|
I am confused about what tag should I use for transcription of Greek street names. In name= I have the street name in Greek but I this makes it useless for a foreigner. So I want to add multilingual tag, but which one is correct? Thank you |
|
See https://wiki.openstreetmap.org/wiki/Multilingual_names#Greece (→ If this is a transcription which can be done automatically by computers: I would say, no additional tag at all. See https://help.openstreetmap.org/questions/7492/transliterating-arabic-names . The transcription can be done by the map maker's software: One example is the "Mapnik" style by the German OSM section: compare (only a Greece name tag for the street in the middle … but still a transcription is shown; the tertiary in the upper part has a |
|
For Serbian (name:sr) a tag for latinised names exists: name:sr-Latn. I would have assumed that the appendix "-Latn" is a common way to add latinised names of languages with not-latin names to OSM but I see no documentation so far. Though I'd encourage you to use name:el-Latn - it can't be wrong. :) 2
-Latin, _Latin or similar would make sense, -Latn would seem to be a typo that has propagated a bit (Ltn would make more sense if you actually cosidered 1-2 characters worth saving).
(12 Jan '15, 11:43)
SimonPoole ♦
1
-Latn sounds funny but there is some standard which mandates this: "The locale tags have been chosen according to BCP 47." It's also used for China and Japan. https://wiki.openstreetmap.org/wiki/Multilingual_names#Serbia https://wiki.openstreetmap.org/wiki/Names says simple transliteration should be avoided.
(11 Feb '15, 18:59)
Jojo4u
|
|
"For Serbian (name:sr) a tag for latinised names exists: name:sr-Latn. I would have assumed that the appendix "-Latn" is a common way to add latinised names of languages with not-latin names to OSM but I see no documentation so far. Though I'd encourage you to use name:el-Latn - it can't be wrong. :)" That is a horrible tagging scheme, as it makes the transliteration tag's key language depended, meaning there could potentially be hundreds of such "language dependent" tags, each with their own key. This makes the use of such keys and transliterations pretty much impossible in any renderer, as it would require dealing with, and tracking hundreds of keys. I would really strongly urge to stick to int_name or name:en as established keys for transliteration to English / latin alphabet. |
Thank you for replies! According to this link https://wiki.openstreetmap.org/wiki/Multilingual_names#Greece I have to use int_name. I checked my program (OsmAnd), it makes the transcription of not multilanguage tagged street names but wrongly: Instead of Vyronos it the street is Buronos. I suppose it is a programs fault.
@hariskar: Our wiki does not really set hard rules, but suggestions and documentation of usual practise or what most mappers agree to (it is a bit hard to define what our wiki really is). So you do not "have" to.
Yes, if there is no transcription in our data, then it is OsmAnd's "fault" (or the one of a used library). Maybe several transcription schemes exist (e.g. one usually used in English and one usually used in German – this is the case for Russian)? You could report to / discuss this with the programmers (e.g. in the Issue Tracker).