Before we go with community translations, I'm currently discussing about how it can be done right. There are some considerations. Like potential inaccuracies in translations or translations that don't get updated often enough. So it's most likely that the english texts are going to be proof-read many times in the future before being published (and translated). Some translation guidelines are also needed.
Personally, I want the website to be translated in as many languages as possible (as long as each transaction is accurate and good quality). So you can count on me to come back about this if nobody opposes to this idea.
Thanks for offering yourself!
Congratulations on the upgrade of your work to bitcoin.org!
About the guidelines and the translation process there are the following issues I see:
1) The translations must be kept up-to-date
2) The quality must be high
As I understood this project is set up this way (on github where pull requests are invited) to be able to translate everything, including things like urls and rewrite rules thereof. But there are downsides to not using a tool like transifex: Especially issue 1) becomes a problem I think.
Is there an easy solution to let translators know which sections in the source (English) have been updated since their last translation efforts? Maybe a very detailed summary should be written every time something changes in English? I've noticed in bitcoin-qt that sometimes a wording in the source changes only slightly but results in a big difference in meaning. If you don't know there has been such an update it is very easy to miss it when you compare a source and its translation side by side.
Regarding the quality issue (2): Should we maybe require a small team per language, so that we only let a language go live if there are at least say three people vouching to look after the translation?
Surely there is some experienced web developer here who has managed a high-profile multi lingual website?
I'd be very happy to contribute the Dutch translation, I've done also the bulk of bitcoin-qt so I can make them consistent.
Github clearly is awesome to track updates in files. So I would say, yes, it's a little geek, but it's very efficient. I didn't go with transifex because this is only .html files so most of the content is only pure human readable texts to translate. As long as a translator doesn't touch the rare tags (
), then everything is ok. And even if it happens, it's easy to fix.
The second reason is that the structure of the site might actually needs to be adapted in some cases. And this setups makes all translated version independant of each other. So that if only one translation is not up-to-date, it won't prevents us to add more content in all others. All the opposite of a software that only have a few dialogs and fallback to english when nothing else is available, the website contains long texts and I think this would not give a good result.
Also, someone does not necessary needs to be a github expert to contribute. In the worst scenario, simply downloading the .html files, translating them and sending them back by email can do the trick. Github pull request only saves us time and allows us to work more efficiently.
Indeed, I think that having a more than one translator for each language (at least one native translator and another native speaker to review) sounds pretty good. However, I have almost no experience at leading such things. So any help will be appreciated.