As mentioned in a prior note, here's a current comparison of the translation
status for all locales that have a corresponding ispell dictionary.
<pre>
Latin-1 KDE GNOME Moz Abi Star MSFT
------- --- ----- --- --- ---- ----
en-US English - US 1 1 1 1 1 1
en-GB English - British 1 2 9 1 1
fr-FR French 1 1 2 2 1 1
de-DE German 1 1 1 1 1 1
es-ES Spanish 1 1 2 1 1 1
du-NL Dutch 1 2 9 2 1 2
it-IT Italian 2 2 1 1 1 2
pt-BR Portuguese - Brazil 1 1 2 2 2
pt-PT Portuguese 1 3 2 1 1 2
sv-SE Swedish 1 1 2 1 1 2
da-DK Danish 1 1 2 1 1 3
fi-FI Finnish 2 1 9 1 3
nb-NO Norwegian - Bokmål 1 1 9 1 3
nn-NO Norwegian - Nynorsk 1 2 1 1 3
ca-ES Catalan 2 2 2 1 4
af Afrikaans 4 9
de-CH German - Swiss
</pre>
Not surprisingly, we have excellent coverage for these Latin-1 languages --
as good as any comparable Open Source project, and a bit better than the
proprietary alternatives.
Note, however, that our support of the corresponding ispell dictionary files
isn't so hot:
http://www.abisource.com/bugzilla/show_bug.cgi?id=824
http://www.abisource.com/bugzilla/show_bug.cgi?id=902
In short, the problem seems to be that we only recognize one of the various
hashfile encodings out there, and balk at any of the others (the "magic
number" error).
<pre>
Latin-2 KDE GNOME Moz Abi Star MSFT
------- --- ----- --- --- ---- ----
pl-PL Polish 2 2 2 1 1 2
cs-CZ Czech 1 3 3 1 2
ro-RO Romanian 1 1 9 4
sl-SI Slovenian 1 2 9 4
Cyrillic
--------
ru-RU Russian 1 1 1 1 1 2
uk-UA Ukrainian 2 1 9 1
be-BY Belarusian
</pre>
We're starting to see more translations for Central European and Cyrillic
locales. AFAIK, these have only been tested on Linux, and I have no idea
whether the corresponding ispell dictionaries Just Work. Unless everyone is
using Unicode fonts for these locales, I suspect that there are lingering
charset issues to be dealt with. (Just a hunch.)
<pre>
Mediterranean KDE GNOME Moz Abi Star MSFT
------------- --- ----- --- --- ---- ----
el-GR Greek 2 1 1 3
BiDi
----
he-IL Hebrew 2 9 1 3
Other
-----
eo Esperanto 1 2
</pre>
These last three locales are all still TODO items for AbiWord. Presumably
once we've addressed charset issues satisfactorily for Latin-2 and Cyrillic,
then el-GR and eo shouldn't be much additional work.
Finally, I have a strong impression that the he-IL locale will take care of
itself as soon as Tomas' BiDi work gets more prominently featured. ;-)
Paul
