ubuntu-phone team mailing list archive

[Peter Bittner <peter.bittner@xxxxxxx>]

The problem of context for translators is not new, yet still it has
not been tackled to the full extent possible.

I have discussed this, e.g., with Transifex who try to provide
"frontend translation" on Websites using JavaScript. It's clear that
when they use JavaScript to extract texts (as an alternative to the
tool-driven approach to extract texts from source code) they can
relatively easy provide context by displaying the text origin of the
Website in an IFrame next to the translation panel.

This is not all that easy when you extract texts from source code. Yet
still the extraction tools have built-in features that would allow
adding necessary information. The magic word is "contextual markers":

- Example for C/C++ (and others):
https://www.gnu.org/software/gettext/manual/gettext.html#Contexts
- Example for Python/Django:
https://docs.djangoproject.com/en/1.8/topics/i18n/translation/#trans-template-tag

Instead of simply the message context field could accommodate an URL
to a site that provides up-to-date screenshots. Or one could build a
(REST) service that yields you a screenshot (or an informational page)
for a given resource path that doesn't necessarily change; the
screenshot could therefore be shot even after extracting the texts,
e.g. when front-end (visual) tests are running.

Consequently, translators could work on a platform that automatically
displays the context for a translation resource.

How does that sound?
Peter



2015-10-22 12:40 GMT+02:00 Ivo Xavier <ivofernandes12@xxxxxxxxx>:
> Hi everyone,
>
> I'm also a translator of the ubuntu portuguese team. In fact, the
> translations process or the "process of translating" is not easy as it
> seems. Because very often we don't understand the context of the string (at
> launchpad), sometimes we get things correct at first attempt, sometimes we
> don't. And errors have a high price... in the portuguese (form Portugal) we
> have a bug in the ubuntu store (for months), because, until the moment, the
> package hasn't been update yet, and yes... we've opened a bug, but nothing
> change since.
>
> I know how hard the devs are working on, but I think it would be very
> helpful if, at least, we receive an email (mailing list?) with some
> screenshots when new things are added. Example, the address-book-service
> received new strings to translate, how can I reproduce them to analyse the
> best way to translate it? I'm on Stable channel, and my laptop isn't the
> fastest to play with the emulator..
>
> Also, I think that, we could have a person that was responsible for the
> translations in general. Someone that we could contact if something isn't
> correct, and that person could be responsible to manage the mailing list to
> send the screenshots. Why I'm teeling this? Because, in some cases, opening
> bugs for "errors in translations" on apps or something else, we get some
> answers that makes us think: why I have a Launchpad account? If, that person
> already exists, please tell us...
>
> Off course, having two phones would help a little bit, but I can't buy
> another to use it on rc-proposed. Although, we felt very happy with our
> performance, from system-apps, core-apps, system-settings, scopes and
> third-apps (uNav, Dekko, Telegram, uReadit and more) everything is down to 0
> on untranslated strings.
>
> Keep the good work people!
>
> Thanks
>
>
> 2015-10-22 10:07 GMT+01:00 Peter Bittner <peter.bittner@xxxxxxx>:
>>
>> (just a comment)
>>
>> ... the same is true about changes to source code, be it bug fixes or
>> added features. What is the process here in detail?
>>
>> User facing components could update also in between releases, given
>> the idea that such changes typically don't entail incompatibilities
>> (read: "API changes"). Translations would fall in this category.
>>
>> Just for a common understanding: After all, Ubuntu Touch core apps are
>> just apps, like any other, with the exception the they carry the
>> promise of "Canonical safeguards platform conventions here", and
>> Canonical is expected to do the QA for them.
>>
>> A question that come naturally: How far is the process of test
>> automation? Are there, say, acceptance tests written in Gherkin
>> language (for each and every feature, one day)? If this is not the
>> case or not planned would it make sense to come up with a volunteer
>> workforce dealing with writing acceptance tests (specs +
>> implementation), and Canonical could focus on the automation of the
>> whole process? The end result (vision!) could be that we have such a
>> high confidence in the stability of, say, the core apps when all
>> lights are green that intermediate releases can happen automatically
>> or at least semi-automatically with a dramatically reduced effort of
>> human QA supervision.
>>
>> Would be interesting to see this happen. Any thoughts?
>> Peter
>>
>>
>> 2015-10-22 9:11 GMT+02:00 Krzysztof Tataradziński <ktatar156@xxxxxxxxx>:
>> > Hello,
>> >
>> > I'm working everyday as hard as I can translating strings into Polish.
>> > But
>> > very often when new OTA is released, there are some texts to be
>> > corrected.
>> > Sometimes because of our mistake and sometimes that we do not know how
>> > environment of the text looks like, is it a button, etc.
>> > For now, when we found some error with OS, we must wait for next OTA and
>> > live with that 4-8 weeks.
>> > Is there a possibility to add some small updates for translations only
>> > after
>> > OTA (maybe 3-5 days after OTA)?
>> >
>> > Best regards,
>> > Krzysztof Tataradziński
>> > https://launchpad.net/~ktatar156
>>
>> --
>> Mailing list: https://launchpad.net/~ubuntu-phone
>> Post to     : ubuntu-phone@xxxxxxxxxxxxxxxxxxx
>> Unsubscribe : https://launchpad.net/~ubuntu-phone
>> More help   : https://help.launchpad.net/ListHelp
>
>