Well, the fact that the corpus is under CC-BY means you have to mention Tatoeba if you are going to reuse it, not that we have to indicate the original contributor in the dump.

It works like this : people provide their work under CC-BY, so the attribution of the work of each user is mentioned in Tatoeba itself, through the logs. But then we're reusing the work of everyone to make something else: the corpus, which we also redistribute under CC-BY.

But now, if you also need to have the username of the original contributor, or any other information that is not provided by default, you can just ask and I'll see what I can do =)

Hey, hey! I'm bourdu, collaborating with parent poster to have a finnish-french-english-japanese corpus. Keep up the good job!

Hi vbkun, and welcome :)

We haven't considered adding Brazilian Portuguese, for one thing : because no one requested it. But also, it's not exactly a new language.

I feel it is a problem similar to the distinction between British English and American English. Perhaps the difference is more significant in the case of Portuguese "Brazil vs. Portugal" though, but technically, it is still considered as the same language...

So right now, the best I solution I can suggest is to add a tag [Brazil] at the end of sentences that are in Brazilian Portuguese.

Merci, c'est corrigé.

No, you don't have to preserve them when adding new translations. They used to be used, and they may be used again someday, but right now they don't mean anything anymore to the system.

You can find a short explaination at the very bottom of this page:
http://tatoeba.org/eng/pages/do...mple-sentences

True. We installed our own pinyin converter and started using it recently, but perhaps the data we used is not complete...

Anyway we switched to the old way so it is now working again for the traditional Chinese characters.

As for all other language, we're opened, we just need people adding some sample sentences in this language, as soon as you can provide us around 5 sentences, we can consider adding it.

That's because back at the time when I introduced the flags as the indicator of the language of a sentence, I felt that the real colors were too strong and needed to be lightened a little bit to fit more to the overall colors of the site.

But you're right, it leads to issues such as having a flag of a country looking like the one of another country.

But anyway if someday luxembourg language is added, its flag too, and it will look lighter than dutch flag :) (and the contrast between red and blue is kept as the red is also lighter)

Thanks for the note. I've corrected them.