I would like to be the reviewer for Arabic translations.
The Arabic translation is managed by the translation agency
They are not doing a great job. They are very protective to some suggestions from community as it will look bad if someone improved most of their translations.
I understand it is hard to guarantee crowd contributed translations to stay on-top of latest changes (which it seems this agency are not even keeping up), but I suggest looking into making some community members to become a reviewer too as Discourse is all about being community ran.
Can you provide some examples that have been translated poorly? are there a significant number of issues?
Crowdin (see comment), the translation is “subscribe” instead of of signing up (which usually means creating an account).
Crowdin but seems like I also have suggestion similar to the current translation? I hate how crappy Crowdin, but correct translation is “لا يظهر للعامة أبدًا” by Discourse translator Bot
see comments in Crowdin and please let me know if it indeed an issue on your end too.
I don’t want to look like as if I’m saying that all translations are terrible but some needs a fix and waiting for months till someone sees it (like this post caused a review of my suggestions) doesn’t help.. Of course it is hard to keep big projects like Discourse translated from community and im not calling for cancellation of translation agency contract. But if there is something I notice is that these agencies inherit the bad translation of mega corps and not always translate according to context as regular user of the software.
I know you also added comments on
The process of changing approved translations often seems very difficult to me, too. And I am not sure what the best way to report them actually is. I know there are similar texts with comments in German, too. What works quite well is to comment immediately when the proofreader changes the translation (but it’s not that easy to review all their changes made during the review within a few hours before they disappear for the next 6 months). And this doesn’t help for older translations that need improvements. Sometimes I report them in the topic about mistakes in the German translation here on Meta, sometimes I only add a new translation and downvote the existing one, often I also add a comment. But so far nothing seems to be a good strategy ![]()
Proposing an improved translation is absolutely the right approach. Adding a brief note explaining why your version is preferable also helps a great deal. Tt gives reviewers the necessary context and makes the evaluation much easier.
At the moment, suggestions on strings that have already been approved often don’t get revisited. This is particularly true for reviews handled by the translation agency, where approved strings are typically treated as final.
To improve this, we plan to extend our internal tooling with a button that removes approvals from strings that have received new suggestions. It won’t run automatically; instead, we’ll use it intentionally before handing work over to the agency so those strings are clearly returned to the review queue.
The aim is to have this in place within the next one or two releases. I’m assigning this topic to myself to ensure it doesn’t fall off the radar.
I definitely think it would make sense to have some Discourse users have some kind of role as « subject matter experts ». We may not be professional translators but we do use Discourse pretty much daily in our language, so that does give us a certain amount of expertise on the question.
Who is better placed to determine if a translation works or not, somebody in a translation agency who might not actually use the tool, or active users of the tool in a given language?
That’s one of the reasons I think it makes sense to create small « collectives » of users for various languages. For example, when I stumble upon a bad translation and want to improve it, it would be nice sometimes to be able to briefly consult with my fellow French-speaking Discourse admins.
I have dealt with similar situations in my work-life — I work in a big multilingual company, with a team of internal translators and a whole network of agencies they outsource work to. Most of the translations of technical or niche business topics are really bad, because the translators do not have (cannot have!) sufficient context or background knowledge to do the job well. In my discussions with the translation team to see how we could improve things (I am part of the linguistic minority), they pointed out how important it was that the person providing a document to be translated make available « subject matter experts » in the destination languages, so that the translators could contact them if needed to understand what the source material is about and have input on the vocabulary used and understood by the target audience in that language.
I think there is really a role to play here for us.