Zanata assumes that the source language to be translated is US English.
However this is not necessary true.
For most work, when translating TO english, this is not a issue.
However there are a few places that this makes difference (other than the GUI).
For example, the glossary.
The glossary is a great place to add possible translated options for an ambiguous word. However, its functionality is diminuted when the target language is en_us.
I can see two milestones to solve this issue.
The first is to select a system wide source language.
This solves problems for local installations.
However it is not useful for shared installations like translate.zanata.org.
The second milestone is to allow a per project, per version, per document source language.
This would allow for the greatest granulaty and many db structure updates.
Just for reference, I am using my installation to translate books and texts from pt_BR to en_US and others,