Property talk:P17

From Wikidata
Jump to: navigation, search

Documentation

country
sovereign state of this item
Represents country (Q6256)
Data type Item
Template parameter e.g. "subdivision_name" in en:Template:Infobox settlement, if subdivision_type = Country
Domain geographic features/places: all (note: this should be moved to the property statements)
Allowed values Subclasses of:
  1. sovereign state (Q3624078),
  2. client state (Q1151405),
  3. dependent territory (Q161243),
  4. disputed territory (Q15239622),
  5. occupied territory (Q2577883),
  6. provisional government (Q59281),
  7. fictional location (Q3895768),
  8. for historic civilizations:
    1. ancient civilization (Q28171280),
    2. polis (Q148837),
    3. ancient Chinese state (Q836688),
    4. realm (Q1250464),
    5. state in the Holy Roman Empire (Q26830017).
(note: this should be moved to the property statements)
Example Madrid (Q2807)Spain (Q29)
Wusung Radio Tower (Q2270034)People's Republic of China (Q148)
Robot and gadget jobs
Tracking: usage Category:Pages using Wikidata property P17 (Q20116958)
Tracking: local yes, WD no Category:Country not in Wikidata, but available on Wikipedia (Q22158803), Category:Infobox road articles without wikidata country (Q25348518)
See also country of origin (P495), country for sport (P1532), country of citizenship (P27), territory claimed by (P1336), culture (P2596), applies to territorial jurisdiction (P1001), located in the administrative territorial entity (P131), indigenous to (P2341), cuisine (P2012)
Lists
Proposal discussion Property proposal/Archive/1#P17
Current uses 7,850,845
[create] Create a translatable help page (preferably in English) for this property to be included here
Conflicts with “instance of (P31): human (Q5), given name (Q202444), female given name (Q11879590), male given name (Q12308941), ocean (Q9430), taxon (Q16521), Wikimedia disambiguation page (Q4167410), Wikimedia category (Q4167836), Wikimedia template (Q11266439): this property must not be used with listed properties and values.
List of this constraint violations: Database reports/Constraint violations/P17#Conflicts with P31, hourly updated report, SPARQL
Value type “sovereign state (Q3624078), client state (Q1151405), occupied territory (Q2577883), disputed territory (Q15239622), state in the Holy Roman Empire (Q26830017), polis (Q148837), ancient Chinese state (Q836688), realm (Q1250464), fictional location (Q3895768), ancient civilization (Q28171280), provisional government (Q59281), country (Q6256), dependent territory (Q161243): This property should use items as value that contain property “instance of (P31)”. On these, the value for instance of (P31) should be an item that uses subclass of (P279) with value sovereign state (Q3624078), client state (Q1151405), occupied territory (Q2577883), disputed territory (Q15239622), state in the Holy Roman Empire (Q26830017), polis (Q148837), ancient Chinese state (Q836688), realm (Q1250464), fictional location (Q3895768), ancient civilization (Q28171280), provisional government (Q59281), country (Q6256), dependent territory (Q161243) (or a subclass thereof).
Exceptions are possible as rare values may exist.
List of this constraint violations: Database reports/Constraint violations/P17#Value type Q3624078, Q1151405, Q2577883, Q15239622, Q26830017, Q148837, Q836688, Q1250464, Q3895768, Q28171280, Q59281, Q6256, Q161243, SPARQL
Qualifiers “start time (P580), earliest date (P1319), point in time (P585), latest date (P1326), end time (P582), applies to part (P518), subject of the statement (P805), statement disputed by (P1310), reason for deprecation (P2241), located in the administrative territorial entity (P131), participant (P710), diplomatic mission sent (P531): this property should be used only with listed qualifiers.
Exceptions are possible as rare values may exist.
List of this constraint violations: Database reports/Constraint violations/P17#Allowed qualifiers, SPARQL
Property “country (P17)” declared by target items of “country (P17): If [item A] has this property with value [item B], [item B] is required to have property “country (P17)”.
Exceptions are possible as rare values may exist.
List of this constraint violations: Database reports/Constraint violations/P17#Target required claim P17, SPARQL, SPARQL (by value)
Value England (Q21) will be automatically replaced to value United Kingdom (Q145).
Testing: TODO リスト
Value Scotland (Q22) will be automatically replaced to value United Kingdom (Q145).
Testing: TODO リスト
Value Wales (Q25) will be automatically replaced to value United Kingdom (Q145).
Testing: TODO リスト
Value Wallis and Futuna (Q35555) will be automatically replaced to value France (Q142).
Testing: TODO リスト
Value Easter Island (Q14452) will be automatically replaced to value Chile (Q298).
Testing: TODO リスト
Value Guam (Q16635) will be automatically replaced to value United States of America (Q30).
Testing: TODO リスト
Value Northern Mariana Islands (Q16644) will be automatically replaced to value United States of America (Q30).
Testing: TODO リスト
Value Mayotte (Q17063) will be automatically replaced to value France (Q142).
Testing: TODO リスト
Value Antarctic Treaty area (Q21590062) will be automatically replaced to value Antarctic Treaty area (Q21590062) and moved to located in the administrative territorial entity (P131) property.
Testing: TODO リスト
This property is being used by:

he:תבנית:אישיות_תקשורת, he:תבנית:בחירות, he:תבנית:מדינה במשחקים האולימפיים, he:תבנית:מפה אוטומטית/מדינה, he:תבנית:מפת מיקום מוויקינתונים, he:תבנית:עיר, he:תבנית:עיר/מדינה, he:תבנית:אנדרטה, he:תבנית:בית חולים, he:תבנית:בית מלון, he:תבנית:מבנה, he:תבנית:מוסד אקדמי, he:תבנית:מערה, he:תבנית:ספרייה, he:תבנית:קו רכבת תחתית, he:תבנית:קו רכבת, he:תבנית:תחנת רכבת


Please notify projects that use this property before big changes (renaming, deletion, merge with another property, etc.)

This property is being used by:

Please notify projects that use this property before big changes (renaming, deletion, merge with another property, etc.)

This property is being used by:

Please notify projects that use this property before big changes (renaming, deletion, merge with another property, etc.)

External uses[edit]

This property is being used by:

Please notify projects that use this property before big changes (renaming, deletion, merge with another property, etc.)

Pictogram voting comment.svg countries should agree: ATE in single country
current countries should have best rank, historic or disputed - normal, absolutely wrong - deprecated
Violations query: SELECT DISTINCT ?item WHERE { ?item wdt:P31/wdt:P279 wd:Q15916867 . # ATE in single country ?item wdt:P17 ?countriesOfItem . ?item wdt:P31 ?class . ?class wdt:P17 ?countriesOfClass # disagreement FILTER NOT EXISTS { ?item wdt:P17 ?countriesOfItem . ?class wdt:P17 ?countriesOfClass . FILTER(?countriesOfItem = ?countriesOfClass) } }
List of this constraint violations: Database reports/Complex constraint violations/P17


Discussion[edit]

Archive[edit]

Older discussions archived under Property talk:P17/Archive.

Change this to be about modern countries[edit]

Some geographical items are listing more historical periods than necessary and it's a bit arbitrary to decide which historical governments should count as countries and which shouldn't. For example, there are German towns that have P17= German Empire, Weimar Republic, Nazi Germany, Allied-occupied Germany, East Germany, and Germany. I don't think towns need to list every type of government they have been uneder. I propose that we make this property specifically about what modern country the item is in. When we want to list historical countries associated with an item, there are other ways of doing it, like applies to territorial jurisdiction (P1001) for government bodies and country of citizenship (P27) for people. --Arctic.gnome (talk) 19:08, 11 January 2014 (UTC)

I understand the problem, but it will look strange to have "Ancient city of Karthago country (P17) Tunisia". -- Lavallen (talk) 19:16, 11 January 2014 (UTC)
Wikidata should not cover the present only, but also the past. We do not remove information about the profession of individuals just because that individual has retired or died and does not work in his profession any more. Likewise, we should not remove (or forbid) information about what countries an item was related to in history. When we want to find out the current country, we can use date properties on the country and/or date qualifiers on the country (P17) statement. --UV (talk) 20:30, 11 January 2014 (UTC)
Agree. Al lot of information in infoboxes depends not on present, but on past, including, for example, place of birth and death. It should be possible to retrive correct person->place of birth->country information depending on person date of birth, that is used a lot in ruwiki infoboxes. -- Vlsergey (talk) 04:11, 13 February 2014 (UTC)
I liked the single value constraint, see https://www.wikidata.org/w/index.php?title=Property_talk:P17&direction=prev&oldid=90834252 . I don't see advantages in splitting the property according to modern vs. historical states. And I believe it would become insonsistent from the beginning. On the other hand, I think that not every German town has to repeat the German history in its P17 values. --Zuphilip (talk) 22:11, 11 January 2014 (UTC)
The problem is that each German town can have different German history... And I suppose it should be revealed. Infovarius (talk) 05:41, 12 January 2015 (UTC)
  • @UV: I don't think that analogy is apt. We should list the past professions of a retired person because that's what made them notable. For geographic features, on the other hand, their political history isn't what's notable about them. For example, on an item about a random lake in Italy, the fact that it used to be part of the Italian Republic and the Italian Monarchy and the Papal States and the Roman Republic and the Roman Empire is true, but what makes the lake notable is it's geographic location and its depth and its out-flowing rivers. I doubt many people using the data care about the political history of the county of every geographic feature. Maybe removing all historical countries isn't the answer, but I think we should at least set some rules about when a historical country needs to be listed and when we would be better off listing the modern country or using country of citizenship (P27) or applies to territorial jurisdiction (P1001) instead. If, on the other hand, we do allow all past countries, we need to add hundreds of new items to our "one of" list to include all the countries that lasted for five years before being merged into something else. --Arctic.gnome (talk) 22:36, 11 January 2014 (UTC)
    If we want to be able to use Wikidata to answer queries (such as: longest river in the Austro-Hungarian empire, deepest lake in the German Democratic Republic/in Czechoslowakia), then the easiest way is to add all historic countries. For many geographic features, their fate is not strongly linked to any country (a mountain does not change is height whether it happens to lie in China or Brazil), still we add the country (P17) information. If we do add country information to geographic features, why confine ourselves to the present and diminish the possibility to use Wikidata for queries? --UV (talk) 23:30, 11 January 2014 (UTC)
Current implementation in ruwiki will show all values, but I an not sure, this is not my implementation. Some thinks: today we add historical information. Tomorrow somebody will add information about non-official division. After that alternative history information (Moskow <country> German with qualifier <wanted by> Adolf). After that fictional (Moskow <country> Galaxy Union Federation). And more another cases... Every new case will cause changes in every data client... I think this is way from database to datahell. Is there real task where is needed historical information? — Ivan A. Krestinin (talk) 16:59, 12 January 2014 (UTC)
If qualifiers are added correctly, You can design a module that only gets the "current" or "latest" information. I have templates on svwp who will need both "present" and "historical" information, in separate parameters. Not P17 (yet), but P131. -- Lavallen (talk) 17:17, 12 January 2014 (UTC)

Bad translation[edit]

In catalan language there is a bad translation. Country is translated as State. I tried to fix it, but it is protected. Please, correct the error. The correct translation would be País. Thnak you. Gavanzo (talk) 00:24, 14 January 2014 (UTC)

Country property for international treaties?[edit]

Is this property the correct one to be used to list those countries which have signed a treaty? I would think not, that would obviously violate the single value rule, but is there any other property yet? So far the only item which lists the participating country is Anglo-Dutch Treaty of 1824 (Q224794), but that used the equally wrong has part (P527). Do we need a new property for this? Ahoerstemeier (talk) 14:16, 23 January 2014 (UTC)

The single value rule is often violated, and I do not think it should be considered a large problem to have such exceptions. Norway–Russia border (Q2661853) have two values for example, and I cannot see that it could be otherwise. -- Lavallen (talk) 15:27, 23 January 2014 (UTC)

Description[edit]

Shouldn't the description be "sovereign state of this geographical feature/location"? --SamB (talk) 02:31, 26 February 2014 (UTC)

Can we change this constraint to "value type"?[edit]

The property constraints for this item have lots of cases to go through and decide whether they belong on our list of countries. Is there an easier way of doing this? It seems like it should be possible to say that objects must be a certain type. Is there any class, in English or another language, that includes sovereign states and dependencies, but does not include constituent states? --Arctic.gnome (talk) 17:27, 9 April 2014 (UTC)

At the current time, no this is not possible since no single Q item groups all allowed "countries", and template:Constraint:Value type only supports one parent. However the good news is that statements on properties will be supported very soon; the domain might then be specified by an explicit API request, giving much more flexibility than the existing scheme. For example, a domain comprising sovereign state (Q3624078) and country of the Kingdom of the Netherlands (Q15304003) and member of (P463) United Nations (Q1065) could be specified like this:
Domain: "claim[31:3624078] or claim[31:15304003] or claim[463:1065]"
Stay tuned. - LaddΩ chat ;) 19:28, 9 April 2014 (UTC)

This should be sovereign state[edit]

"Country" is not synonymous with "sovereign state. eg. Ireland is a country, Eire is a sovereign state. Northern Ireland is just as much a part of Ireland, indeed people from Northern Ireland can claim Irish citizenship. Country is geographical concept with political aspects.Leutha (talk) 23:34, 12 January 2015 (UTC)

Ireland (Q22890) is not a country. Northern Ireland (Q26) is (though it is not a sovereign state), and Ireland (Q27) (which is also a sovereign state) is as well. Ireland (Q22890) is an island on which both of these entities exist. This database is supposed to mostly embody things written on the various Wikimedia projects. See en:w:Ireland_(disambiguation). I agree that it should be "sovereign state", though, seeing as Northern Ireland (Q26) is a country but not a sovereign state. --BurritoBazooka (talk) 05:39, 20 September 2015 (UTC)
I don't think it should be sovereign state. There are already lots of things explicitly allowed above which are not sovereign states, we can already say which items are sovereign states and which aren't via instance of (P31) on the relevant items and we can say which item something is part of via located in the administrative territorial entity (P131), so there's no need for this to be overly restrictive. In fact, I would go in the opposite direction and allow pretty much anything with an ISO 3166-1 country code (which would include overseas territories and (more of the) highly autonomous places) instead of picking and choosing which ones are autonomous enough and which aren't (e.g. why is Greenland allowed but not Hong Kong?), because this is a generic property used for a wide variety of things and sometimes it makes sense to allow a broader range of values. - Nikki (talk) 09:23, 19 November 2015 (UTC)

Should countries have statements with country: themselves?[edit]

For example, the following statements currently exist:

Are these correct? They seem redundant and not really reasonable to me. —DSGalaktos (talk) 09:18, 31 January 2015 (UTC)

It depends on how you read the statements.
  • If you read "Germany is located in 'Germany'", this doesn't seem logical. (P:P131 would imply that)
  • However, the meaning of P17 is broader: If you read: "Q183 relates to the sovereign state 'Germany' ", this makes sense. It helps to determine which country the statements on Q183 relate to.
This makes it easy to allocate statements to countries that can appear on all types of geographical entities, such as timezone, "list of monuments", etc.
Hope this is makes it clear. --- Jura 10:24, 31 January 2015 (UTC)
Okay, that makes sense, thanks. I added a statement to United States of America (Q30) as well. —DSGalaktos (talk) 10:49, 31 January 2015 (UTC)
Sorry Jura and DSGalaktos, but it doesn't seem logic to me.
My problem is not with country (P17) (which is obvioulsy not located in the administrative territorial entity (P131)) but the the self link itself.
Germany (Q183)country (P17)  Germany (Q183) can be read as « Q183 relates to the sovereign state Q183 » or « 'Germany' relates to the sovereign state 'Germany' » but not « Q183 relates to the sovereign state 'Germany' » (if you make a distinction beetween Q183 and 'Germany' then 'Germany' is not Q183 but an other item).
Plus, I don't see how it help nor can I see a case where is easier to say « something relates to itself ». Could you be clearer on the usefulness of this self link?
Cdlt, VIGNERON (talk) 18:42, 14 February 2015 (UTC)
These selflinks does not make sense for me. --Jklamo (talk) 21:25, 14 February 2015 (UTC)
For the reasons why this self-link was introduced, see the previous discussion at Property talk:P17/Archive#Add a self-link to sovereign states? --UV (talk) 21:48, 14 February 2015 (UTC)
Sorry but I'm still not convinced. Maybe I miss something or misunderstand the whole thing but it seems to me that this self link is just an illogic hack which is not needed.
Jura : « It helps to determine which country the statements on Q183 relate to. » make no sens to me. Q183 doesn't relate to a country, Q183 is a country and P31 is the easiest, logicest and cleaniest way to indicate it. How a self link could be « an elegant way to describe that an item describes a sovereign state » (as UV said in August 2013) when there is plenty other ways to do it with existing properties and without breaking a basic rule of common sens and database organisation ?
Cdlt, VIGNERON (talk) 20:33, 15 February 2015 (UTC)
Other than that some don't like them, are there cases where having them generates problems? --- Jura 11:50, 16 February 2015 (UTC)
It's not that I don't like them, it's that I like logic and common sens. There maybe a logic behind that but noone explain it clearly to me.
A self link is kind of a problem in itself that why is a contraint that shouldn't be violated without reason (and accepting them without good reason is a slippery slope to others self link, like Albert Einstein (Q937)instance of (P31)  Albert Einstein (Q937) - wich look exactly analog to Germany (Q183)country (P17)  Germany (Q183) to me).
Redundancy is a problem too. If you want to indicate that France (Q142) is a sovereign state (Q3624078), why not simply use France (Q142)instance of (P31)  sovereign state (Q3624078) ? It seems easier and more explicite to me than a self link.
Cdlt, VIGNERON (talk) 19:19, 17 February 2015 (UTC)
Given the odd "Einstein" example, I take it that it has no practical implications for you. --- Jura 21:01, 17 February 2015 (UTC)
Odd? The whole idea is odd, therefore Vigneron's example is odd as well. Because there are a dozen countries which are a NUTS unit as whole, these twelve countries apparently showed up in the constraint violation list of NUTS code (P605), and the self-link was proposed as a solution. But the self-link is added to other countries like Germany etc. as well, and that makes no sense at all. Marking a country as a country (for the goals Jura mentions like timezone, list of monuments etc.) can be done with P31, as said before. Far more elegant.
Jura also mentions P131, that is not useful in these cases either.
Let's quote Bjung: "That's a very bad idea that no experienced database manager would ever implement. Technical problems should not be solved by polluting the database with properties that are useless for the end user. Such kludges are also difficult to remove later when users get accustomed to them and programs depend on them." Bever (talk) 23:45, 8 June 2015 (UTC)

Maybe fix the NUTS code constraint or NUTS code data layout if that caused a constraint violation instead of allowing self links in the form "country of country X is country X"? Eldizzino (talk) 12:00, 7 July 2015 (UTC)

Add me to the list of those who don't understand the rationale behind this. I read the discussion at Property talk:P17/Archive#Add a self-link to sovereign states?, and the only reasons given there are:

Can someone please elucidate me? --Waldir (talk) 01:14, 24 July 2015 (UTC)

The problem arises with any property that can apply to items for geographic locations of different levels including countries. --- Jura 05:27, 5 September 2015 (UTC)

Those quirks should be removed, if a constraint really need this, the constraint should be fixed. Akeron (talk) 08:10, 1 August 2017 (UTC)

P17 in Antarctica[edit]

I have seen statements like: "P17:Argentina" in Antarctica, and to be onnest, it feels wrong. I can accept that research-stations and the handfull of settlements who exists there uses such statements, but for islands and mountains etc, it looks terribly wrong.

How about located in the administrative territorial entity (P131)/location (P276):Argentine Antarctica (Q646243) instead? -- Innocent bystander (talk) 13:54, 29 November 2015 (UTC)

I think I once read something that suggested that the difference between Argentine stations and those of any other country is that Argentinians sit there mainly to assert their territorial claim ..
Personally, I think it would simplify things if we used P17="Antarctic Treaty area" for Antarctica. --- Jura 14:30, 29 November 2015 (UTC)
Agree! -- Innocent bystander (talk) 14:37, 29 November 2015 (UTC)
I made a corresponding item: Antarctic Treaty area (Q21590062) --- Jura 16:46, 29 November 2015 (UTC)
That seems completely wrong to me. "Antarctic Treaty area" isn't a country any more than Antarctica is. I also don't understand the point of doing that. We already have continent (P30) which can be set to Antarctica (Q51). Using "Antarctica Treaty area" doesn't link it to the country either, which was presumably the whole point of setting P17 to Argentina in the first place. If you really want to say that it doesn't belong to a country, we can already use the special value "no value". - Nikki (talk) 05:11, 30 November 2015 (UTC)
It's a sort of condominium (Q734818) with country codes and other things. As you mentioned elsewhere, using continent creates various systematic problems. --- Jura 06:34, 30 November 2015 (UTC)
@Jura1: Lots of things which aren't sovereign states have an ISO code. Most of them are currently not considered valid values for this property, so the existence of an ISO code does not automatically mean an item should be used with this property. The only issues I've had are not with the data itself, but with the constraint reports being inflexible. The solution to inflexible tools is never to change the data, it's to fix the tools. Even if we do want to add P17 to make it clear that something in Antarctica is not in any country, we already have novalue for saying that. I don't see any reason why "Antarctic Treaty area" should ever be used as a value for P17, in the same way that I don't see any reason to set all items on the Moon to "country: Moon Treaty area" or all items in outer space to "country: Outer Space Treaty area" - none of them are countries, just areas defined by treaties which some countries have agreed to. If we want to say something is located in the Antarctic Treaty area, location (P276) would be the appropriate property. - Nikki (talk) 11:59, 13 March 2016 (UTC)
ɾ Ok, sounds like we haven't solved the problems you had at the same time. P276 can seem suitable, but we use that mainly for events and movable objects. As far as I can tell, codes for the treaty area are the only ones that refer to a condominium, so it was a good candidate for P17. This is similar to subnational entities used in P131 that are administrated by some central or joint entity, but that are not the constitutent parts of sovereign states. Using P17:Argentina can be tempting - as Argentina is a country -, but P1336=Argentina reflects better the generally agreed view on the status of these areas. --- Jura 08:17, 2 April 2016 (UTC)
While we are talking about Antarctica: This document (page 17) from the Norwegian Goverment can be of interest. From the maps on Commons, it looks like the extent of the Norwegian claim is not well defined. This document partly confirms that, but it also states that the Norwegian goverment do not oppose any interpretation of their claim, that it could include of all the area from the Coast to the Pole, including the Pole itself.
"Norske styresmakter har difor ikkje motsett seg at nokon tolkar det norske kravet slik at det går heilt opp til og inkluderer polpunktet."
-- Innocent bystander (talk) 08:40, 30 November 2015 (UTC) (And no, I am not Norwegian, and I do not support any claim of Antarctica, no matter its purpose.)
Supposedly we could include the Norwegian claim in P17 as well (we currently do for a few others, even one of overlapping territorial claims), but including a distinct value in P17 would indicate more clearly the nature of the area. It's not in Norway (or in France or whatever), just because some national statistical authority included it. --- Jura 06:58, 1 December 2015 (UTC)
"we currently do for a few others". Where?
Overlapping areas is not a big deal, as you say, we have many such.
If I remember correctly, there are areas within these Antarctic claims that are not affected by the Antarctic Treaty. -- Innocent bystander (talk) 08:42, 1 December 2015 (UTC)
Have a look at the items listed with "applies to part" at Q21590062#P1336. --- Jura 08:48, 1 December 2015 (UTC)
It seems that quite a few P1336 values need to be defined. Except for installations, probably not much use to duplicate them in P17.
--- Jura 09:14, 3 January 2016 (UTC)
  • One of the questions brought up recently is if ATA would be "a region that is identified as a distinct entity in political geography". Clearly it's not physical geography. We (Innocent bystander and me?) do have now the issue that people added P30 with the continent to places that are not located on the continent, but only part of ATA. [1]
    --- Jura 16:20, 4 July 2016 (UTC)
    I think it is better for us to have a consensus to use "country:novalue" than having repeated questioning of "country:ATA"-statements. My main issue here is that I do not want to see statements like "country:Norway/UK/Australia/Chile/etc" in the Antarctic. I am open for adding it in the few settlements that actually exists here, but otherwise not. And of course, adding Continent:Antarctic is an option, but you first have to check if the place really is on the continent. Many Lsjbot-items are related to places outside of the coast, some even on the bottom of the sea. -- Innocent bystander (talk) 18:20, 4 July 2016 (UTC)
    • People question what is new to them, no problem with that. Wikidata is new. Many people may also discover the treaty that way. There isn't really any issue with that. We add this by consensus so lets attempt to evaluate if it meets the suggested criterium "distinct entity in political geography". Interestingly, if one reads on that page it can include ATA. In any case, we need a way to identify locations that are within ATA. I can't really think of a better way to do that. Using P30 leads to the problem highlighted by here.
      --- Jura 19:04, 4 July 2016 (UTC)
The Antarctic Treaty Area is not a country. If anyone can find a source stating otherwise, I'll be happy to change my mind, but in the meantime, we should be using "country:novalue" for all locations in Antarctica. Kaldari (talk) 20:05, 6 July 2016 (UTC)
@Kaldari: In your opinion, is it possible to classify ATA as a "distinct entity in political geography"? Jura1's description of it as as a condominium (Q734818) looks valid to me! My opinion is that all ISO 3166-areas can be a "distinct entity in political geography". (Even if the entity "Svalbard and Jan Mayen" looks very fabricated to me.) That said, I am not 100 % convinced that this is a perfect solution. If > 75 % of all users who come across such a statement will remove it, it is probably waste of time to add and watch such statements. -- Innocent bystander (talk) 05:27, 7 July 2016 (UTC)
@Innocent bystander: "Distinct entity in political geography" is not a valid definition of "country". (Yes, I know that's the definition that Wikipedia uses.) A country is the territory of a nation or state (according to pretty much any dictionary or native English speaker). I have no opinion on whether it's a condominium (Q734818). There is no book, journal, website, or database that calls the ATA a country besides Wikidata. According to [Wikidata:Verifiability]], claims must be verifiable in secondary sources. Kaldari (talk) 17:03, 7 July 2016 (UTC)
ISO 3166-1 includes in countries. To make it clear that it should be used as value we could just add note it on the description of P17.
--- Jura 04:15, 8 July 2016 (UTC)
I agree with Kaldari. ATA is not a country. We should use P17=novalue and P131=Argentine Antarctica (Q646243). --Pasleim (talk) 19:53, 7 July 2016 (UTC)
Using the special value "novalue" doesn't define that a given location is located within ATA. We still need a structured way to do so. Merely changing the label of Q21590062 to "novalue (ATA)" could achieve that. An alternative could be to create a new property for condominia.
--- Jura 04:15, 8 July 2016 (UTC)
@Kaldari, Jura1, Pasleim: That ATA isn't a country in the meaning "the territory of a nation or state", I fully agree. From what I can see, the item about ATA does not claim that. In my opinion, adding "country:ATA" therefor do not add anything against any verifiability-policy. The problem I see here is Principle of least astonishment (Q22668), not that the claim adds anything against any good source. The meaning of "P17:Q123456" is not decided by the P17-label in English, especially not when it is contradicted by the English description, the label in many other large languages, and the constraints. -- Innocent bystander (talk) 07:12, 8 July 2016 (UTC)
The meaning of a property is defined by its labels and descriptions! All the labels and descriptions of P17 I'm able to read mention the word "country" or "state". As there is no source stating ATA is a country or a state, the claim "P17:Q123456" adds false information and is against any verifiability-policy. --Pasleim (talk) 08:10, 8 July 2016 (UTC)
I think we agree that it's verifiable that these places are located within the treaty area, that it is a structured way of expressing this and that the applicable country codes are ATA or AQ. The question is now what would be the best way to structure this. I think we are all open to reasonable suggestions.
--- Jura 08:25, 8 July 2016 (UTC)
@Jura1: What about located in the administrative territorial entity (P131):Antarctic Treaty area (Q21590062)? Kaldari (talk) 17:39, 8 July 2016 (UTC)
That sounds odd to me and it doesn't seem like the normal use of that property, but I can't quite explain why. I would rather use location (P276) over located in the administrative territorial entity (P131). - Nikki (talk) 19:05, 8 July 2016 (UTC)
P276 would be inconsistent with other uses of that property (see discussion above). Given that ATA is a "Distinct entity in political geography", it should go into P17 or maybe P131. P17 has the advantage that it's equivalent level to most other values used with that property, but maybe P131 could do as well. Those Wikipedia versions that make use of it seem to use it as P17 (e.g. nnwiki). A simpler solution could be to improve labels and descriptions of the current property/value. For those who prefer not to follow ISO or nnwiki, can we have samples and definitions for the suggested alternate approaches?
--- Jura 18:09, 10 July 2016 (UTC)
P131 looks more correct than P276... -- Innocent bystander (talk) 18:38, 10 July 2016 (UTC)
So far P276 isn't used for geographical objects, so I would really go with P131. --Pasleim (talk) 08:52, 21 July 2016 (UTC)

OK, so it sounds like the consensus is to replace country (P17):Antarctic Treaty area (Q21590062) with country (P17):no value and located in the administrative territorial entity (P131):Antarctic Treaty area (Q21590062). Does that sound correct? Kaldari (talk) 13:09, 27 July 2016 (UTC)

As far as I can see, Yes! If you intend to fix this, feel free to close the thread on "Bot requests" about this topic! -- Innocent bystander (talk) 13:48, 27 July 2016 (UTC)
The problem with "no value" is that it's virtually impossible to edit efficiently. Can we do an item instead?
--- Jura 13:54, 27 July 2016 (UTC)
Kaldari offered to fix it with his bot. So you don't have to work. Moreover, it doesn't take more time to add a novalue instead of an item. --Pasleim (talk) 14:01, 27 July 2016 (UTC)
I already did all the work so: can you illustrate your last point with PetScan, QuickStatement, Autolist?
--- Jura 14:04, 27 July 2016 (UTC)
We shouldn't adapt the data structure of Wikidata to functionalities of external tools. Also quantities with units and monolingual texts aren't supported by Magnus' tools but still we use them.
I agree with Pasleim. External tools should adapt to Wikidata, not vice versa. Kaldari (talk) 18:09, 13 August 2016 (UTC)
Currently we are following an approach consistent with ISO. If we change this, we would need to provide a reference for it. Can we see this before any change?
--- Jura 14:07, 27 July 2016 (UTC)
Where was consensus made to follow the approach by ISO? And if there was consensus why weren't places like Norfolk Island (Q31057) or Bouvet Island (Q23408) not added to the above country list? --Pasleim (talk) 11:18, 31 July 2016 (UTC)
The current approach was discussed last year. If you think we need to check P17 for other regions, this should obviously be discussed as well. -- Jura 05:33, 14 August 2016 (UTC)
I didn't ask when it was discussed but where. Please point me to the discussion. --Pasleim (talk) 07:51, 14 August 2016 (UTC)
Ah .. here. -- Jura 15:41, 21 August 2016 (UTC)
We don't provide references for any of our current statements that places belong to the country of ATA, so I see no reason to require references for changing it. I also don't see how ISO is especially relevant here. The purpose and scope of ISO codes is different than the purpose and scope of P17. Kaldari (talk) 18:04, 13 August 2016 (UTC)
It's normal that you have to provide a reference for your change. I can understand that it can be tempting to sign up to do some change without being able to substantiate it, but that's not what how it works. Please research your suggested approach and provide the requested reference for it. -- Jura 05:33, 14 August 2016 (UTC)
@Jura1: You want me to find a source for every place in Antactica that says that it isn't located in a country? There are no such sources. Would you prefer that I just remove all the country claims and leave none? Kaldari (talk) 20:49, 14 August 2016 (UTC)
A reference for the values that illustrates that it's consistent with the scheme you advocate for Treaty Area places. Similar to the reference and the definition we provided for the current approach. -- Jura 15:41, 21 August 2016 (UTC)

@Jura1: Here is a reference:

  • Saul, Ben; Stephens, Tim, eds. (2015). Antarctica in International Law. Portland, Oregon: Hart Publishing. p. 885. ISBN 1849467315. "Antarctica can properly be characterized as something of an international anomaly. It is a large continent which has never been and is not now subject to the sovereignty of any nation. Under the Antarctica Treaty of 1959 the signatory nations agreed not to exercise sovereignty in Antarctica, although their claims to sovereignty were not extinguished." 

Kaldari (talk) 21:44, 21 August 2016 (UTC)

  • Reading the context, is this a US judicial view for the 1970s focusing on its installations on the continent in relation to some US act? It's not clear how this supports your approach for ATA in 2016 in general. It might mean they'd use P17=United States of America (Q30) or at least P1336=Q30 for areas with installations. Interesting find. I agree with what Innocent bystander wrote earlier: we shouldn't add such statement in P17 (even it's the US here). It could support adding Q30 to P1336 of some items. -- Jura 17:43, 30 August 2016 (UTC)
    • @Jura1: It supports the approach that was agree upon because it says that no country has sovereignty over any part of Antarctica. It goes on to say "Antarctica is not a foreign country, it is not a country at all; and it is not under the dominion of any other foreign nation or country." No one has suggested that we add statements about any US claims, so I don't know why you're creating a red herring about that. Kaldari (talk) 00:36, 31 August 2016 (UTC)
      • The problem is that it goes on to say that the US makes such claims and it is an official US view of the 1970s. This doesn't match the current condominium view and I don't think it's sufficient to change from the view we agreed on last year and currently implemented.
        --- Jura 21:22, 7 September 2016 (UTC)
        • @Jura1: How does the fact that the US has made claims on Antarctica mean that the ATA is a country? That's a complete non-sequitur. Even if Antarctica was a condominium, that doesn't automatically make it a country. The Gulf of Fonseca is also a condominium, but it's obviously not a country. So far, no one else here agrees with your position and you haven't done much to build any kind of alternative consensus or present any references backing up your position that the ATA is a country (other than citing a convoluted definition of "country" from Wikipedia, which isn't a reliable source, and using original research to argue that that definition makes the ATA a country). At this point, it doesn't seem like you are really interested in building a new consensus, but are just throwing up roadblocks to obstruct the current consensus. Kaldari (talk) 09:00, 12 September 2016 (UTC)
          • ATA could be used as P17, if we form a consensus to do that! "Country" as a property does not have to fit any definition in any dictionary in any language. We are here trying to form a relation between two items, not to replace the function of international law. It would be useful if the use of P17 could fit how "country" is used as a parameter in Wikipedia. Currently I do not think P17 does that. "Greenland", "Puerto Rico" et al are there often used. But if we demands such a thing as "sovereign state" we cannot use such items as ATA and Greenland in P17. -- Innocent bystander (talk) 09:18, 12 September 2016 (UTC)
          • I think a condominium fits well the definition brought up earlier in the discussion of the request for P17: "region that is identified as a distinct entity in political geography" (not by me BTW). As it's the only definition brought up so far and the US view for their potential territorial claims in the 1970s seems hardly sufficient as a reference for a solution in the 2010s, I think you primarily fail to provide what we expect from users.
            --- Jura 10:14, 12 September 2016 (UTC)
            • @Jura1: So you're saying that things like the Gulf of Fonseca, Pheasant Island, Greenland, Puerto Rico, and maybe even the Pacific Ocean (which could be considered a condominium) should all be listed as countries even though no other source on the planet calls them countries? That doesn't seem sensible. If you have no actual sources to present to support this argument, I don't think you're going to convince anyone. Kaldari (talk) 19:26, 12 September 2016 (UTC)
  • The quote on p. 885 of Saul & Stephens is a bit of an odd one to pick - as noted, it's a quote from a US court ruling in the 1980s, dealing with a specific point of US law (is Antarctica included in "foreign countries"?). It's not particularly wrong, and it's certainly not a US-only interpretation as has been suggested, but it's a bit limited in context - a better summary of the situation is from the ATS:
Among the signatories of the Treaty were seven countries - Argentina, Australia, Chile, France, New Zealand, Norway and the United Kingdom - with territorial claims , sometimes overlapping. Other countries do not recognize any claims. The US and Russia maintain a “basis of claim”. All positions are explicitly protected in Article IV, which preserves the status quo:
"No acts or activities taking place while the present Treaty is in force shall constitute a basis for asserting , supporting or denying a claim to territorial sovereignty in Antarctica or create any rights of sovereignty in Antarctica. No new claim, or enlargement of an existing claim to territorial sovereignty in Antarctica shall be asserted while the present Treaty is in force."
In effect, under the Treaty, those countries which have a claim suspend it without losing the claim; those countries which reserve the right to make a claim continue to do so; and those countries which don't make a claim agree not to make any new ones. This is all a bit fuzzy and doesn't really map perfectly to the way we structure things on Wikidata, but it seems the P131 approach works regardless of what we interpret the Treaty Area to be - it's an area, yes, but it's certainly not a country and it's not quite a condominum or a disputed territory, either.
I agree that "country: Treaty Area" or "country: Antarctica" would also make sense if we defined these as a special case for P17. Really, I'd be happy with either - we just need to pick one, declare it standard, and get on with it. P131 has the benefit that it avoids anyone saying "oh, it's not really a country, remove" later (which will inevitably happen somewhere), and it's unarguable that *some kind* of administrative concept exists; P17 has the benefit we don't have to mess around with novalue properties and it's consistent with the way we handle every other landmass in the world. Neither is perfect, but either is defensible, and we shouldn't wait around for years trying for conceptual perfection. Andrew Gray (talk) 22:20, 13 September 2016 (UTC)
We actually did pick one last year and implemented that. Changing it now just complicates things going forward.
--- Jura 07:15, 14 September 2016 (UTC)


Proposal for solution I[edit]

The argument above has gone in circles for months. I think it's time to assess whether or not we have consensus. I propose the following solution:

Please indicate below whether you support or oppose this proposal. Pinging interested parties: Pasleim, Jura, Innocent bystander, Nikki, Andy Mabbett, Thryduulf, Ymblanter. Kaldari (talk) 19:39, 12 September 2016 (UTC)

  • Symbol support vote.svg Support, since this is the solution which most people seem to agree with in the discussions above and it seems like a sensible approach. Kaldari (talk) 19:42, 12 September 2016 (UTC)
  • Symbol support vote.svg Stöder I think there are other ways to do this, but this is a solution we can find a consensus around today. -- Innocent bystander (talk) 19:44, 12 September 2016 (UTC)
  • Symbol support vote.svg Support since I would never call Antarctic Treaty area (Q21590062) a country. --Pasleim (talk) 19:52, 12 September 2016 (UTC)
  • Symbol support vote.svg Support - this looks like the cleanest solution but I think we should also add some clear definitions on how to handle edge cases - comments below so not to derail this. Andrew Gray (talk) 20:18, 12 September 2016 (UTC)
  • Symbol oppose vote.svg Oppose - Wikidata shouldn't adopted the US view for the 1970s. Stating that's it's no man's land is just erroneous. The treaty area as a condominium matches the agreed definition "region that is identified as a distinct entity in political geography". As we still haven't seen a valid reference for the suggested model, I don't see what justification we would have to adopt it. We should definitely retain all country claims in the region. P1336 has been made for this specific reason.
    --- Jura 22:23, 12 September 2016 (UTC)
    • I agree that we should leave all P1336 country claims. Kaldari (talk) 04:54, 13 September 2016 (UTC)
  • Symbol support vote.svg Support One of good solutions for this delicate problem. --Jklamo (talk) 23:59, 12 September 2016 (UTC)
  • Symbol support vote.svg Support. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:32, 13 September 2016 (UTC)

Proposal for solution II[edit]

The proposed definition for P17 is "region that is identified as a distinct entity in political geography". The Antartic Treaty Area is generally considered a Condominium. The solution we currently agreed on is the following:

This use is consistent with country lists by various organizations. Please indicate below whether you support or oppose this. -- Jura 00:10, 13 September 2016 (UTC)

  • Symbol support vote.svg Support -- Jura 00:10, 13 September 2016 (UTC)
  • Symbol oppose vote.svg Oppose -- I like this idea, I even prefer it. But I cannot see that we can form a consensus around it when we are so focused around the literal meaning of the label of our properties. -- Innocent bystander (talk) 05:22, 13 September 2016 (UTC)
    • It depends on how you read "country: none": I think most agree that it's incorrect that this area can be appropriated by any state. -- Jura 10:12, 13 September 2016 (UTC)
  • Symbol oppose vote.svg Oppose Antartic Treaty Area is not a country. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:33, 13 September 2016 (UTC)
  • Symbol oppose vote.svg Oppose since I would never call Antarctic Treaty area (Q21590062) a country. --Pasleim (talk) 17:07, 13 September 2016 (UTC)
  • Symbol oppose vote.svg Oppose - I don't agree with defining the ATA as a country. Kaldari (talk) 23:47, 13 September 2016 (UTC)

Suggested definitions & related properties[edit]

I feel like I should have put my Antarctic hat on and commented on this earlier - this is dangerously close to my day job :-). As noted above, there are some edge cases it would be good to have a clear line on.

  • Nationally-associated places (mostly research stations and other things with an owner). We need some way to say that Neumayer-Station III (Q7933) is associated with Germany (Q183). We should probably use operator (P137) for these, either listing a country or an organisation - this seems to be what embassies do and that's a good analogy.

If people are happy with these, I can start putting together a script to run a constraint report for them. Andrew Gray (talk) 20:18, 12 September 2016 (UTC)

If this is linked to your day job, can you advise us about any territorial claim your employer may have?
--- Jura 22:25, 12 September 2016 (UTC)
@Andrew Gray: Those all sound like reasonable suggestions to me. If you can generate a list of all the locations south of 60S, I would be happy to use it. Kaldari (talk) 04:57, 13 September 2016 (UTC)
I believe they rent an office in Cambridge. HTH. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 15:35, 13 September 2016 (UTC)
  • @Jura1: - I work in scientific support for British Antarctic Survey (Q918757). The UK (the country, not the organisation) has a territorial claim in the Peninsula area (British Antarctic Territory (Q165783)). As with all the national claims, this claim is effectively suspended under the Treaty. I hope this isn't considered a COI - I am certainly not commenting here in any official capacity, just as someone who has to think about Antarctic issues daily and hopefully understands some of the issues :-)
Territorial claims are complex and I don't pretend to understand the nuances fully, but I can't think of a better way to deal with them than P131, other than leaving them off Wikidata entirely, and I feel that wouldn't be very useful. Either way, they're secondary to the core issue here.
I've added some comments above re the quote Kaldari found, which seems to have caused a bit of unnecessary confusion. Andrew Gray (talk) 22:11, 13 September 2016 (UTC)
I think it might have if you had supported solution II. How about @Kaldari:: do you have a conflict of interest?
--- Jura 07:11, 14 September 2016 (UTC)
My only conflict of interest is that I'm tired of talking about Antarctica and want to wrap this up one day :P Kaldari (talk) 07:22, 14 September 2016 (UTC)
I think the monuments in Antarctica could use some curating. I finished protected areas some time ago.
--- Jura 07:27, 14 September 2016 (UTC)
  • @Kaldari: - I've asked about a suitable SPARQL query. There's a sample one at query.wikidata.org (based on an arbitrary 3000km north of the Pole) but it feels like directly querying coordinate values would make a lot more sense than converting to distance. Once I've figured out how to do those I'll write a query for "everything below 60S" and one for each of the seven territorial claims. Andrew Gray (talk) 22:11, 13 September 2016 (UTC)
For future use reference, this is everything with P30:Q51 between -40S and -60S (seems to mostly be sub-Antarctic islands and a bit of South America; will probably remove all these at some point) Andrew Gray (talk) 21:02, 15 September 2016 (UTC)

One more thing that's just occurred to me - most (but not all) Antarctic places are currently tagged with continent (P30):no label (Q50), which seems to have originated as a way to get around the P17 problem. Once we have everything marked as "Antarctic Treaty Area", it feels like this won't be needed any more. I think we can safely remove them - thoughts? Andrew Gray (talk) 20:45, 16 September 2016 (UTC)

@Andrew Gray: Did you mean continent (P30):Antarctica (Q51)? That seems fine to me. Antarctica is a continent after all. There doesn't seem to be any controversy about that. Kaldari (talk) 23:39, 16 September 2016 (UTC)
@Kaldari: this is what I get for doing properties from memory - yes, Q51! It feels like this is really a hack to of express the same thing; it's not really consistent with the way P30 is used in the rest of the world and I'm not sure we need it alongside the P17/P131 solution (whichever one we go with). P30 is currently a bit strange - the definitions say it should be assigned to countries, places on two continents (eg Istanbul?), and anywhere in Antarctica. At the moment, some continents are reasonable (a few hundred items), some are a bit high (a few thousand), and Antarctica has 48% of all P30 values. But this is a discussion to have over there once we've sorted this bit out, I suppose :-). Andrew Gray (talk) 09:34, 17 September 2016 (UTC)
Summarised

I've put together a summary of the recommendations here for future reference at Wikidata:Antarctica. Andrew Gray (talk) 21:55, 23 September 2016 (UTC)

Akrotiri and Dhekelia[edit]

How should we do about Akrotiri and Dhekelia (Q37362)? I see Q60509 has Cyprus in P17, when it maybe formally is United Kingdom. Thoughts? @Tom Morris: who added that statement? -- Innocent bystander (talk) 13:26, 9 May 2016 (UTC)

Innocent bystander: I was just adding country properties based on category membership. It wasn't a grand political gesture and I have no strong opinions on this matter. Feel free to revert/remove the claim. —Tom Morris (talk) 20:21, 9 May 2016 (UTC)
It is always tricky when we have articles and categories that are dubious in their character, like both islands and nations. One essential question here is if Akrotiri and Dhekelia can be treated as a nation in its own, like the Channel islands, Gibraltar and the Indian Ocean Overseas Territories sometimes are. -- Innocent bystander (talk) 07:04, 10 May 2016 (UTC)
I think that locations in Akrotiri and Dhekelia (Q37362) should have P17 set to either Cyprus or United Kingdom - I have no preference which (maybe the former for civilian places and the latter for military?) - the areas are not a country themselves. Thryduulf (talk) 19:20, 4 July 2016 (UTC)

P17 with end dates but no replacement.[edit]

We have discussed if independent nations should have self-referencing statements like: "Canada: Country:Canada" or not. I do not remember what conclusion we came to, but my experience is that most such statements have sooner or later been removed.

A related question is how we should act with statements like this. The statement is fully true, but it is not replaced with anything, making it the most valid P17-statement in the item at the moment. Leaving the question if Western Sahara is a "country" independent enough to use with P17 aside for a moment, we have to solve this for many nations. Norway became fully independent in 1905 after almost a century as subject under Sweden. Some kind of statement have to be added next to "P17:Sweden, end date:1905". Should it be "P17:Norway, start date:1905" or "P17:novalue, start date:1905"? -- Innocent bystander (talk) 18:32, 4 July 2016 (UTC)

Well, even the genetics project has issues with people merging completely unrelated items. It's likely to happen.
At mw:Wikibase/Indexing/SPARQL_Query_Examples#List_of_countries_in_1754, I added a query based on the discussion on Wikidata:Project chat. I think we should attempt to make such queries work.
--- Jura 19:12, 4 July 2016 (UTC)
@Jura1: Was I unclear, how is that thread (which I remember) related to my Q above? -- Innocent bystander (talk) 05:14, 5 July 2016 (UTC)
If one starts out with that query for a given year, one should eventually be able to find parts of these countries.
--- Jura 04:19, 8 July 2016 (UTC)
But it was the P17-relations after 1976 for Western Sahara or after 1905 for Norway I was talking about. -- Innocent bystander (talk) 07:14, 8 July 2016 (UTC)
To query the first one for 1980, Q6250 needs an inception date and a second statement with P17 (P17=Q6250). For 1850, Q20 (or an item with a similar geographic scope) would need a P17-statement related to Sweden.
--- Jura 08:33, 8 July 2016 (UTC)
I think we need to work on better mapping of changes in geographic extensions .. your agglomeration proposal is a start.
--- Jura 21:25, 7 September 2016 (UTC)

P17 for places in Jersey (Q785)[edit]

What's the value to use? For some reasons, it should probably not be United Kingdom (Q145) or Normandy. Jersey (Q785)?
--- Jura 13:50, 10 January 2017 (UTC)

Use United Kingdom (Q145) with country (P17). You can use Jersey (Q785) with applies to territorial jurisdiction (P1001) or country for sport (P1532) or located in the administrative territorial entity (P131). --Arctic.gnome (talk) 18:23, 16 January 2017 (UTC)
@Arctic.gnome: Why would P131 be used here? What does that qualifier indicate precisely? --Yair rand (talk) 23:13, 18 January 2017 (UTC)
@Yair rand: We use P131 for subdivisions of countries. For example, the Empire State Building is P17 United States and P131 New York State. But P17 can be used for client states and vassal states. British crown dependencies are in a grey area that could be a P17 or P131 value under our current parameters. --Arctic.gnome (talk) 06:20, 19 January 2017 (UTC)
In the meantime I went ahead and used Q785 as "novalue" didn't quite work out. What do you base your recommendation for P17 on?
--- Jura 07:12, 17 January 2017 (UTC)
I would use Jersey for P17. We already allow various things which are not sovereign states for P17 and I think the British Crown dependencies should be added to that list, because they are not part of the UK (as stated in this government document). - Nikki (talk) 01:41, 19 January 2017 (UTC)
@Nikki: It looks like our current set up does allow British crown dependencies to be P17 values because they are a subclass of client state (Q1151405). I'm not sure if it would stay like that if we did a thorough review of our political geography hierarchy, but for now it works. --Arctic.gnome (talk) 06:20, 19 January 2017 (UTC)
Ok, as we seem to agree they should be there, I added them explicitly. As the hierarchy keeps getting changed without any explanation, we might otherwise loose focus what actually should be there.
--- Jura 06:32, 19 January 2017 (UTC)

Jersey is not part of the United Kingdom, which is clearly stated in w:en:Crown dependencies. Hence it is very sensible to use Jersey itself as the P17 value. --Anvilaquarius (talk) 15:19, 21 August 2017 (UTC)

Should countries have a country? (i.e. self reference)[edit]

There is a constraint on this property that says they should. But many countries do not, and some that did have had it removed. --99of9 (talk) 00:22, 5 August 2017 (UTC)

To use P17 in that way is a good way to show that you have reached the top level. You will not get any further. -- Innocent bystander (⧼Takpagelinktext⧽) 04:48, 5 August 2017 (UTC)
I also would prefer it. But it was removed from Australia (Q408), and when I checked United States of America (Q30) it wasn't there either. --99of9 (talk) 11:22, 5 August 2017 (UTC)
Sometimes I wished we had a way to comment a statement, like we do in templates at Wikipedia: <!-- Please, do not change here, because bla bla --> -- Innocent bystander (talk) 12:45, 5 August 2017 (UTC)
We had comment (DEPRECATED) (P2315), but it's been deprecated. I'm not sure how it would work, though, on a multi-lingual project. Andy Mabbett (Pigsonthewing); Talk to Andy; Andy's edits 17:40, 5 August 2017 (UTC)
One option could be to have item-datatype for those comments, allowing everybody to translate them as labels, (if they want to). reason for deprecation (P2241) is partly used in that way. -- Innocent bystander (talk) 18:54, 5 August 2017 (UTC)

Belongs to vs. situated in[edit]

I'm wondering how to use P17 for an Australian monument which is situated in theUK. From a "situated in" perspective P17 should be UK, but the monument "belongs to" Australia.

In guessing embassies have much the same problem. Any suggestions are welcome. /Lokal Profil (talk) 07:40, 2 September 2017 (UTC)

Well, if I remember correctly, operator (P137) is used for embassies to tell who "owns" the embassy, while P17 tells where it is located. Which nation is served by the embassy is another delicate problem. (For example embassies in Italy serving The Holy See and/or San Marino, or embassies in Spain serving relations between Andorra and the owner.) But I guess the latter is seldom a problem for a monument. -- Innocent bystander (talk) 08:26, 2 September 2017 (UTC)
Thanks! operator (P137) it is. / Lokal Profil (talk) 18:35, 2 September 2017 (UTC)