Re: Tools to help with mnemonic l10n?

[Chris Leonard <cjlhomeaddress gmail com>]

On Wed, Jan 11, 2012 at 5:00 AM, Byrial Jensen <byrial vip cybercity dk> wrote:
> Hi,
>
> When I translate button labels and menu texts etc., it may be difficult to
> select good (that is unique) key mnemonics (the letters with a prepended
> underscore) as I often do not know which options will be grouped together in
> the same menu or window.
>
> I wonder if there is any tools to help with that.
>
> When the UI is made with glade, maybe the program which extracts the
> translatable strings from the glade XML file could analyze the file enough
> to insert comments into the pot file about which strings with mnemonics are
> in the same group. This could help the translator, and it would  be possible
> for other tools such as for instance pofilter to check for possible mnemonic
> conflicts in the translated po files. Do anything like this exist, or has
> anyone plans to do like this?
>
> Best begards
> - Byrial


I don't know of any such tool, but I like your idea.  I particulalry
like the notion of a pofilter check, but on what basis will the
algorithm know that the entries are in a common pulldown?  One
intermediate step would be improved translator comments in the PO
file.  This is still imperfect as depending on the sorting of the PO
file as presented (some people go for alphabetical instead of
appearance in code) , even the best manual annotation is not going to
group the menu items as needed.

Having reviewed many PO files with accelerators in many languages, I
can tell you that localizer understanding of and implementation of
accelerators is very uneven.  Romance languages are generallty better
(probably due to common root word origin) whereas languages in
non-latin scripts tend to either skip the accelerator or simply leave
it in place at the beginning.

Don't forget the use of ampersand & as an accelerator in addition to
underscore in various settings.

http://translate.sourceforge.net/wiki/guide/translation/accelerators

cjl