Languages aren't quite in alphabetical order

[Matthew Paul Thomas]

Binary package hint: gfxboot-theme-ubuntu

Ubuntu Hardy alpha 3, live CD

1. At the initial menu, choose "Language".
2. Observe the order of the options presented.

The languages are almost, but not quite, in alphabetical order. For example, "Español" comes after "Esperanto", "Suomi" is between "Euskaraz" and "Français", and ""Latviski" comes after "Lietuviškai".

Perhaps this is caused by alphabetical sorting of some underlying language code, though even that shouldn't cause "Norsk bokmål" and "Norsk Nynorsk" to be separated by two unrelated languages. Regardless, it looks like a mistake.

I suggest that languages presented in Latin characters be sorted by those Latin names, and languages presented in non-Latin characters be sorted by their English names. So for example, "Português" should be listed before "Português do Brasil", and the two Chinese options should be listed between "Čeština" and "Dansk".

[Colin Watson (cjwatson)]

The languages are sorted by their ISO-639 language code (i.e. the first bit of $LANG). nb < ne < nl < nn (Norsk bokmål, Nepali, Nederlands, Norsk nynorsk).

I acknowledge that it is not perfect, but I'm not convinced that your proposal is fully-specified, and it's exactly this woolliness that has held me back from doing it that way in the past. See bug 39299. The short answer is that there is no sort order that makes sense to speakers of all languages, even if we as English speakers have an intuition that says that it makes sense to sort similar-looking Latin letters together. The current behaviour is a compromise.

[Colin Watson (cjwatson)]

You can experiment with different national language sort orders by running the 'sort' program with LC_COLLATE set to various locale names in the environment.

[Colin Watson (cjwatson)]

This is also bug 40144.