GNOME Bugzilla – Bug 149982
gweather locations need non-ASCII in their names
Last modified: 2005-01-09 08:03:52 UTC
It seems all the English names in the gweather po-locations are ASCII-only. This isn't really helpful, since the names of many locations are really only known by their proper names, not their ASCII-mangled counterparts. ASCII-only names don't help translators, since the translator of language A rarely would realize the proper name for obscure airport Z in country B after the name has been mangled though an ASCII-only filter. As an example, the Swedish rather small city "Trollhättan" should most likely be spelled the same way in German. It's however unlikely that a German translator realizes this though if the name is mangled to "Trollhattan". I would argue that ASCII-only names here don't even help English-speaking people, since if they are looking for a small airport, they'll be a lot more successful using it's proper name (if they don't have access to the code). I don't know if something like this would be possible: <location code="ESIB" name="Satenas"> <!-- ESIB Satenas --> <_name>Såtenäs</_name> </location> This would still allow the entry to be accessed both by code and by (ASCII) name, while still give translators a big hint to be able to translate it correctly. Also, "en_US" or "C" users will benefit from having the names properly spelled.
You're so much right! These corrections aren't feasible for 2.8, but 2.10 should really be patched to have sane location descriptors.
Bug 151675 contains a patch with the UTF-8'ified names of the Swedish locations. Perhaps others could contribute similar patches.
In my opinion, Locations.xml.in should contain names the names in en_US, with comments giving the localized name. The translators will then be able to use the name they use in their locale. That said, I want to move the locations database into it's own file anyway, so that all programs will be able use the same location database. Perhaps we can move it to freedesktop.org, so that we can share it with KDE.
Yes, the names should be the en_US ones. After all, we do require all messages to be written using en_US English. As an example, perhaps "Göteborg (Säve)" in my patch should really say "Gothenburg (Säve)", since Gothenburg is the name of the city in English. I mainly used "Göteborg (Säve)" in my patch since currently it says "Goteborg (Save)" which is a city and an airport that doesn't exist. It's either "Göteborg" or "Gothenburg", never "Goteborg". Similarily, the name of the airport is "Säve" and nothing else. If anyone hasn't figured it out already, what I mean is that using the en_US name doesn't mean that the name should always be in ASCII. Some locations have established English names, like "Göteborg"->"Gothenburg" and "München"->"Munich". Then there's no question that the English name should be used, and the original name be a comment only. Many other locations do not have established English names, like "Düsseldorf", "Borlänge", "Växjö", "Umeå" and so on, and then the original name should be used, and no ASCII-munging take place. Of course one might want to limit these names to latin script, so that only latinized versions of locations that are otherwise natively spelled in other scripts like cyrillic, CJB etc. are used. It would be good though if native names in native script still could be added as comments though.
Just found this bug. I have a patch for German cities in Bug 155202 (-> add dependency)
Can't this be closed now that we've had locations love action going for a while?
Yes, I think so.