#10797 NORM Not Tri: Integrate Language packs into olpc-os-builder
Zarro Boogs per Child
bugtracker at laptop.org
Thu Mar 31 14:46:01 EDT 2011
#10797: Integrate Language packs into olpc-os-builder
--------------------------+-------------------------------------------------
Reporter: reuben | Owner: martin.langhoff
Type: defect | Status: new
Priority: normal | Milestone: Not Triaged
Component: not assigned | Version: not specified
Keywords: | Next_action: never set
Verified: 0 | Deployment_affected:
Blockedby: | Blocking:
--------------------------+-------------------------------------------------
From an old thread with Sayamindu:
Hi,
On Wed, Mar 24, 2010 at 7:35 PM, Reuben K. Caron <reuben at laptop.org>
wrote:
On Mar 23, 2010, at 4:48 AM, Sayamindu Dasgupta wrote:
Hi,
Of recent, there has been OLPC sponsored translation efforts for
languages like Brazilian Portuguese, Hindi and Haitian Kreyole (all of
these are for potential deployments).
I'd also like to add Arabic.
While we can easily integrate
the work done by these translation teams in the core sugar packages of
our builds (we maintain these packages) - there are no straightforward
ways to do this for activities. Some activities see regular releases,
but there are some (like Paint) without any proper maintainer at all.
I was wondering if it makes sense to make the build system pull in the
translation of the activities directly from the translation server and
replace the older ones with it.
I was surprised when I found out we already did not do this.
Does it sound like a sane idea ? It's
a bit ugly in the sense that we would be blindly overwriting older
translations, but I think it is better than shipping translations that
were last updated in 2008, especially when OLPC is investing in making
the translations up to date.
Can you elaborate on what peer review or approval is being done on
translations to ensure the accuracy of translations? This may decrease the
fear of "blindly overwriting" translations.
Normally translations done via contractors tend to be of higher
quality. I also try to follow their progress and work personally and
encourage them to use Pootle's built in quality check mechanisms
(which catch at least half of the common mistakes translators make).
As for peer review, usually its a one person team (eg: Robson for
Brazilian Portuguese or Ravishankar for Hindi), so it is a bit
difficult to do that.
I think that with the Pootle checks, along with checks via msgfmt
(which convert the translations into the final MO files), we can be
reasonably sure about the quality. The msgfmt checks will also ensure
that wrongly formatted translation files will not make things crash.
Thanks,
Sayamindu
--
Ticket URL: <http://dev.laptop.org/ticket/10797>
One Laptop Per Child <http://laptop.org/>
OLPC bug tracking system
More information about the Bugs
mailing list