[Localization] Picking up additional strings

[Chris Leonard]

Yama,

I have some Aymara string suggestions for the Write Activity for you to
consider for inclusion!  How does the gringo manage this you may ask?  Well,
I have been looking into the overlap between strings in Write and Abiword
and about half of the 42 UI strings in Write are present exactly in AbiWord,
which has a partially overlapping set of language projects in their Pootle
system, including lang-ayc and lang-aym strings, where OLPC has lang-ay.

On a system with the GNU translation toolkit and gettext installed, run
these commands.

mkdir test

#get the Abiword lang-aym and lang-ayc PO files
wget http://www.abisource.com/dev/strings/stable/aym-BO.po -O aym-BO-abi.po
wget http://www.abisource.com/dev/strings/stable/ayc-BO.po -O ayc-BO-abi.po

#get the Write lang-ay PO file (from fructose82)
wget --no-check-certificate
https://dev.laptop.org/translate/ay/fructose82/write.po -O ay-write.po

#merge the identical strings from Abiword PO into Write PO
msgmerge -N test/ayc-BO-abi.po test/ay-write.po -o test/temp-ayc-write.po
msgmerge -N test/aym-BO-abi.po test/ay-write.po -o test/temp-aym-write.po

#clean-up some extra strings that get carried over
msgattrib --no-obsolete test/temp-ayc-write.po -o test/new-ayc-write.po
msgattrib --no-obsolete test/temp-aym-write.po -o test/new-aym-write.po

You now have two PO files new-ayc-write.po and new-aym-write.po, there are
some subtle differences that an Aymarista will have to resolve.

Of course, simple string matching isn't sufficient in all cases, context is
important, but in the AbiWord > Write instance, the context should be highly
preserved.

The same thing can be done for Quechua (although there is again lang-quh and
lang-qul to resolve).

Similar string pickups from Abiword are possible for Amharic, Czech, Danish,
Korean, and they are simpler in that there are not multiple variants
present.

Picking up about half of the Write strings may be a small thing, but it
points out the possibility of more effectively leveraging
upstream/downstream community interactions, which is what open source is all
about (at it's best).

While OLPC Write strings are a small subset of AbiWord strings, it would be
possible for AbiWord to benefit from picking up a few strings from langs
where OLPC coverage exists (in the absence of AbiWord coverage), such as
Bengali, Hindi, Marathi, Sinhala, Telegu, Mongolian, Kreyol, Khmer,
Macedonian, Kinyarwanda, so while the benefits mostly flow from upstream to
downstream, it is not entirely one-way.
I am somewhat curious about how extensively such a strategy might be
leveraged for other activities that are essentially cut down versions of
bigger projects (evince>Read, xulrunner>Browse, etc.) as these are the
instances where a semi-mechanical approach like this can somewhat rely on
conserved context.


cjl
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.laptop.org/pipermail/localization/attachments/20081106/c7958b35/attachment-0001.htm 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: new-ayc-write.po
Type: application/octet-stream
Size: 3549 bytes
Desc: not available
Url : http://lists.laptop.org/pipermail/localization/attachments/20081106/c7958b35/attachment-0002.obj 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: new-aym-write.po
Type: application/octet-stream
Size: 3547 bytes
Desc: not available
Url : http://lists.laptop.org/pipermail/localization/attachments/20081106/c7958b35/attachment-0003.obj