[Localization] AbiWord Pootle status update

[Chris Leonard]

Dear AbiWord devs and localizers,

I have completed my QA review of the AbiWord PO files uploaded to our
Pootle instance:

http://translate.sugarlabs.org/projects/upstream_POT/

Many of the Western European languages (perhaps not surprisingly)
needed to be re-encoded into UTF.  I would remind localizers that
UTF-8 is the preferred encoding for i18n/l10n and ask them to consider
doing their off-line work in UTF-8.

The only languages remaining to be resolved are the Aymara and Quechua
languages.  I'm not neglecting them, I just want to work with Amos
Batto carefully to make sure we get the long-term set up correct.

I've made a few observations from running through all of the PO files.
 In many cases the automated version-to-version migration has resulted
in the population of certain strings with very poor translations due
to the fuzzy-matching algorithm involved.  I would strongly encourage
all localizers to review their PO files on Poolte using the Review
tab, with particular attention paid to those entries flagged "fuzzy".

Future steps of quality review may include doing a poconflicts
analysis between the two versions.  Another potentially useful step
would include doing a reverse poconflicts analysis which would
identify cases where the same string is used to translate two
different messages, this would be particularly helpful in identifying
artefacts introduced by the version migration process.

Another observation is that there are some opportunities to improve
the internationalization (i18n) of AbiWord to make it easier for
localizers.  I will be summarizing these ideas in an actionable form
and I will file it in the bug tracker for review / action by a
developer, I may share it with the list as well; when I have the
opportunity to do that work.

cjl