames item into the more accurate GeoNames item and delete the less accurate GeoNames item (I don't recall offhand if a merge capability exists there).
• Go to cebwiki and svwiki and replace the entire content of the page corresponding to the less accurate GeoNames item with "#REDIRECT [[{name}]]" (where "{name}" should of course be substituted with the name of the page on that wiki corresponding to the more accurate GeoNames item).
• Delete the sitelinks for the page corresponding to the less accurate GeoNames item on those sitelinks' associated Wikidata item, delete the statement for that Wikidata item's GeoNames ID, and merge that Wikidata item with the more accurate GeoNames item's Wikidata item.
I imagine this would be a good course of action for those people who frequently get ticked at the plethora of cebwiki/svwiki-only geographic Wikidata items. @Abductive, Ahoerstemeier: from a recently archived June discussion, @Jura1, Pigsonthewing, Hsarrazin, Liuxinyu970226, YMS, GZWDer: from a March complaint. (I will most likely ping folks from the two discussions archived here at some point.)

1. I say this because I have noticed many conflations of Indian administrative divisions, such as taluka/village or tehsil/census town for example, on other wikis.
2. I use 'accurate' instead of 'detailed' here because the French commune GeoNames items imported by INSEE have geoshapes that the GeoNames items imported by World Gazetteer do not.

Mahir256 (talk) 16:42, 18 June 2018 (UTC)

• Keep in mind that occasionally the administrative excludes the settlement its named after for example Q1002828 is not in Q21347409. There should still be 2 items then, I don't know if France has any like this though. Lucywood (talk) 18:04, 18 June 2018 (UTC)
• Sounds fine. No doubt there are some real duplicates and some erroneous duplicates. Abductive (talk) 18:15, 18 June 2018 (UTC)
• If you like to proceed that way, this is fine for me. I'm not sure it scales though, but I'd love to see if it did. I don't think it's suitable as a general suggestion on how to proceed to other contributors:
• Wikidata should be editable without creating accounts at other Wikimedia websites or third party websites.
• Quality improvements at Wikidata (or lack thereof) should be possible without correcting errors everywhere else.
There are couple of possible approaches that have been identified in earlier discussions of the problem. In the meantime, I suppose most people just try to skip cebwiki/svwiki only items.
--- Jura 07:07, 19 June 2018 (UTC)
I think we should also ping cebwiki sysops @Bentong Isles, Harvzs, Jordz, Josefwintzent Libot: and post the Swedish translation of this section to sv:Wikipedia:Bybrunnen (@Ainali: can you please help us?)
IMHO the GeoNames datas aren't wrong, the wrong is that Lsj himself doesn't know how to insert data links to the existing items (or especially don't know how to find existing articles to add em). --Liuxinyu970226 (talk) 07:59, 19 June 2018 (UTC)

To my experience, we have several problems:

• duplication of administrative subdivision and populated places - something done by geonames, but (usually) not by Wikipedias. In some cases this is handy, e.g. when a municipality consists of several distinct populated places, then Wikipedia (usually) omits the one named same-named as the municipality.
• Elevation data imported from ceb often wrong due to inaccurate coordinates, and even worse often imported here without adding imported from (P143). Especially monadnock (Q1139493)-like hills are affected.
• Items imported by bot, but without the coordinates, thus making it even more difficult to spot the duplication.
Each of these points need different action, as in principle the geonames database isn't that bad as it seems from the recurring discussions here. Ahoerstemeier (talk) 08:33, 19 June 2018 (UTC)

## Representation of Wikidata at the Wikimedia movement strategy process

Since nobody spoke up so far, I would like to represent the Wikidata community at the Wikimedia movement strategy 2018-20, more specifically at the working group of Roles & Responsibilities. Besides of my individual Wikimedia movement-wide concerns, my intention is to bring to the forum whatever the community wants to be discussed in that area, and to bring from the working group whatever it has been discussed that might be relevant to the Wikidata community. If there are questions or suggestions, please do comment. I also would appreciate individual endorsments. I will notify the participants of the Wikidata Community User Group, if you are still not in that group you are invited to join.--Micru (talk) 18:28, 18 June 2018 (UTC)

• @Micru: Can you please develop your user profile with at least ~3 sentences to state more about yourself and your positions on Wikidata strategy? I am not able to evaluate your current user profile with regard to your Wikimedia policy positions. Blue Rasberry (talk) 18:47, 18 June 2018 (UTC)
@Bluerasberry: Thanks for your question and for your support. I have updated my Wikidata profile with some relevant things I've been doing for the project, plus some information about my stance on global policy. I hope it is now more clear for everyone what I would like to accomplish.--Micru (talk) 20:05, 18 June 2018 (UTC)
@Micru: Yes, this is what I wanted. Thanks! Blue Rasberry (talk) 22:42, 18 June 2018 (UTC)
Support I have not worked closely with Micru. I also think this user is overly modest and could do a better presentation of their Wikimedia projects on their user page. However I have had some overlap of my projects with theirs and to the extent of my knowledge Micru is a passionate and thoughtful Wikimedia contributor who is appropriate to sit in a discussion as representative of the Wikimedia community and advocate for good things. Blue Rasberry (talk) 18:56, 18 June 2018 (UTC)
• I'm not sure that's how this works? For affiliates, it seems that there can be representatives, but that doesn't seem to apply to the online projects. (Details of the application are at m:Strategy/Wikimedia movement/2018-20/Working Groups/Apply/Form - for disclosure, I'm planning on applying as a Wikimedian volunteer, not as a rep of any sort of group). Mike Peel (talk) 19:03, 18 June 2018 (UTC)
• @Mike Peel: On this page you can see that Projects can be represented too.--Micru (talk) 20:05, 18 June 2018 (UTC)
• I must be missing it? I can see "We recommend that movement organizations and groups create internal processes for the selection of their candidates", and you can select "Project community" as "your relation to Wikimedia", but nothing about specific project representation? Thanks. Mike Peel (talk) 20:09, 18 June 2018 (UTC)
• Well, you can consider this the "internal process for the selection of a candidate". If there is the need for something more formal, we can for sure come up with something.--Micru (talk) 08:23, 19 June 2018 (UTC)
• @Micru: Hi, thanks for proposing you. Bus as Blue Rasberry said, this is difficult to support someone we don't know: I met you in the different talk's pages, but I never worked with you. So my proposition is to create a simplified version of the application form for the selection process in one of your profile's subpage to present your skills and to post some possible answers for the questions of section Describe your expertise and motivation. I don't think personal details like gender or old living locations are important, but more what kind expertise highlighted by personal baseground you can bring. Thank you Snipre (talk) 20:16, 18 June 2018 (UTC)
• @Snipre: I have started User:Micru/Expertise and motivation with my answers to the questions. I have centered my answers on my volunteer career and not on my private career. For me it would be easier if you (or others) would ask questions about any issue that you want to know my stance about or if there is any aspect that I should elaborate more (like my perspectives on consensus-making, community governance, commons, WMF, etc). There is a lot I can write, but for me it is important to know what is relevant for you. Thanks for your interest in getting to know me better :-)--Micru (talk) 08:23, 19 June 2018 (UTC)
•  Oppose sorry Micru but I don't feel like you're the best person to represent the Wikidata community. Cdlt, VIGNERON (talk) 13:32, 19 June 2018 (UTC)
• @VIGNERON: Can you please elaborate?--Micru (talk) 13:38, 19 June 2018 (UTC)
• This is my personal feeling and it's difficult to elaborate precisely but your contribution to Wikidata is quite low (you're the 820th editor by number of edit) and too episodic. Plus, I'm not convinced by what you wrote on User:Micru/Expertise and motivation. Cdlt, VIGNERON (talk) 13:51, 19 June 2018 (UTC)
• @VIGNERON: I never got the feeling that the number of contributions meant that a person is more or less capable to represent a community. It is true that I normally rely on manual edits and comments, which take me more time/effort and are not so "visible" (in the way that those contributions don't put me up in the scale you use to measure contributor's worth), but as you can see on my profile I've been serving the community in other ways. What exactly didn't convince you about my Expertise and motivation?--Micru (talk) 14:04, 19 June 2018 (UTC)
•  Comment I don't see anyone else stepping up to volunteer to represent Wikidata in the Strategy process. Micru gets points in my book for being willing to take this on. - PKM (talk) 17:50, 19 June 2018 (UTC)

## List of Wikipedias by number of properties they use

Is there a list of Wikipedias by number of properties they use? E.g. on German Wikipedia someone claimed that "Wikidata is no source for Wikipedia." Is that true if it would mean that no value is included in that article names space in dewiki? Are there more Wikipedias like that? 85.180.90.183 19:59, 18 June 2018 (UTC)

• I don't understand your question. In what sense does a Wikipedia "use" a [Wikidata, I presume] property? What exactly is the relation between using a property and accepting Wikidata as a [presumably citable] source [or do you mean something else]? The next sentence I don't follow at all: what does it mean for a [Wikidata] value to be "included in article name space"? For that matter, what do you mean by "value"? A Statement (Q-item + property + value) or something else? And what do you mean by "included": explicitly present, driving content via a template, or something else? And then "like that" in what way: having some contributor who asserts that Wikidata isn't a valid source, or what?? Very confusuing; you may need something two or three times this long to express what you actually mean to ask, an if there is a different language where you can express it more clearly (e.g. German), feel free. - Jmabel (talk) 21:31, 18 June 2018 (UTC)
• There are some statistics about Wikidata usage in Wikipedia, see grafana. But they are not very fine graded, for example they do not distiguish between statements which are used to create maintenance categories and statements which are displayed. There is also Category:Templates using data from Wikidata (Q11985372) which lists some templates that use data from Wikidata. --Pasleim (talk) 21:48, 18 June 2018 (UTC)

## Info Commons draws from Wikidata

I'm guessing that there is something I could do at Camlin Hotel (Q2935180) so that Commons:Category:Camlin_Hotel would say "Seattle, Washington" rather than "Washington, Seattle" in its Wikidata-based infobox, but what would that be? - Jmabel (talk) 21:18, 18 June 2018 (UTC)

Yes. Problem was two P131 values. The usage notes for located in the administrative territorial entity (P131) says use the lowest level and check that it is linked via a chain of P131s to the higher levels. So the solution was to delete the higher-level location. --Tagishsimon (talk) 21:31, 18 June 2018 (UTC)
Ah. But then you get "Seattle, United States of America" which is not how anyone writes a U.S. location: state is always included. At least for Seattle it's unambiguous, but imagine if it was "Springfield" or "Freeport". Do we maybe need some specific property for U.S. states so that we can end up with a saner result? Or should the template on Commons be specialized to handle U.S. differently and trace up from a U.S. municipality to a state as part of the place name? - Jmabel (talk) 21:37, 18 June 2018 (UTC)
Several ways of skinning that cat; agree that the current output is suboptimal. --Tagishsimon (talk) 21:48, 18 June 2018 (UTC)
The line in the infobox template beginning "location -->" needs to be modified so that P131 values are recursively displayed until a value that is instance of (P31) country (Q6256) (or something like that) is reached. Mahir256 (talk) 21:58, 18 June 2018 (UTC)
@Tagishsimon: This is one of the tensions between Wikidata's structure and the ability to access it through Lua modules. The simplest thing here would be to have both "Seattle" and "Washington" as P131 values, but in the right order - then we can access those with one call to a Lua module. The easiest way to do that seems to be to remove the statements then add them back in the right order, which is awkward. The alternative is to try to trace Camlin Hotel (Q2935180) -> located in the administrative territorial entity (P131) -> Seattle (Q5083) -> located in the administrative territorial entity (P131) -> King County (Q108861), but that requires an expensive query through arbitrary access, and doesn't actually give us what we want. An extra step of King County (Q108861) -> located in the administrative territorial entity (P131) -> Washington (Q1223) does, but with another expensive query, and we don't know to stop there unless we do a third expensive check instance of (P31) - which here returns state of the United States (Q35657), which is US-specific, so we need to do all of this for the US, and have a different set of (expensive) queries for other countries. I'm not sure what the best approach is here. (pinging @RexxS: for ideas here!) Thanks. Mike Peel (talk) 22:10, 18 June 2018 (UTC)
I wondered about (mis)using Series Ordinal to denote the preferred order for a set of P131 values; would be rather artificial. --Tagishsimon (talk) 22:14, 18 June 2018 (UTC)
@Mike Peel: Since I updated Module:WikidataIB to use getAllStatements/getBestStatements instead of getEntity, it is obliged to use the entity-id (qid) for every call. As far as I am able to tell, all calls are now equivalent and arbitrary access is not expensive if it uses getAllStatements/getBestStatements. If I'm right about that, it would allow us to use far more complex chains of fetching information. I think I probably need to write a new function call that iterates through "located in the administrative territorial entity (P131)" until it finds no P131, then looks for a "short name (P1813)", failing that a label. That would give us Camlin Hotel, Seattle, King County, Washington, USA. I'll try to knock that up tomorrow and we can test it out. If anybody has a better algorithm, or a list of exceptions that can be programmatically implemented, I could use that instead. A list of entities to use as test cases wuld be handy as well. --RexxS (talk) 22:37, 18 June 2018 (UTC)
(ec) located in the administrative territorial entity (P131) should only refer to the lowest, most local level; otherwise all sorts of queries will become unpredicatable. Additional higher levels will always be in danger of being removed by bot sweeps.
U.S. states are instances of first-level administrative country subdivision (Q10864048), which should be applicable to a lot of other countries; though in some countries one might prefer second-level divisions, such as UK counties or French departments rather than regions. Jheald (talk) 22:41, 18 June 2018 (UTC)
Maybe a property equivalent to or the same as OpenStreetMap's admin_level key? Jc86035 (talk) 06:37, 19 June 2018 (UTC)

──────────────────────────────────────────────────────────────────────────────────────────────────── I've created a basic function that returns the chain and started a discussion at c:Module talk:WikidataIB #Function to return location chain. Sorry for the cross-project discussion, but the module exists on Commons, not on Wikidata and I need it to demonstrate the function. --RexxS (talk) 16:55, 19 June 2018 (UTC)

## "owned by"

I'm finding it very difficult to see the line between owned by (P127) and parent organization (P749). Some cases are clear (a work of art uses the former, a subsidiary company uses the latter) but which applies, for example, to a branch library such as Capitol Hill Branch Library (Q48807472)? - Jmabel (talk) 01:22, 19 June 2018 (UTC)

The property proposal for parent organization (P749) at Wikidata:Property_proposal/Archive/12#P749 gives some information. Ghouston (talk) 08:02, 19 June 2018 (UTC)
Yes, "Wikidata:Property_proposal/Archive/12#P749 gives some information," but it fails to answer my question as to which is more appropriate here. - Jmabel (talk) 16:03, 19 June 2018 (UTC)
I would not use <owned by> here. Instead, I would make Seattle Public Library (Q7442157) <instance of> library system (Q28324850) and use <part of/has part> relations between the system and the branch libraries. However, <parent organization> seems perfectly fine since a library system is clearly an organization (and I would mark Seattle Public Library as a library system in either case). - PKM (talk) 17:58, 19 June 2018 (UTC)
Thanks, edited accordingly. - Jmabel (talk) 16:06, 20 June 2018 (UTC)

## Park designed by...

Right now, constraints on designed by (P287) and movement (P135) have a type constraint that gives a warning if used on instances of urban park (Q22746). Most urban parks are works of landscape design, and so have a designer; some are clearly identified with a movement. For example:

I think the constraint should be altered to consider these statements appropriate, but I'm not sure how best to do this. Perhaps garden (Q1107656) (of which Q22746 is a subclass) should be considered a subclass of creative work (Q17537576)? That would solve it, I believe- Jmabel (talk) 03:59, 19 June 2018 (UTC)

## Death in episode number...

How to better model that a character was killed in specific episode/book/film of series? I use manner of death (P1196) with qualifier described by source (P1343). Or better as reference? Or some other property? And how to distinguish from the case when a death of the character was described in episode (as a flashback, not a main storyline)? --Infovarius (talk) 11:20, 19 June 2018 (UTC)

Infovarius, in general you can use stated in (P248) as a reference to indicate the work where a fictional event is mentioned (both as part of the main storyline and as part of a (in-narrative true) story within this work).
Sometimes present in work (P1441) is used as a qualifier to express that a certain event takes places in a certain work, but mostly to restrict in-narrative facts to a certain story-world (e.g. if Dracula has a son in one work but not in others). So this one would also not be useful to emphasize that a certain event takes place in the main storyline of the work.
It seems to me that start period (P3415)/end period (P3416) as qualifiers could work for this purpose, if one accepts the idea of a storyline being a period (see the property proposal). - Valentina.Anitnelav (talk) 18:24, 19 June 2018 (UTC)

## Proposal for a musical notation datatype for properties

Back in August I had inquired about the possibility of a musical notation datatype for use with new properties related to musical compositions (analogous to the mathematical expression datatype for defining formula (P2534)). There is greater interest from those involved in WikiProject India in improving items related to Indian musical modes and Indian musical meters, in the former case requiring properties for separate ascending and descending scales and distinguishing motifs, as well as dominant notes—please bear in mind the links I have made—and I am sure those who work with Western musical compositions may find such a datatype useful for their own purposes (chord progressions, leitmotifs, opening and closing phrases by analogy with first line (P1922) and last line (P3132), perhaps). The use of such a datatype would work well with the Score extension, most likely with LilyPond notation enabled (ABC notation might work if line breaks were allowed in Wikidata property values), as it is necessary to generate appropriate staves and may be useful to sidestep the use of audio (P51) in generating Vorbis files. Since last proposing the datatype I am becoming increasingly convinced that neither Parsons code (P1236) nor a custom format, as suggested in the previous discussion, will ultimately be portable enough for other uses and would make the generation of staves impossible (with the existing property) or needlessly difficult (with a custom format). As a result, I wish to float the idea once more of creating a musical notation datatype before creating a Phabricator task for it. Mahir256 (talk) 14:02, 19 June 2018 (UTC)

@Shlomo, ChristianKl: from the previous discussion. @Bodhisattwa: as someone who has been taking a recent interest in this. @Lydia Pintscher (WMDE), Lea Lacroix (WMDE): as those who may be interested. Notified participants of WikiProject India Notified participants of WikiProject Music

Note the datatype for defining formula (P2534) is a subtype of the "string" datatype, which is limited to 400 characters (which has proved a problem for some uses). What do you anticipate would be the rough size of these musical notations? It might be something that would be better handled by a new Commons file format? ArthurPSmith (talk) 14:27, 19 June 2018 (UTC)
@ArthurPSmith: If we need a larger piece (or multiple smaller pieces—akin to relegating numerous uses of population (P1082) to tabular data) of musical notation from a composition, and if it isn't in anyone's interest to host the entire transcribed composition on Commons in some way, the piece(s) could be relegated to some new data format on Commons or elsewhere. For the moment, though, the properties I envision using this datatype would most likely be smaller samples—maybe a few measures at most—specifically to stay under the limit for the string datatype and to help avoid the possibility of too much unstructured data entering Wikidata (a complaint I have seen regarding some proposed string properties). Mahir256 (talk) 14:40, 19 June 2018 (UTC)
@Mahir256: We already have MuseScore ID (P4097), although there are only several accounts run by the site which actually publish official scores (all scores are associated permanently with their uploader, so we can't really use most of them) so its use would be very limited unless Wikidata somehow acquired an "official" account with which to upload files in some manner. Jc86035 (talk) 15:58, 19 June 2018 (UTC)
@Jc86035: I am aware of that property, but it is primarily intended for the entire score of a particular composition. This datatype would be intended for a few measures' worth of music, as I noted to Arthur above and as I hope was implied by the sorts of applications I listed in my original post. Mahir256 (talk) 16:54, 19 June 2018 (UTC)

@ArthurPSmith, Jc86035: Some examples of the sorts of properties that would benefit from this explicit datatype:

(For completeness, the primary dominant note is no label (Q12434661) and the secondary dominant note is no label (Q12423481).) There are some melodic modes for which the scales and motif are much longer, but I believe none would exceed the length imposed for the string datatype. Mahir256 (talk) 14:12, 20 June 2018 (UTC)

• Intuitively, it seems to me that storing the data at commons would be preferable. Having a property that can only store the notation of short songs but not of long songs would likely be annoying to people who use it. Are there arguments against putting the data on commons? ChristianKl❫ 18:24, 20 June 2018 (UTC)
• @ChristianKl: Apologies if I'm drawing too many unneeded parallels to the present status of LaTeX equations, but I think the situations here are very similar. We have migrated quite a bit away from storing images of equations on Commons to simply rendering them on wikis where they are needed (whether on Wikipedia, on Wikisource, and now here), even as we don't store full-fledged mathematical derivations in statements here due to issues with the maximum length of the string datatype. Similarly we have the ability to migrate from storing images of music snippets on Commons to simply rendering them where they are needed (whether on Wikisource, on Wikipedia, and potentially here), and I am sure it will not be feasible to store full scores of pieces as statements here for the same technical reason. As such the use cases I have suggested above are intended to exclude the possibility of storing entire scores, just as entire derivations are excluded from being stored here. Mahir256 (talk) 18:54, 20 June 2018 (UTC)

## Merge pages

Can platform (Q27921656) and Platform economy (Q50867887) be merged? Both are about the "economic and social activity facilitated by platforms": the former focuses on the economic actor while the latter covers the economic activity. That's more a difference of form than substance though. Freedatum (talk) 15:41, 19 June 2018 (UTC)

## Try the prototype of the new termbox on mobile

Hello all,

As a next step of the termbox project, following the first feedback loop about the termbox, the Wikidata team developed a first prototype of the termbox on mobile. We would love to have your feedback on it. Please have a look at the information on this page and give us feedback on the talk page. Thanks! Lea Lacroix (WMDE) (talk) 16:23, 19 June 2018 (UTC)

## "Cross Lang Conflicts"

Some editors may be interested in this discussion on the English Wikipedia, where Xinbenlv has demonstrated results from a partly closed-source system to find conflicts in data between Wikipedias. I asked him to post here but he has not yet, so I am linking to the original discussion. Please post comments there. Jc86035 (talk) 18:27, 19 June 2018 (UTC)

## Continent of cities

Is the continent cities belong to represented in Wikidata? For example: Does Wikidata know that Ankara (Q3640) belongs to Asia (Q48)? As far as I see Wikidata does know that Ankara (Q3640) belongs to Turkey (Q43), but Turkey (Q43) belongs to two continents. --Jobu0101 (talk) 19:11, 19 June 2018 (UTC)

@Jobu0101: This should ideally be inferred from the continent Ankara Province (Q2297724) is on (contrast with the continent Edirne Province (Q83102)—also in Turkey (Q43)—is on). Mahir256 (talk) 19:24, 19 June 2018 (UTC)
@Mahir256: Ideally because it isn't? Do you also know the answer to my continent related question here? --Jobu0101 (talk) 19:58, 19 June 2018 (UTC)
This is not limited to the continent property. There are also properties like located in time zone (P421) or official language (P37) added to every village, even they are same all up the hierarchy to the country. As there is nothing like a calculated statement within Wikidata (not sure about what's possible with Lua in Infoboxes) we probably have to live with the redundancy. Ahoerstemeier (talk) 21:27, 19 June 2018 (UTC)

## Quick Statements not creating items

Hi, today I'm trying to create some items through Quick Statements, but it is stuck forever in "running". Anyone know what is the problem? Ederporto (talk) 20:15, 19 June 2018 (UTC)

In QS2, this may indicate a bug in the syntax - QS1 would skip the error but QS2 gets hung up on it. Could you post an example of what you're trying to create? Andrew Gray (talk) 20:39, 19 June 2018 (UTC)

I also have problem with adding statements. It is stuck. For example, Q4827920 P1082 47 Xaris333 (talk) 22:12, 19 June 2018 (UTC)

Quick Statements is not working for me either. User:Magnus Manske is that something you can restore? --Jarekt (talk) 01:43, 20 June 2018 (UTC)
Still frozen :-( Jheald (talk) 07:35, 20 June 2018 (UTC)
It doesn't work for me either, still stuck forever in "running". There aren't errors in JavaScript console. @Magnus Manske:. --Rotpunkt (talk) 08:08, 20 June 2018 (UTC)

Server response:

<b>Fatal error</b>:  Cannot access empty property in <b>/data/project/quickstatements/public_html/quickstatements.php</b> on line <b>561</b><br />


QS was last modified on Friday. Matěj Suchánek (talk) 08:34, 20 June 2018 (UTC)

Fixed now. --Magnus Manske (talk) 11:56, 20 June 2018 (UTC)
 I think that this discussion is resolved and can be archived. If you disagree, don't hesitate to replace this template with your comment. Matěj Suchánek (talk) 06:41, 21 June 2018 (UTC)

## How to add to an existing item the label in a new language?

I tried with the gadget "Labels List" and in the end got a message "API error". Then I tried with the Beta version "List of Headers". In the end I got: ✘ Error : (permissiondenied) You do not have the permissions needed to carry out this action. Tue Jun 19 2018 23:32:07 GMT+0300 (FLE Daylight Time)

It is true that on this page (https://www.wikidata.org/wiki/Q4233718) there is a padlock saying "this page has been semi-protected". Can this be the reason for the above message or something else?

Best regards Evgeni

@Dimitrove.tmp: Welcome to Wikidata! The padlock icon means that you need to contribute a bit more; in particular, your account needs to be at least four days old and have at least 50 edits on it before you can edit that item. Mahir256 (talk) 21:15, 19 June 2018 (UTC)

Thank you Mahir256.

## About the empty Query namespace

Currently we have Query: namespace, along with its Query talk:, but both are having no pages.

As we now use WQS, I think both namespaces can be safety dropped, anyone oppose? --Liuxinyu970226 (talk) 23:39, 19 June 2018 (UTC)

I don't think that was the idea of the query namespace. Sjoerd de Bruin (talk) 08:10, 20 June 2018 (UTC)

## General topic properties

Another new navbox:

{{General topic properties}}

-- Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 14:00, 20 June 2018 (UTC)

## How to indicate a topic intersection when not a category

Medical facilities of Seattle (Q6806529): undoubtedly I was wrong to use category combines topics (P971) here because this is not a category, but I don't see how else to express this. - Jmabel (talk) 16:11, 20 June 2018 (UTC)

Similar issue would arise for Museums and galleries of Seattle (Q6941146). I see someone there tried "is a list of" but that's not OK because neither of these is a "list article". - Jmabel (talk) 16:15, 20 June 2018 (UTC)
Not sure if this is what you need, but facet of (P1269) serves a similar purpose I think. ArthurPSmith (talk) 17:19, 20 June 2018 (UTC)
ie subclass of (P279) = medical facility (Q4260475) + facet of (P1269) = Seattle (Q5083). Also location (P276) = Seattle (Q5083). Though, is that sufficient? Jheald (talk) 17:30, 20 June 2018 (UTC)
It's a list, so say so, like this. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 18:12, 20 June 2018 (UTC)
@Pigsonthewing: I usually agree with you on most things, but in this case I think you are wrong. Look at the en-wiki article. It is not a "list article" or anything like. It is a prose overview of medical facilities in Seattle; it doesn't even contain a list. - Jmabel (talk) 22:41, 20 June 2018 (UTC)
"Other hospitals in the community include Swedish Medical Center/Ballard (formerly Ballard General Hospital), Swedish Medical Center/Cherry Hill (formerly Providence Seattle Medical Center), and Swedish Medical Center/First Hill (Swedish's original location); Virginia Mason Hospital, on First Hill; the Seattle Division of the Department of Veterans Affairs' Puget Sound Health Care System on Beacon Hill; the Fred Hutchinson Cancer Research Center in Cascade; and Kaiser Permanente's Capitol Hill campus (outpatient only, formerly Group Health Central Hospital and Family Health Center)." Looks like a list to me. Also, the page is in en:Category:Washington (state)-related lists. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:56, 20 June 2018 (UTC)
OK, there's one list-like paragraph in the middle (I didn't notice that at a quick scan) but "First Hill is widely known as 'Pill Hill' for its concentration of hospitals and other medical offices. In addition to being the current home of Harborview, Swedish, and Virginia Mason, it is also the former location of Maynard, Seattle General, and Doctors Hospitals (all of which merged into Swedish) and Cabrini Hospital," and "In 1974, a 60 Minutes story on the success of the then four-year-old Medic One paramedic system called Seattle 'the best place in the world to have a heart attack.' Some accounts report that Puyallup, a city south of Seattle, was the first place west of the Mississippi River to have 911 emergency telephone service,' hardly seem like list content to me. Still, if en-wiki counts it as a list, I guess that suffices. - Jmabel (talk) 23:22, 20 June 2018 (UTC)

When I click the  link next to a second-level head on this page, the editor is opening a different section. Can someone look into this? Thanks! - PKM (talk) 18:54, 20 June 2018 (UTC)

@PKM: This often happen when a section has been added or deleted, between you fetching the page and starting to edit it. Try refreshing the page in your browser, then editing. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 22:54, 20 June 2018 (UTC)
@Pigsonthewing: Thanks, Andy. - PKM (talk) 00:24, 21 June 2018 (UTC)

## "individual aircraft"

I raised this question before, but as part of a recently-archived discussion of several different matters. This particular question was never answered, at least not to my satisfaction. Really, the only answer I had was an IP contributor telling me my question wasn't reasonable. I still believe it was.

Coming in as an experienced Wikimedian and database developer but relatively new to Wikidata, I would expect that individual aircraft (Q21051516) would be a class that is to be used for individual aircraft. Since it has no description, and it is not that, it is very hard for me to imagine what it is. - Jmabel (talk) 22:49, 20 June 2018 (UTC)

I interpret this item in the same way as you. This means instance of (P31) should be replaced by subclass of (P279). --Pasleim (talk) 01:00, 21 June 2018 (UTC)
@Pasleim: I was informed by several experienced users in the prior exchange that my interpretation was not how it should be used, and there is not a single individual aircraft that uses it as the target of instance of (P31). There had been a few; one result of the prior discussion is that those were removed. - Jmabel (talk) 03:43, 21 June 2018 (UTC)
• Some blocked user delete the statements. Obviously they should be restored.
--- Jura 05:45, 21 June 2018 (UTC)

## Requesting a merge

What's the right way to request an item merge? no label (Q49655286) is certainly a duplicate of Lummi Island (Q937204) - Jmabel (talk) 23:25, 20 June 2018 (UTC)

You don't need to request an item merge, you can do it yourself. For instructions, see Help:Merge. --Pasleim (talk) 01:03, 21 June 2018 (UTC)

## Elevation above sea level

What can it possibly mean to have the (current) statement

elevation above sea level (P2044) < 23 metre >

The islands, obviously, have a minimum elevation of 0, and their highest point is the peak of Mount Constitution (Q6920229), at 734 metres. Is there some useful way to express that in San Juan Islands (Q1196315)? Because the current statement is like saying "part of this house is 34 cm off the ground." - Jmabel (talk) 23:43, 20 June 2018 (UTC)

## Bot question

This is from a long time ago, so maybe it's something long since fixed, but how did someone born in Nebraska and who made her career mainly in Seattle get designated as a "British artist"? - Jmabel (talk) 05:19, 21 June 2018 (UTC)

https://en.wikipedia.org/w/index.php?title=Nellie_Cornish&diff=next&oldid=594725354
--- Jura 05:46, 21 June 2018 (UTC)

In short Garbage in, garbage out (Q1569381)? - Jmabel (talk) 06:15, 21 June 2018 (UTC)