The format of VOMP localisation files is described below.

Each file may contain one or more languages, and a language may be contained
in more than one file.

A valid file must start with a line of the following form for each language
in the file:

vomp-l10n:[<wsp>]<lang-code><wsp><lang-name>

<wsp> represents a whitespace string (spaces and/or tabs).
<lang-code> can be any string of visible ASCII characters. It is intended to
be a recognised language code or locale code.
<lang-name> can be any UTF-8 string.

Example:

vomp-l10n: cy Cymraeg
vomp-l10n:ga-IE	Gaeilge
vomp-l10n:	gd-GB Gàidhlig

The first line not of this form begins the main body of the translation file.

A line of the following form introduces a key to be translated:
x:[<wsp>]["]<text-key>["][<wsp>]

<text-key> represents an ASCII string.
Quotation marks (ASCII 34) are optional and would normally be used to include
whitespace at the beginning or end of the string.

A line of the following form provides a translation for the previous key.
If there is no previous key, the translation is ignored.

<lang-code>:[<wsp>]["]<translation-text>["][<wsp>]

<lang-code> must match one of the codes introduced by the vomp-l10n:
section; otherwise the line is ignored.
<translation-text> can be any UTF-8 string.

Any other lines (blank or otherwise) are ignored. Nonblank lines should be
treated as comments and preserved by tools written to manipulate these files.

Example:

x:	One
cy:	Un
ga:	Aon
gd:	Aon

* This is a comment. Note the trailing whitespace below.
x:"two "
gd:		"dà "
ga:		"dó "
This is a comment, perhaps giving a good reason why the translation below
has no trailing space. The quotation marks have no effect here.
cy:	"dau"

(Thanks to http://home.unilang.org :-)