Our data is released under various Creative Commons licenses.More information
If you love this content, please consider a donation.
Menu
MessDjaaf, why not just translating existing English and French sentences?
Well, I do do that from time to time. When I find interesting sentences. Because I don't like writing ANYTHING just for the sake of writing. I love my language and therefore I try to write the best sentences with different styles & vocabulary. That way I will contribute to its development, because actually the written form of Berber / Kabyle, using Greco-Latin characters, is very recent. In addition to this, as I said in my last post, Tatoaba is linked to an online dictionary that uses Berber (Kabyle), French & English.
(Sorry for answering to a question directed to someone else, but it’s asked on the wall and not in private messages, so I assume you wanted us to participate in the discussion.)
> why not just translating existing English and French sentences?
I’m very opposed to this point of view, because this way, English and French speakers will end up deciding what gets translated. I find this both:
- detrimental to the quality of the corpus (because English and French speakers don’t know many things that are common in everyday life of speakers of other languages, so many things will be ovelooked and neglected), and
- unfair (because you shouldn’t get extra privileges just because of your place of birth, and the current system means English speakers get the privilege of deciding what will be translated just because they were born in some kind of family).
I think the ‘only contribute sentences in your native language’ policy should be abandoned, and native speakers of other languages should proofread other speakers’ sentences instead of discouraging us from contributing.
I’ve wrote more about this in my comment to CK at #4132859
I couldn't agree more.
+1,000
Well, everything you said is correct and it should be so... in theory.
There are several other issues but just as an appetizer: MANY people don't even read the rules of the website. I think it's quite an obvious thing, and it doesn't take to be a corpus maintainer to see it.
Tatoeba is not lang-8: every other day, a new guy will come and add his sentences waiting for them to be translated.
We want complete sentences: well.... (And I don't think about nominal sentences on which people can disagree about the definition of "complete sentences")
Adopt orphan sentences only if they are correct: well, being the corpus maintainer for the French corpus, I can tell you that some people didn't really pay attention to that point xD
etc., etc.
Personally, if I cannot trust new users to even read the quick start guide, I cannot really trust their first contributions (let's call it their trial period). Hence, not encouraging newcomers to contribute sentences not in their native language is not outrageous to me.
For users with more experience, it may be a different story. But how to gauge this "experience" is a whole different topic. For example, I can easily imagine people adding foreign sentences they found somewhere else on the web, forgetting the copyright infringement. Or, to avoid it, replace some words, at the risk of making the sentence awkward without realizing it (because of lack of understanding).
> As long as he is able to adequately assess his level of English and abstain from adding sentences he’s not sure about, I don’t see any problem with his contribution.
> and native speakers of other languages should proofread other speakers’ sentences instead of discouraging us from contributing.
For "popular" (ugh, what a bad word) languages, it is a lost battle. There would never be enough proofreaders to ensure corpus sanity. There are already not enough while the current policy is underway...
But in my opinion, the problem would be solved if the policy was changed to promote (actually promote) "ORIGINAL contributions in your native language". That is, people are encouraged to actually WRITE their own sentences, and not only translate them. Hence, the problem of language-biased corpus you mentioned will be reduced, there will be more variety in general. Then, people could translate their sentences into any other language if they are sure about that (as you mentioned), or ask somebody else to check it.
However, this then goes back to people not reading rules anyway and doing whatever they want. Some people will say that it's a design problem, or a matter of finding the information. But really, it is just people never reading the rules.
> I think the ‘only contribute sentences in your native language’ policy should be abandoned, and native speakers of other languages should proofread other speakers’ sentences instead of discouraging us from contributing.
That would be OK if we had a lot of proofreaders, but this is not the case. We all know that errors often stay in the corpus for several years. Many languages have only one corpus maintainer and most languages don't have one at all. I don't have time anymore to write my own sentences. All my time goes to proofreading, to correcting, and - what is most time consuming - to trying to explain why the sentences are wrong.