Plone 5.1 translations

I'd like to say that I feel uncomfortable with the quality of translation as well.

What the end user sees is the frontend and the wild mixture of languages. I know from my experience that this makes it much more difficult to argue pro Plone as the impression he gets is the one of poor quality even if you have the most sophisticated techstack.

I opened a topic a few days ago and the screenshot there is taken from Plone 5.08 - so this problem exists for several years now. The answer I got was to do some coding.
I am not quite sure if the need to be a developer is adequate just to edit some strings in the frontend.

I would do some translations english->german to get a better coverage if this is helpful, but I won't take a course in coding for this task.

Finally I'd say that as being multilingual is a promoted feature of Plone this problem is a marketing desaster.

Plone is great and has a wonderful community but it lacks the focus on the customer (the enduser) in my opinion.

Thanks for reading my thought and for rephrasing them in a perhaps more friendly way.
This is exactly the point besides other user-facing parts of Plone.

For contributing translations you do not have to be a programmer. The process is straight forward although you need to some git and Github knowledge. Translating is one thing, quality translations are a different topic. In particular you need to know the context of a text string and you need to have an eye on the wording of existing translations in a particular..otherwise you get a wild mixture of wording, spelling etc.

https://docs.plone.org/develop/plone/i18n/contribute_to_translations.html

My recommendation is to ship Plone only with languages that have a high translation rate. 83% for German ist not high for. Something like >98% would be a criteria to include a language with Plone. Better ship Plone without a localization than with a half-baked or completely unfinished localization (that's the case for German).

-aj

Thanks for making positive suggestions.

Because we are a volunteer driven project, we compromise on our releases. The thinking goes like this: "While we are in alpha or beta or release candidate stage, very few people test or work with the version. Until we actually release, we get few eyeballs on the version." That is why we have to ship before some people would argue it is "time". However, if you bear in mind that we control no one in this project, we only get the best that volunteers provide.

We have also heard many gripes about Plone 5.1 taking too long (does that sound familiar, @zopyx?)

I know we are all multifaceted individuals and we hold internally (and sometimes loudly public) inconsistent beliefs simultaneously.

So to go back to your suggestion, Andreas, if we were to follow that, we would ship Plone in English, and then you can imagine that other people (not you!) would complain that Plone isn't professional because it doesn't include any translations, Plone developers are English bigots, Plone is dying, Plone is used only in English countries, etc.

Now that we've beaten this horse well beyond death, we have some things we can do in the short term:

  • please help improve Plone's translations
  • Plone 5.1 is real! It's been released. Let's make it better.
  • Can we get some help getting in-progress versions tested and completed (including translations) earlier in the process?

What is the status of using transifex.com for translations? This would make it much easier for non-developers to be able to contribute. A developer/maintainer of plone.app.locales then can sync and commit the changes.

There is a document available, but also already "outdated": https://github.com/collective/plone.app.locales/wiki/Workflow-between-GitHub-and-Transifex.

We are using transifex.com very successful also some Plone add-ons and some other Zope applications. I think @macagua did the setup and can sync right now.

Any thoughts?

Hi @tmassman

You are right, I must take up this initiative, the reason for not finishing it in March of 2017, is because I lost my computer with all my data and backups. So let me review the issues to be solved and update the procedure and the files to transifex.

I disagree completely.

The Plone marketing claim about support for 40 languages is nothing but marketing, a marketing lie. There is only full support for a handfull of languages. Do we really have need for doing marketing with lies? Please explain if and why you see this differently?

I need to stress out over and over again that partly translated user interfaces are a major foul. They are counterproductive for the creditibility of any software. Existing Plone users can perhaps life with but it is unacceptable orgs evaluating Plone.
Your marketing works against itself.

Yes, it would be better to ship with a small number of supported and complete languages.

-aj

That's for sure!

Don't be afraid. My translations would pass your quality gate.

I agree.
It doesn't matter to me if it's all in english, but the non-technical enduser expect the frontend translated to their language.

This sounds very promising! I'm looking forward to it.

I agree completely. We should remove that promise and only ship in english. All translations become plugins and community supported. We need to be more realistic about our resources now.

1 Like

thanks, I always like your practical approach; these are the things I think we need to do in order to enhance the situation:

the following languages have enough Plone community members that can easily (in an hour or two of work) rise the bar to 100% translation rate: French, Italian, German, Netherlands and Brazilian Portuguese (I'm already working on that); I don't know about Traditional Chinese, Japanese or Danish, but seems doable also.

(on a side note, I don't understand how Finnish, Romanian and Catalan have so low rate of translations; any country looking for its independence for now on should have 100% Plone translations to be taken seriously by the international community.)

also, we need to start working to get all those new kids trying to participate on GSoC and coming from India, to start translating Plone into Hindi, Bengali, Telugu, Marathi, Tamil and Urdu, at least; its hard to understand how such widely spoken languages are not done yet.

we need to find out community members that speak Arabic and Russian (seems doable in Europe) to rise the bar on that languages also.

finally, let's face reality: Plone is not available in more than 40 languages and counting, so let's fix that claim on the front page of plone.org.

I would like to make my friend Thiago Tamošauskas happy and translate Plone into Lithuanian, but that's not going to happen any time soon.

covering the first 20 languages by native speakers in the world would reach almost 60% of the globe's population: that's good enough for me.

UPDATE:

be careful what you wish for… :rofl:

1 Like

I just completed 100% the translation of Plone 4.3 into Brazilian Portuguese and started the one for Plone 5.1; to be honest, it took me more that 2 hours, more likely about 4:

in my experience is not possible to translate the whole thing because there are dozens of msgids that have no default value, so it's impossible to know what they stand for; take for instance this:

…
#: plone.formwidget.recurrence/plone/formwidget/recurrence/browser/i18n.py:10
msgid "add_rules"
msgstr ""

#: plone.formwidget.recurrence/plone/formwidget/recurrence/browser/i18n.py:60
msgid "additional_date"
msgstr ""
…

how did you solve this, @erral?

1 Like

I loved and hated on the same level to translate the recurrence widget.

I did a trial-and-error approach: I installed the develop-egg of plone.app.locales in a plain Plone site and checked what all those msgids meant, and then I translated them.

To tell the truth the recurrence widget is very hard to translate without seeing it on the browser. And you need to take into account that not every languages' structures are the same.

I think the lack of default values is a bug we must fix:

I have no further time for this task now; anyway, translations to Brazilian Portuguese are merged and Plone 5.1 is now 96% translated.

1 Like

Quite off topic, but anyway:

When translating some of the norwegian I tested 'automatic Translation using TM'.
The result was not useful at all, not because it was bad, but because very often the result seemed to come from 'another field' (lets say a field with 'upload file' would be translated to 'your file could not be uploaded'.

Anyway: it made me think, and maybe this is possible already somehow?

When I translate (for example with google translate) from english to (lets say) dutch, the result is not correct.
If I then translate from german to dutch, the result is not correct, but it is different (probably) than the english translation.
If I then translate from french to dutch, I end up with 3 translations which are not correct (probably), but I am pretty sure that 'combined' they would make much sense.

Is there any possibility that when for example the first 5 languages have been translated another languages po-file could 'get a good translation based on those ' ?

Please don't do that, that's the worst you can do to get a translation. If you want to get a decent translation hire a professional translator, they have their own tools, to ensure quality not quantity.

My wife is a professional translator, and I've been involved for +10 years in community translations (for GNOME in my case).

Machine or semi automatic translations could help in some cases, but most of the time are actually slowing down the translation process.

2 Likes

GSoC 2018 project number 357:

1 Like

Statistics generated daily:

https://jenkins.plone.org/view/Translations/job/plone-5.1-i18n/

Click on the files to see for each domain how is the translation status.

And of course, there are jobs for 4.3, 5.0 and 5.2!

https://jenkins.plone.org/view/Translations/job/plone-4.3-i18n/
https://jenkins.plone.org/view/Translations/job/plone-5.0-i18n/
https://jenkins.plone.org/view/Translations/job/plone-5.2-i18n/

4.3 and 5.2 jobs are currently failing, some work needs to be done on them to make them work.

Enjoy!

3 Likes

it's possible to use the -t option instead? it will give us a better idea of what's going on:

2 Likes

Some thoughts:

  • Marketing is 'aspirational'... it reflects the reality we want it to embody. If there is no aspiration (please, no one make "aspiration sucks" jokes), there is no objective, and without objectives we stagnate.

  • If you're going to quibble about the "40", what would you want it to be? (/me slowly retreats from the argument)

  • Rather than quibble about the 40, let's spend the energy improving the translations. I know there is a PR already for the German translations by @jensens and @agitator that will be in the next release.

  • I love the ideas on making the status of translations more immediately visible! Those nice graphs are easily comprehensible. If we can display that status in an easily seen place, together with perhaps an easier mechanism for translations to be contributed, that would go a long way.

  • Next time, when we have a release in multiple alpha, beta, and RC versions, and our release manager asks for them to be tested, that would be a good time to jump in and contribute translations. I for one am eager to help with the English Canadian translation :wink: :canada:

2 Likes

sure, let's update the plone.org front page to this then:

I found it highly aspirational, but it's still horseshit.

the only way you're really going to motivate me is finishing the translations of Plone to Vietnamese, the sixteenth most spoken language in the world.