en-US language presumption (was...)

[Jan Claeys]

Op vrijdag 28-08-2009 om 23:26 uur [tijdzone +0100], schreef Paul
Sladen:
> If this presumption is being made, then it is a bug.  People should be
> free to code their human-readable appliation output in whatever
> (human) language that group of developers feel most collectively
> comfortable with.
> 
> I personally write my error messages in UTF-8 formatted en-GB---it
> would be an unfair bunden to expect me to be familiar with en-US (or
> en-CA, en-AU, en-*)---and I would be more likely to make mistakes in
> attempting to do so. 

I agree that that should be the way it works, but unfortunately as far
as I know the GNU gettext string extraction utilities assume that the
source language is always US English.

(IIRC the gettext specifications don't mention this, so maybe we should
consider it as a bug in the GNU gettext utilities implementation?)


-- 
Jan Claeys