Title says it all. This would be useful, just in case someone has already translated some strings and submitted to language maintainer and strings are waiting for approval.
Just to save another zealous translator's "labour in vain".
Sorry, koen, I do not see what you mean by those "new" or "review" statuses.
Sorry Joseph - it is possible that I'm the only one having the overview of all contributions where nothing is done with in all language packs. I have to explain myself:
At the moment there are over all the language packs about 280 contributions in the status "new" or "in review", so those strings are not in the language pack. All together about 32600 strings. That is really sad.
I don't understand how it got to that state: Amos sends an email when there is a contribution, there is the contributions link that shows the contributions, when there are contributions, it saves a lot of typing, ... I don't know.
What I meant was that anything that can help to migrate these 32600 strings towards the language packs in a responsible way or can reject the contributions who are not good, gets my support.
So from my point of view, if the "waiting for approval" strings filter you propose makes it through development, it can be switched on by default for me
thanks for the information and numbers of contributions waiting for confirmation. I think our common interest is that Moodle languages are as complete as possible. We also know that most of the translators are translating from time to time and not very continously. May be we also have a group of maintainer that are inactive .
My idea is to make differences between types of contributions:
- core strings that are translated the first time
- core strings that should be changed (spelling errors, better translation)
- third party module (first translation)
- third party modules (changes)
Core code strings are more relevant than third party modules. Relevant means we should be interested that core is translated as complete as possible. If third party plugins are not translated well it is not really a big problem.
Sometimes I have the problem with checks of third party module translations if I don't know the tool and can't check if the terms are used in the good way. My practice is to check the quality of German language in translation, correct grammar and spelling errors and wrong use of core terminology (i.e. role names).
Sometimes I think, getting a bad translation is better than a string that is not translated. If someone sees a bad translation he will offer a better one. This is also easier than to start a translation from the ground.
What do you think about this idea:
- core strings that are translated the first time are accepted automatically if they are not checked by lang maintainer after 4 weeks.
- core strings that should be changed (spelling errors, better translation) are not accepted automatically.
- third party module (first translation) are accepted automatically if they are not checked by a maintainer after 4 weeks
- third party modules (changes), accepted after 4 weeks
I don't know if it possible to implement a check with this criteria.
Does it make sense?
Otherwise I think that we need more professional translations of the packs or a concentration on the 30-40 most relevant languages. But there is no money for more professional translation and I'm not able to decide about the relevant languages.
Thanks for the idea Ralf. I must disagree though. We can't blindly approve anything.
Imagine there is a language pack with the maintainer who is not active or is on a longer holiday or whatever. Now somebody creates an account in AMOS and translates new 2.7 Moodle core strings. They are hidden to other contributors, thanks to the new great feature I will implement based on this thread. Four weeks later, sites in that country are upgraded and the automatically approved language pack is distributed from official Moodle servers. What a surprise - people read things in their own language like "Angela Merkel and Barack Obama are capitalist bastards who made your country poor" because that is the spam translation that got into language pack. I am not naive enough to think that media would even try to be fair in this case. They would love to report about what kids are being taught in this Moodle software.
Sorry for being rude here intentionally, I wanted to illustrate the seriousness of the risk. Moodle is translated into many languages. My computer does not even have fonts required to display all of them. Having translations as crowdsourced as possible is our current strategy. But we still need people that review, approve and sign off things. People that we trust to.
Professional translations (like yours) are great, for sure. But let us admit, it's unavoidable luxury for most languages we have. So we really must focus on supporting the community contributions here And indeed, AMOS has big potential for such improvements.
not sure about the automatic check in of strings, as I think it is better using original English strings rather than a bad or misleading translation, not to mention the David's extreme cases. The issue anyway, at least for Italian lang pack "core" strings is almost non existent as from October 2010 till now the contributions accepted has been slightly over 1% of the total strings and are mostly typo fixes, so review for core strings contributions up to now has been quite fast and are absolutely not a problem.
For contrib modules, things could be different but again I still think better to review them rather than automatic check in. Since I often find myself asking for improvements of those contrib strings, maybe an AMOS contribution status type improvement could facilitate the workflow.
I agree with David and Andrea that automatic approval is dangerous for several reasons. After all, the language packs end up in php-files that run on a server. Apart from unwanted messages, it could get really bad.
Deciding about relevant languages to concentrate on is indeed impossible: any language is relevant for the people who use it.
What about a limit of how many strings that can go in one contributed chunk (or is it stash)? I've seen contributions containing over a 1000 strings. I wonder how you start reviewing something like that. I usually suggest 30 max, so I can review them in 5 minutes if there are not too many problems in it.
I can understand your concern. I made this suggestion based on our experiences with German translation where we never had any problems.
Koen, we get contributions from universities that translated 3rd party modules internally complete and put the translated php file into AMOS. This creates easily contributions with 200-400 strings. A limit will not make sense here.
As a language pack maintainer, I must admit that when it comes to 3rd party add-ons, I am much more benevolent and usually just skim the stage before committing it. As said above, I often do not know the plugin and also I know that it will typically have much less users (if not only those who submitted the translation).
When i first read your suggestion, i thought it was a good idea, but when reading the whole thread, i admit it can be a great risk!
I think that, better then making an automatic validation, AMOS should better re-send a new mail to the language pack maintener after some weeks.
And perhaps create a dashboard to see if some language pack maintener still don't validate/reject several propositions after some more weeks. It would indicate we should try to find someone with more time to do that...
Just an idea.
I like the resend idea Severin.
It would solve the problem of a contribution been send to a non active language pack maintainer who is later replaced by a new language pack maintainer that didn't get the email.
Dashboard: "Your language pack has 2 new contributions and 4 contributions in review" - Cool