[Localization] Polish Etoys translation mistake

[Chris Leonard]

On Tue, Mar 20, 2012 at 3:48 AM, Bert Freudenberg <bert at freudenbergs.de> wrote:
> I don't think so. You can mix RTL and LTR text, and Latin characters would always appear left-to-right.
>
> Btw, for now, Etoys doesn't even use that field, it's for when we get around to actually  change the UI layout based on the setting. Maybe we should just drop it. The YAGNI rule would suggest so ...


Bert,

It is actually not too difficult to find reliable sources of language
names in their own language and script.

My best trick for doing this is to look at the ISO-639 PO files (for
that language code, if available) and then look up the language name
msgid and copy the corresponding msgstr.  One can generally expect
that the language name from the corresponding language code file will
be translated correctly.

ISO-639 code PO files
http://anonscm.debian.org/gitweb/?p=iso-codes/iso-codes.git;a=tree;f=iso_639;hb=HEAD

In some cases, you might need the ISO-639-3
http://anonscm.debian.org/gitweb/?p=iso-codes/iso-codes.git;a=tree;f=iso_639_3;h=3ac1e50ca8af1af1a6c1c579f29cc8388c76c34a;hb=HEAD

My second best trick is to look up the language in English wikipedia
(where it will typically appear in the original script) or to look
down the left hand bar for the link to the language in the
corresponding language wikipedia.  It is not unexpected that every
language wikipedia has an article on it's own language and these tend
to be linked by the inter-wikipedia language links and easy to find,
even if you may need to use the language code to recognize the
language name links as you mouse over them.

As for LTR or RTL, this is pretty easily determined from the set up in
Pootle by whether the text is left-justified or right-justified.

I've gone through all of the eToys project languages and made
corrections or additions where needed, although in some cases Ieft
them flagged as "fuzzy",

cjl