English keyboard on a Google Pixel phone, I press and hold ‘a’ to get æ, ã, å, ā, à, á, â or ä. Press and hold ‘s’ and all I get is ß.
I am amused at you accidentally picking one of the Virgina towns most likely to gain names in other languages, as Ashburn is the nominal (named) location of AWS’s us-east1.
German layout is fine for accents, although they are not used in German, you type them by combining the accent with another letter, worked like this also in times of mechanical typewriters.
Don’t know about the ç though ![]()
Microsoft distributes the Canadian Multlilingual Layout with Windows. Direct keys for French and support for a lot more, including the Romanian ă (but not the ș). There’s also the US International layout, which is less disruptive to someone used to normal US keyboard layout (=less keys have their main function changed).
Meanwhile we haven’t gotten Microsoft to support the nationally standardized Finnish-Swedish multilingual variant (and physical products don’t exist either, bit of a chicken and egg problem there). The Windows default Finnish keyboard layout doesn’t even support all Finnish characters.
Sorry for the off-topic keyboard talk, I’ll stop here.
Windows supports the keyboard layout, but hardware manufacturers don’t widely distribute the actual, physical keyboards (save perhaps in Quebec (Québec
) and parts of New Brunswick…?). You can’t effectively use the “Canadian Multilingual Standard” without a compatible physical keyboard, the “multilingual” keyboard having several additional keys compared to unilingual English keyboards. Despite having done all my primary and most of my secondary education in French, having lived my entire life in Western Canada: I’ve never seen a multilingual keyboard in person.
And then “Kraków” is “Kraków” not “Krakow”
But if you think about it, in Poland we have a special word for München (Monachium) but we don’t have it for Düsseldorf.
def get_display_name(tags):
if "name:en" in tags:
return tags["name:en"]
for key, value in tags.items():
if key.startswith("name:") and key.endswith("-Latn"):
if tags.get(key.removesuffix("-Latn")) == tags["name"]:
return value
return tags["name"]
I’m sure those OpenStreetMap-based apps with thousands of lines of code could deal with this level of complexity.
You missed the point. It’s not the code that would find it complex, it’s the complexity of tagging that creates uncertainty about what tags to include in the code. That’s a different problem set.
I’ve finally identified one of things that was bugging me about this – I think one big problem here, @apm-wa, is that you are using the information the US Embassy has/is receiving, as authoritative for all English speakers.
This is solidly in conflict with the consensus we came at for the Gulf of Mexico and Lake Ontario – when the government picks a weird name, we shove it into official_name and keep name as the common name.
If there is no common name in English, then name:en should be omitted.
Don’t worry, they are still Gulf of Mexico and Lake Ontario to me (and I suspect to all but a vanishingly small number of OSMers). What you describe is what the name:en-US tag is used for.
To put it in perspective when I was a diplomat in Turkmenistan, Palestine opened an embassy there. The U.S. government does not recognize Palestine so I received instructions to address the Palestinian ambassador as “Mr. Ghanem”, not “Ambassador Ghanem”. When we met, I asked him to use my first name and we had a cordial if unofficial relationship. When I mapped the Palestinian Embassy in Ashgabat, I mapped it as an embassy since the Turkmen government recognized it as such. No problem.
From a UK perspective, that was bothering me also. Here there is certainly a tolerance for accents in names sourced from other languages. A UK-based web search for Dusseldorf (without the umlaut) gets a definite majority of UK-targeted sites with the umlaut rather than without it (and the occasional “ue”). “Skoda” (without the háček on the “S”) is similar, and even if you exclude the direct resellers following corporate branding, you still get plenty returning text with “Škoda” in it.
Somewhat related discussions have been had in Ireland and New Zealand about names of Irish and Māori origin respectively. I even seem to remember San Jose been mentioned previously too (currently the alt_name for that has the accent in it).
I get the feeling that the acceptance of non-English characters in English is less in the US, especially outside the large Spanish area of influence in the southwest?
I don’t think that it makes sense to make up and store in OSM a name:en for places for which there is no common name in English because the presence of that tag means more than “here is something to show to English-speakers”; it means “OSM thinks that there is a distinct English name for this place”.
I have yet to see a U.S. typewriter keyboard with umlauts or accent keys that don’t move the carriage, as exist in other parts of the world. My U.S. computer keyboard has no diacritical marks if I select the ENG-US keyboard (I also commonly use the RUS and TKM keyboards). Yes, loan words of foreign origin exist in English, but they are borrowings and though in common usage are not recognized as “English” words. C’est la vie dans le monde anglophone. ![]()
I just fundamentally disagree with this take. If no diacriticals can exist in English, what language is the name of the city La Cañada Flintridge, California in? Obviously part of the name came from Spanish originally, but it’s in common English usage by English speakers and has no other name (spelling or pronouncing it “La Canada” would be regarded as an error). This makes it English in any way that matters for OSM in my opinion.
I’m not sure what typewriter keyboards have to do with this, although maybe if I’d ever seen a typewriter that wasn’t in a movie or museum it’d mean more to me
. Using my Mac’s US keyboard, I wrote “ñ” using option+n, it wasn’t so hard.
Yeah, well, we’ll have to agree to disagree. English is English and Spanish is Spanish, and loan words in English from Spanish are still Spanish. Most Americans if asked which states of the union have Spanish-origin names cannot name them (no, New Mexico isn’t one of them, and neither is Texas, both of which derive from indigenous languages, not Spanish). That is why Montana is spelled in English Montana and not Montaña.
Not to start a linguistic prescriptivist vs descriptivist holy war, but as a native English speaker I make a point of trying to respect diacritics whenever they’re present, at least in Latin alphabets
Leaving them off feels lazy and disrespectful of the language I am robbing borrowing from
E.g., if Türkiye believes Türkiye is preferable to Turkey, then I should make a point of including the diacritic, even if that means I have to hunt down the code point somewhere when typing (via an alt code, long press, search engine, or whatever other resource/mechanism). For the country, I’d expect something like this for tagging: name=Türkiye, name:en=Türkiye (or, if we must, simply leave this particular tag out), and alt_name:en=Turkey for the non-preferred name. Alas, that’s not how it is, but hey.
because the Turks know better how to call their country in English than the native English speakers, or why?
Why do you limit it to latin alphabets? Couldn’t you hunt down the correct code points also for other scripts? What’s wrong with name:en=北京 and alt_name:en=Beijing?
Do you also study phonetics of these languages, to pronounce Türkiye like a Turk, Pyšely like a Czech, and 北京 like a Chinese person?
I think that that happens organically, as more English speakers hear names spoken by native speakers, It also varies with context (e.g. a conversation with people who only speak English, or native speakers).
Of your examples, I’ve never heard of Pyšely before, but according to OSM and Wikipedia there isn’t an English name for it. I’d use the Czech spelling and try and use an approximation of the Czech pronunciation (English “sh” sound rather than “s” or “ss”). It might not be a very good approximation, but I wouldn’t feel the need to make up a new English name and use that.
“Türkiye” can be, and is, approximated by the average English speaker in speech, as after removing the diacritics, it’s in Latin script which they are accustomed to. The average English speaker is simply unable to pronounce 北京, because they have no phonological approximation for it. They cannot associate the written form with the spoken form without explicit instruction.
@SomeoneElse, Thank you for this answer; it’s an interesting perspective – and one that’s completely unexpected for me. It seems more like the development of a new international language, in which local names would be both written and pronounced in a way that stays close to the original.
