[Localization] [ANNOUNCE] Pootle migration

[Sayamindu Dasgupta]

On Mon, Jan 11, 2010 at 9:08 PM, Chris Leonard <cjlhomeaddress at gmail.com> wrote:
> On Mon, Jan 11, 2010 at 7:40 AM, Sayamindu Dasgupta <sayamindu at gmail.com>
> wrote:
>>
>> Hello,
>> Just a small heads-up - we are tentatively going to upgrade to Pootle
>> 2.0 at the end of this week. I don't have a fixed date and time yet (I
>> just uncovered a problem with Pootle's symlink handling which causes
>> issues with our entire git<-->pootle workflow), however, I do expect a
>> downtime of around 24 hours (since the metadata and indexing data
>> associated with each PO file need to be refreshed as a part of the
>> migration process).
>> I will keep this list posted on how things go
>> Thank you,
>> Sayamindu
>>
>
> Sayamindu, thanks for all your work on keeping the SugarLabs / OLPC
> localization infrastructure current.
>
> Everyone should be aware that the Pootle 2.0 offers a substantially
> different look-and-feel on the Home page to the Pootle ver 1.x servers we
> have been using to date.  To get a sense of what the Pootle 2.0 interface
> will look like, you can look at the Pootle server where Pootle itself is
> localized.
>
> http://pootle.locamotion.org/
>
> Those of you who may have been following along on Pootle users list will
> know that Pootle 2.0 is truly a major revision of this tool.  It still uses
> the Translate Toolkit, but the web-serving elements have been completely
> re-based on Django.  While the actual string translation interface is
> prettier, it is still fairly familiar looking, so there should not be much
> of a re-learning curve for localizers.  The re-basing on Django is said to
> improve the ability to perform administrative tasks via the web-interface,
> including better account and password management. which would be a win for
> both admins and users.  I'm looking forward to working with the new version.
>

Ah yes - among other things this will _finally_ let users reset their
password if they forget :-)

Pootle 2.0 should be _much_ _much_ better than what we currently have
now, from both a UI perspective, as well as performance-wise. We will
be switching to MySQL from sqlite as the backend store for user data,
metadata of PO files, etc. This, plus the use of memcached will help
us a lot in terms of scaling up as we progress (we currently have
~1400 users registered, and serve around ~6300 PO files, which is
itself a pretty large number)

Thanks,
Sayamindu



-- 
Sayamindu Dasgupta
[http://sayamindu.randomink.org/ramblings]