← Back to team overview

lp-l10n-fr-community team mailing list archive

Re: French translation

 

Hi Kip, sorry for the late answer,


I wish there was some way for Launchpad translations to provide a link
for each message to where it is found in the source code for as a last
resort to indicate context if one isn't explicitly provided. There is a
way for gettext to automatically insert file and line numbers into the
generated message catalogue template, but they're not machine readable
by Launchpad as far as I know.

Launchpad can display these just fine.
For example:
https://translations.launchpad.net/pybik/trunk/+pots/pybik/fr/+translate
(notice the "Located in" string at the bottom of each field)

So yes, this is one of the things you could provide to help translators. Though this is not THE solution because not all translators know their way around source code.


Personally, I can't translate this string because I don't know what it
is referring to.
Is this a menu item? A window title? A button? (which does what?)
Is this a kind of editor for hardware? (which hardware?) or is this an
hardware implementation of a kind of "studio"?

In this case, it's just a field heading for in the software credits
listing notable software libre used in its creation.

So, what's the difference with the 2nd string, "Studio Software"?

And, as I said, I just took the first untranslated string ;-)

It's ok. Although most of the successful translations so far on the
software were able to do it, it might have happened faster and easier if
they knew the context better for each string.

I would really like to know how they managed to correctly translate this kind of string without context/asking.
Maybe I'm missing something.


There are two ways for you to improve this situation.
First, you should improve existing gettext contexts and provide them
for most (if not all) strings.

From a programmer's perspective, it's actually more difficult than it
sounds. Keep in mind that the software you are translating was mostly
written first, then i18n'd after.

This is indeed a lot of work, but I think it should be on you TODO list so that the situation can gradually improves over time.

Secondly, you should document an easy way for translators to actually
test the software with and without their translation (gettext context is
not always enough, good translations are often tweaked so that the
language style is consistent across the app and depending of the UI
design where they are actually found).

That process actually isn't unique to this software, but should be
standard across all reasonably compliant GNU Coding Standards friendly
projects. e.g. configuring with --enable-nls, setting LANGUAGE
environment variable, etc. But personally even that standardized
process is still confusing not just for many translators, but
developers too in some cases.

You could, for example, provide a PPA so that translators could then install the latest snapshot with the latest translations from launchpad (and it could be updated automatically daily, as some other projects do). This way translators could easily test their translations without having to know how to build the project.
Don't forget that most translators are not developers.


I'm aware that this may require a lot of work but, well, this is the
cost for ensuring a good translation quality when it has to be done by
third parties who don't know much (yet?) about your project.

I just realized that you probably didn't see our original post and
that's my fault. I probably should have provided that on the
lp-i10n-fr-community list.
[snip]

This is indeed helpful, but I was rather referring to a deeper knowledge level of the project. What I wanted to say is that if a translator wanted to take very good care of Avaneya translation today, she would have to dive into the source code to learn about the context and build the software to test what she translated. In the end, this translator would need to acquire nearly as much knowledge about the project than a potential developer.

There are indeed such translators, but not many :-)
As you probably understood now, my point is that you should do whatever you can to lower the technical knowledge needed to contribute good translations to Avaneya.



Cheers,
Nicolas


Follow ups

References