Consistency of ISO worldview (e.g. Crimea)

[andyjenkinson]

I notice that in the 10m country ISO worldview v5.1.1 Crimea is not present in the map. Please nobody misinterpret this issue as an invitation to re-debate the de facto boundaries policy, however as far as I know the ISO view is perfectly clear that Crimea is part of Ukraine with code UA-43. Likewise Sevastopol. Also missing from the ISO country map are many other areas such as Transnistria in Moldova and Guayana Esequiba but I am not as familiar with the ISO definition of those areas. So my first question is, is this expected? Since other disputed areas are included within their respective ISO country boundaries (e.g. Kosovo, Western Sahara).

I figured perhaps this is intentional and that what I am supposed to do is combine the ISO countries view with the disputed areas view, but I am a bit confused about how it is intended to work? In the disputed areas shapefile the ADM_ISO code for Crimea is -99 and there is no ADM_A3_ISO code, and I don't see an FCLASS_ISO code. The ADM_A3_WB and AD_A3_UN (I assume World Bank and United Nations?) are also -99.

Meanwhile Crimea is not called out separately in the map units dataset, and in the admin_1 dataset its FCLASS_ISO code is null (as it appears to be for all polygons), the iso_3166_2 code is UA-43 but the iso_a2 code is RUS. I'm unsure what these codes are intended to represent (natural earth's view of the ISO code, or ISO's) but it seems to be that one or the other is wrong.

So overall I am unsure how I ought to (programmatically) produce an ISO-aligned boundary dataset that covers all areas. This is important as I want to avoid taking a specific country's view of geography, which I expected to be a common goal. I am happy to contribute a fix but I am unsure how the build process works (e.g. are you automatically creating the ISO worldview using metadata from the disputed areas dataset?)

[rshopa]

This is by any means a "de facto boundaries policy", it is a deliberate declaration of @nvkelso's support of the terrorist state - Russian Federation, which has been discussed multiple times yet without avail. I have sent an official complaint to GitHub and probably should do this again, since after the horrific genocide and atrocities happening right now in Ukraine, sticking to such a "policy" is nothing else as a humiliation of victims and hypocrisy.

[andyjenkinson]

I understand that the de facto policy is controversial and will not use it. However as I made clear in the first sentence of the issue, this is NOT the same issue and is not about Crimea specifically. I really want to keep that issue aside, there are already other issues about that.

This issue is to do with the ISO worldview, and more generally in the usability of the alternative worldview system. It’s (I believe) a bug, not a feature, or at the very least an unexplained or incomplete feature. The de facto view or the choice to use it might be disputable, but the ISO view is not. The ISO worldview should contain every piece of land in the world, and (at the time I raised the issue) it does not.

The reason I want to ISO world view to work correctly is precisely because I do not support the stance taken by the project - I am trying to make an alternative independent global world view viable.

[dolynchuk]

Because of divergence with ISO standards, GeoPandas doesn't trust this dataset anymore.

They made a patch that brings divergence compared to naturalearth. To fix the map according to ISO. BUG PR in GeoPandas: https://github.com/geopandas/geopandas/pull/2670 Release note: https://twitter.com/geopandas/status/1601653591781085185?s=20&t=Mp2b4z-TrJPzcNuZrlcqCg

[WarthogARJ]

10m country ISO worldview

I don't see the issue here.

As long as you say you are "Classifying according to ISO 3166" then you need to follow what it says: explicitly. Right or wrong.

Bear in mind that there are several parts to it, as in ISO 3166-Part 2, classifies Taiwan as CN-TW: as a separate Territory with the overall CN (China) designation.

Transnistria is classed as part of Moldova by ISO.

HOWEVER, there is nothing to say one cannot have a Modified ISO list, as long as you say HOW you modified it. ISO is a series of STANDARDS, and if you follow them EXACTLY, you don't need to justify whatever you do. Just that it follows the given ISO Standard.

[andyjenkinson]

I'm not sure what it is you're trying to say or who your comment is directed to?

You say you "don't see the issue here" but I thought it was quite clear- it is that "country ISO worldview" isn't a worldview at all, because it doesn't contain all the parts of the world. It is missing many territories entirely. You say that Transnistria is classed as part of Moldova by ISO, yet my use of it as an example is because Transnistria does not exist at all according to the natural earth ISO worldview.

This is because the ISO worldview is just broken (or was, I haven't checked Transnistria in the latest version but I can see others are still missing). It has nothing to do with me disagreeing with the ISO worldview or somehow wanting a 'modified ISO list' - the issue is that the natural earth ISO worldview is clearly and unequivocally not representative of the ISO 3166 standard.

In terms of how these missing territories should be included in the map, it's a separate issue which the worldview should of course define. Yes, there are multiple parts to the standard but it is still necessary to follow a logic. It seems fairly obvious that the countries dataset should first and foremost represent all the countries, and so for ISO that would obviously mean each country as recognised by ISO (i.e. 3166-1) is, well, a country. Taiwan is one, to use your example. It's possible there are areas with multiple 3166-2 codes and no 3166-1 code, but I don't know any offhand and there is no dispute within any ISO standard about which country Crimea and Sevastopol are part of - there are no codes for them as a subdivision of Russia, and there are codes for both as subdivisions in Ukraine. In the ISO 'worldview' of countries, Crimea is part of Ukraine, and certainly is not a hole in the world.

[nvkelso]

The discussion here has veered into uncharted territories, as it were. I'm locking comments on this issue for 6 months and will start blocking GitHub users who can't maintain decorum.