Clarify capital info for changed capitals

[emiham]

For capitals that have recently changed (e.g. Gitega), or changed name (e.g. Nur-Sultan), there should perhaps be a limit to how long this is indicated in the note. It's not a problem in and of itself, but I found myself second-guessing myself when translating them, and going into the future I think it would be a good idea to have easy to follow rules for translators and other contributors.

[axelboc]

Do you mean that the old name/capital being mentioned in the capital info makes it difficult to learn the new one? Hmm yeah, I see your point. I'm not sure we'd be able to come up with a clear guideline, though, as each case is different.

We'd have to weigh the learning difficulty you mention against the general knowledge value of keeping an old name/capital in the capital info. This value probably depends on many factors, such as:

• for how long the old name/capital was used;
• whether the change is "official" or specific to Wikipedia (i.e. because the country has no official capital and people disagree on which city deserves this status the most);
• whether the old capital remains significant (e.g. for economic or historical reasons);
• etc.

[emiham]

Do you mean that the old name/capital being mentioned in the capital info makes it difficult to learn the new one?

I don't have any thoughts on that, I'm just concerned about consistency for ease of contributing. The oldest ones in the English version are only from 2019 I think, but the Norwegian translation has one from 2005. There are a few that have mentions of a name change on Wikipedia, but I chose to not add any new fields for them for now. An example of this would be Reval → Tallinn, which is probably a good example of something that doesn't need to be included since it seems to have changed mid-20th century.

I do agree that it's hard to have a clear guideline for this though.

[axelboc]

Oh right, I get you.

Seems to me that the problem is more that the current translations are not as consistent with one another as they should be. It's more noticeable now that each field has its own CSV file. For instance, the Norwegian translation does indeed mention that the capital of Myanmar has moved from Yangon to Naypyidaw in 2005 (btw, English Wikipedia gives 2006), but this information is not mentioned in any other translation, which is a shame!

So I don't think the solution is to define a time limit for old capital names to appear in the capital info field, but to agree on a case-by-case basis on which old name is worth mentioning and to make sure that it's mentioned consistently in every language.

For your translation, just stick with translating the English deck without worrying about other translations. That being said, please do take note of any inconsistency you encounter, or any information you think is missing in the deck, and open issues for them so we can work on fixing them!

[axelboc]

I should have looked at your PR first, sorry. 😄

Looking at some of the old capital names that @aplaice's mentions, there's definitely a need for more precise guidelines. For instance:

• An old colonial name is most likely common to every language, so I reckon it should be mentioned/not mentioned consistently across languages.
• Same goes for capitals that move to other cities.
• A spelling change, however, which could be specific to one language, is rarely worth mentioning at all in my opinion. The only exception would be if the old name remained in common use (i.e. if the official name differs from the common name), but in this case, because of Wikipedia's guidelines, it's probably the common name that would be mentioned in the capital info, not the new official name 😄.

[aplaice]

AFAICT there seem to be four main, general cases:

1. The capital was moved from one location to another. (For instance, the capital of Burundi moving from Bujumbura to Gitega, and that of Myanmar from Yangon to Naypyidaw.)
2. The official name of the capital (or country) was "recently" changed. (For instance, Astana was renamed Nur-Sultan, Macedonia became North Macedonia and Zaire became DRC.)
3. There are alternative names for the entity. This can be due to alternative spelling or transcription (Laayoune vs. El Aaiún, Ulaanbaatar vs. Ulan Bator, Yamoussoukro vs. Yamassoukro), due to the country, city or region having a "traditional name" (e.g. Kaliningrad ("Ryska Östpreussen" in Swedish), Helsinki (Helsingfors in Swedish and partially Norwegian)) or due to alternative "arrangement" of the name (e.g. "Mikronesiens federerade stater" in Swedish or USA/US in some languages).
4. The entity had an old, alternative name. (Jeju (Quelpart), Sri Lanka (Ceylon))

2, 3 and 4 are not really distinct and I'm not sure I made the split correctly.

Many cases are "in-between", since "traditional names" often come about due to the official name changing, a long time ago, but not actually being used abroad. Burma was renamed Myanmar, but most of the world did not take note and still widely use the old name (unlike Zaire → DRC, where the only cases I hear are "formerly known as Zaire"...) The renaming of Kyiv also does not cleanly correspond to a single category it's a different transcription of the same Ukrainian name, but it's been changed due to an "official" campaign. Similarly for Czechia/Czech Republic (it's the same Czech word being alternatively, officially "arranged" in foreign translation...) Also, what is a "recent" change?

1 is usually language-agnostic, so we should apply a uniform policy.

2 is probably language-agnostic.

3 and 4 would need to be considered on a per-language basis.

---

For 3, my preference would be for not including alternatives that differ by a small spelling differences (unless the minor spelling difference was a major issue — e.g. Kyiv vs. Kiev), but including alternatives that are "significantly" different (e.g. Laayoune vs. El Aaiún.) The aim is that if the learner encounters the alternative name in the "real world" they'll be able to identify it as corresponding to the name they learnt in our deck! :)

If, subjectively, an alternative name is very unusual/not used, I'd also avoid mentioning it. (But why would Wikipedia mention it, if it's not actually used?)

For 4 I'd lean towards not including very old names, at all. (The two issues are:

a. What is "very old" and what is recent? (4 vs. 2)
b. How do we know that it's a former name, not an alternative name (4 vs. 3)

For a. I'd say that something from the colonial era is clearly "very old", while something from the 21st century is "recent", but I'm not sure about intermediate cases... We should probably use a similar time cut-off as whatever we decide for the moving of the location of the capital...

For b. I'd rely on Wikipedia. If it says "formerly known as" (e.g. Sri Lanka/Ceylon) or "ett äldre namn" (Jeju/Quelpart), then it's an old name (i.e. 4 or possibly 2). If it's X or Y (e.g. Myanmar/Burma) it's an alternative name (i.e. 3))

---

Edit: As a discussion starting point, perhaps we could use 20 years (i.e. the 21st century) as the cut-off for capital moves and entity renames?

The two main reasons to have this information at all, is that:

A. It's interesting and useful in its own right.
B. It helps to avoid confusing people who had previously learned the old locations or names, and, on encountering the new one in AUG would wonder why they remembered something different to what AUG was telling them.

For A. 20 years IMO makes sense. (A 20 year old change is no longer geopolitical "news" and we're not a history deck :).) For B. 20 years is a reasonable compromise between helping older people and not annoying younger ones for whom even 20 years is longer than their life. :)

As the cut-off between 2. and 4. (according to the above categorisation) 20 years is also IMO reasonable, since it's almost a generation, and within 20 years the old name should have either become history (i.e. 4) or become an entrenched alternative name (i.e. 3). (In other words, we don't care about Bishkek once being Frunze (until 1991), as it's an old, no longer relevant name (a "4"), but we do care about Myanmar once officially being Burma (until 1989), because "Burma" is still being actually used (a "3".)

[axelboc]

20 years does feel right and can easily be applied to categories 1 and 2/4. We'd be removing Zaire for DRC, which seems alright.

I do agree with your assessment that if Wikipedia says "formerly known as" or "historically", then it's definitely former names (2/4), but the case of alternative names (3) is a bit more challenging. Wikipedia uses "or", "also known as", "officially spelled" (e.g. for Naypidaw) and probably others... However, you're right, it does seems that "or" is used for alternative names that are both nearly as common. For such names, I'd argue that the 20-year cut off is not relevant and that we should mention the alternative name for as long as Wikipedia uses "or" in its introductory text. Names mentioned after "also known as" and other phrases may not be relevant-enough to be mentioned at all.

[emiham]

I think I was reasonably diligent taking notes during the process, but something might still have slipped through. These are the ones I found:

• Salisbury → Harare (1982)
• Reval → Tallinn (1918 apparently, mid-20th was a bad estimate)
• Kishinev → Chișinău - This one is interesting because I don't think there's a clear cutoff, but rather a general sentiment of moving to the country's preferred spelling (due to the older name stemming from the Russian name rather than Romanian), but it's also not really a name change as much as a spelling change, albeit one that is different enough that it might actually be helpful to include.
• Agana → Hagåtña (1998) - This one also blurs the line between spelling change and name change. It's also the closest one to the proposed cutoff.

I agree that 20 years seems like a sensible cutoff ­— it makes sense in the current year, and it doesn't create a lot of work for the foreseeable future.

Regarding alternative spellings, I also agree that they should be excluded unless "significantly different", which is of course a difficult line to draw. I think this gets easier the more languages the deck is already translated to, since alternative names that are not just slightly different spellings are often consistent between at least some languages (from a very Eurocentric perspective). Just to be safe I think it's probably good to opt for too many rather than too few at first, and then get a second opinion once it's time to merge.