[OOBE] wizard prioritises US English over UK English

Sorry, while I understand what you're saying, this is per spec:

  https://wiki.ubuntu.com/LanguageAndText#Sorting_languages

You'll need to have the design changed if you want the implementation to do something different.

From the spec...

> The variation that has the unprefixed language code (or an equivalent) should be sorted first. For example, en_US = en = “English
> (US)” (not “U. S. English”), nl = “Nederlands (Nederland)”. (This avoids hiding the most common variation amongst many others.)
So where exactly is 'en' uniquely mapped to en_US?

I don't have a problem with the default being US English, but it seems rather less than intuitive if the list is not sorted alphabetically.

According to the spec, the claimed reason for sorting Interface Languages by name rather than language code is that nobody knows the language code. This is reasonable. However, using that logic, the list of territories for a language should also be sorted alphabetically "since nobody knows the unprefixed language codes".

A nice touch might also be to pre-populate the default language using the GPS location of the device mapped to that countries main language.

Attached is a snippet of the language list (the same list is shown in the wizard).

As can be seen, "English (United States)" comes first, but "English (United Kingdom)" is seven entries below "English (India)" (and thus not displayed as off-screen).

("Opinion" is a closed bug status, so reassigning an "Opinion" bug report has no consequence. That people persistently don't realize this is one of the reasons that "Opinion" should no longer exist, bug 772954.)

It would be inconsistent for the first-run setup to list languages in a different order from System Settings. So this is a general Language & Text bug, not a bug with the setup in particular.

I'm not happy that the language list falsely claims to offer 17 different varieties of English -- for example, English (Zimbabwe) has a grand total of zero strings translated in Launchpad, while English (Zambia) isn't translatable in Launchpad at all. <https://translations.launchpad.net/ubuntu/vivid/+lang/en_ZW> <https://translations.launchpad.net/ubuntu/vivid/+lang/en_ZM> Each spurious entry makes the list as a whole harder to scan (bug 1350275).

However, regardless of which language variations are useful, I don't think it's "an unreasonable bias" to list the one with the un-suffixed language code first. Even the original example of English demonstrates this: US English is used as a first language by more people than all other varieties of English combined. <https://en.wikipedia.org/wiki/List_of_countries_by_English-speaking_population>

English is a muddled example, since it's the default, so most people won't be actively selecting it in the first place. But the same is true of French: there are more speakers of Français (France) than Français (Canada) and all the other variations put together. Similarly with Català (Espanya), Deutsch (Deutschland), and Nederlands (Nederland).
<https://en.wikipedia.org/wiki/Geographical_distribution_of_French_speakers>
<https://en.wikipedia.org/wiki/Catalan_language#Number_of_speakers>
<https://en.wikipedia.org/wiki/German_language#Geographic_distribution>
<https://en.wikipedia.org/wiki/Dutch_language#Geographic_distribution>
In all these cases, a strict alphabetical ordering would obscure the overwhelmingly most common variant below much-less-used variants.

I'm aware of only two languages where this heuristic breaks down: Spanish: Español (México) has more first-language speakers than Español (España), and Português (Brasil) has more first-language speakers than Português (Portugal). But alphabetical ordering wouldn't improve the situation for Español anyway. If there are several other languages where it would, that would be reason to reconsider this ordering.