Weird little quirk I've recently encountered, somewhat related to OSM: the `mkgmap` tool used to convert OSM data to Garmin map files display Chinese pronunciation as English names next to POI names in Japanese, due to 1) Japanese OSM maps have a lot of entries written in Japanese without indication of language, 2) no one is doing anything about it, due to Wikipedia style community nuances or something, and 3) the tool have no "--assume-language=$LANG" style of options(plus Japanese script->pronunciation mapping isn't even remotely bijective).
e.g. There's a bicycle parking next to the parkside exit at Ueno station, "上野駅公園口自転車駐車場". This is:
meaning: Ueno station park entrance bicycle parking ground in English,
read IRL: ue no eki kouen guchi jitensha chusha jou in Japanese, but written
on the map: shang ye yi gong yuan kou zi zhuan che zhu che chang as Chinese(in Pinyin but I don't know how to type it).
...I would imagine that utility of that is limited for non-Japanese users. Makes me wonder what are the reasons why master OSM data hadn't had missing pronunciations autogenerated or mkgmap tool given an auto selected language flag, but I'm not intending to be OSMian nor am I wanting to be the one to define national borders for bunch of nation states, so there's that...
I have a hard time making sense of your post. What do you mean with 'Japanese OSM maps'? There is just one OpenStreetMap, which includes Japan.
On OpenStreetMap all entities which are named have at the very least the tag `name`. There is no need to specify which language that name is in, because this is geographically determined. For the whole of Japan, `name`-tags are considered to be in Japanese.
This specific bicycle parking has no other names but that one `name` tag, which is the Japanese 上野駅公園口自転車駐車場:
This is totally normal. Most named entities only have a local name. A random bicycle parking is not likely to have a name in any other language but the local one. For things which actually have an exonym (like countries and larger cities), you will find additional namespaced language tags like `name:ja` and `name:en`, but a bicycle parking in Japan usually doesn't have an English name (although you can translate the Japanese name).
As for why you are seeing Chinese, that has nothing to do with OpenStreetMap. Some other tool in your chain is doing that (mkgmap?).
Yes, apparently it's a known unresolved issue with mkgmap tool. Looks like there is actually the option but not autodetected/not applicable if the entire planet is converted at once?
The nonsensical reading I posted in the parent comment was what I actually saw in a free map app that uses OSM data. That conversion tool is apparently used for few apps, and it apparently generate English pronunciations whenever "en" resources are missing and bake it into offline map tile files, but uses some Chinese transcription engine for all Kanji/Hanzi names.
---
So, I assume this can be solved two ways: someone can define polygons for national borders, at least for all countries that could experience this problem, such as Japan, China, Korea, Vietnam, etc., and use appropriate transcription engines to generate English readings, either at the OSM side or at the conversion tool side. But that's going to define borders for nations, and therefore I assume that there will be some challenges along that borders since everyone have some territorial disputes.
I don't think this has nothing to do with OpenStreetMap; OSM could have tentative autogenerated name or something for location names that could use English transliteration, as one way this problem can be solved, aside whether it's OSM community responsibility to do so; it's extremely high quality community maintained free/libre map, which is already tremendous. Or it could be done at export tool side, following perhaps more natural division of responsibilities. But regardless, as a non-critical longstanding issue, and a matter of fact, there are actually such exported map files out there with Chinese pronunciations baked in.
Comments
Weird little quirk I've recently encountered, somewhat related to OSM: the `mkgmap` tool used to convert OSM data to Garmin map files display Chinese pronunciation as English names next to POI names in Japanese, due to 1) Japanese OSM maps have a lot of entries written in Japanese without indication of language, 2) no one is doing anything about it, due to Wikipedia style community nuances or something, and 3) the tool have no "--assume-language=$LANG" style of options(plus Japanese script->pronunciation mapping isn't even remotely bijective).
e.g. There's a bicycle parking next to the parkside exit at Ueno station, "上野駅公園口自転車駐車場". This is:
...I would imagine that utility of that is limited for non-Japanese users. Makes me wonder what are the reasons why master OSM data hadn't had missing pronunciations autogenerated or mkgmap tool given an auto selected language flag, but I'm not intending to be OSMian nor am I wanting to be the one to define national borders for bunch of nation states, so there's that...I have a hard time making sense of your post. What do you mean with 'Japanese OSM maps'? There is just one OpenStreetMap, which includes Japan.
On OpenStreetMap all entities which are named have at the very least the tag `name`. There is no need to specify which language that name is in, because this is geographically determined. For the whole of Japan, `name`-tags are considered to be in Japanese.
This specific bicycle parking has no other names but that one `name` tag, which is the Japanese 上野駅公園口自転車駐車場:
https://www.openstreetmap.org/node/3395310504
This is totally normal. Most named entities only have a local name. A random bicycle parking is not likely to have a name in any other language but the local one. For things which actually have an exonym (like countries and larger cities), you will find additional namespaced language tags like `name:ja` and `name:en`, but a bicycle parking in Japan usually doesn't have an English name (although you can translate the Japanese name).
As for why you are seeing Chinese, that has nothing to do with OpenStreetMap. Some other tool in your chain is doing that (mkgmap?).
Yes, apparently it's a known unresolved issue with mkgmap tool. Looks like there is actually the option but not autodetected/not applicable if the entire planet is converted at once?
The nonsensical reading I posted in the parent comment was what I actually saw in a free map app that uses OSM data. That conversion tool is apparently used for few apps, and it apparently generate English pronunciations whenever "en" resources are missing and bake it into offline map tile files, but uses some Chinese transcription engine for all Kanji/Hanzi names.
---
So, I assume this can be solved two ways: someone can define polygons for national borders, at least for all countries that could experience this problem, such as Japan, China, Korea, Vietnam, etc., and use appropriate transcription engines to generate English readings, either at the OSM side or at the conversion tool side. But that's going to define borders for nations, and therefore I assume that there will be some challenges along that borders since everyone have some territorial disputes.
I don't think this has nothing to do with OpenStreetMap; OSM could have tentative autogenerated name or something for location names that could use English transliteration, as one way this problem can be solved, aside whether it's OSM community responsibility to do so; it's extremely high quality community maintained free/libre map, which is already tremendous. Or it could be done at export tool side, following perhaps more natural division of responsibilities. But regardless, as a non-critical longstanding issue, and a matter of fact, there are actually such exported map files out there with Chinese pronunciations baked in.