Comment 11 for bug 47280

Currently there are exactly two codes for which we offer multiple choices:
grep '^[^#]' data/languagelist | cut -d\; -f 3 | sed 's/_.*//' | sort | uniq -d
pt
zh

for:
Portuguese (Brazil)
Portuguese
Chinese (China)
Chinese (Hong Kong)
Chinese (Singapore)
Chinese (Taiwan)

These are the locales currently in the -en langpack:
dpkg -L language-pack-en-base | grep /usr/share/locale-langpack | sed -rn 's#.*/([^/]*)/LC_MESSAGES$#\1#p'
en_GB en_AU en_CA en en_NZ en_US en@boldquot en@quot

I think a wider discussion on this topic on ubuntu-devel@ would be good to clarify whether we want to go from one English in the list to 4 or so; I would guess we want to do this for all languages.

My personal opinion is that while it is more complex, we already expose this complexity in ubiquity during installation (albeit assisted).