Actually at a moment M, the one at the top is the "main" sentence from which you see the sentence directly connected to it (bright green) and indirectly (means it's a translation of one of the bright green translation) connected sentence (pale green)

So after if at one moment you two sentence directly connected that both taken separatly means something, then they're not one "incorrect" that should be changed according to one "original", but more two "group" which need to be split and a new sentence (actually two), to be added.

browse -> browse by audio, ;-)
the audio tag has been created I think for those who do automatic treatment on it, as that's the only to get this information exported for the moment.

I think for common languages such as french , english and so using the two letters code (which is also a standard, iso 639-2 alpha 2), may be easie than their 3 letters counterpart (and anyway that's a one to one equivalence, it's just that the 3 letters code is able to represent much more, but fr is strictly equal to fra , and zh maybe for example more easier to be understood than cmn) after yes I've never said that they will be black and white , some nearly-random colors can be a solution (I say nearly random because we need to take care that in some cultures this or that colors as not a negative cultural meaning, for example in Chinese if you write the name of someone in red, that's something usually used only for dead people, if you see what I'm implying)

after what I think will arrive, is that to somepoint (yet to define, when we will think we have rich a large enough audience and that we have a well known """""authority""""" on languages) we will propose a remplacent system, that will co-exists for sometimes with the current flag system, enough time to let user start to get used with it, in order to avoid the "one day flag, the minutes after our new custom representation system" effect.

anyway I think for the language for which we already have no proposition of flag (if the natives/contributors for that language don't find a suitable one, note: it's not always a "nation" flag, cf the Wu one), we can already start to use that , as anyway both things (3 letters code + arbitrary chosen colors VS a arbitrary chosen language) will both be "unknown and new" to people.

and anyway that was a boracasli fake account

I agree for you with flags, when Tatoeba will be well known enough ( I know, define well known-enough), I will see with everyone to find something to replace them (the 3 letters, or two letters when the two letters, iso 639-alpha2 are well known, such as FR EN) But for the moment I think that will repulse too much "mainstream" user. I mean let's be pragmatic, there's no point on trying to make people become aware of that things ( 1 country != 1 language , there's country with several languages and languages without country etc.) if no one except already convinced people use Tatoeba. So I see that in two times , first become well-known and then starting to militate

As I know some of you work with the Tatoeba data and its graph representation, I've found this yesterday

http://sigmajs.org/examples.html

it's a JS library to display visually graph, if someone feels like doing a proof of concept to display a sentence and its translations graph like this http://sigmajs.org/examples/more_node_info.html ...

I've updated to advanced contributor both of your status, you may need to relog.

It means that now you can look at random sentences in these new languages and search for them in the search engine (I've solved the problem also for the Novian language)

The languages adding process is now finished (for this time, of course now that I've documented for myself the process of adding languages, it would be not that long anymore to add new ones)

I will try ASAP to reupdate the interface files to be able to translate them on Launchpad (as they were spoiled by Bora if you remember),

On the same time I will start to take care of all the unanswered emails/private messages I have (about user promotion etc.)

I will take a look at it after dinner, thanks to have reported the issue.

Yes sorry for this, these days I'm managing Tatoeba all alone, and my professional life is eating all my time leaving few times for my hobbies (included Tatoeba)
Last week I've made some upgrade on the servers etc. and added languages and for this week I will try to unpack the long list of emails (included yours), sorry for the times it takes

actually I've just seen that you're only normal user for the moment, so I think a first step would be to promote it to "advance user" before thinking about becoming corpus maintener.

so it's already possible to start contributing directly in these languages now.

5AM here, the languages have been added, only one step missing, to update the search engine in order to be able to search form them (and also to use the "show random in", that is actually based on the search engine)
tomorrow that should be finished

I'm on the way to add the new languages (3AM here so I'll first go to bed), they should be soon here (I've been busy this weekend upgrading the server, and my personnal server on which I have the test version of the new version of tateba)

Why, not, we'll first ask what the other think about you as a moderator, as last time I said that I will involve more the commmunity in the process of chosing who becomes moderator.

then possibility for other users to say "X and Y consider Z as a good contributor in French", well after it would be sort of eggs or the chicken problem

possibility to add reference to the litterature then ? even if that's simple sentence then you would something that one can trust (if you put a line of "The times" or a quote from a Orwell book) (that should not need to be exported, so we will still be able to do that as the "for education purpose" exception)

actually to my own personnal experience, even as a well-educated French native, I'm pretty sure to get at least 30% of my contributions that contain mistakes, (ok not huge ones, but still)

For the moment we have the "OK" tag, far away from being a long-term solution, we all agree on this. In the future, as I will permit people to enter in which language they're native /not native , and the proeficiency too, maybe too permit the natives to click on a "I think this sentence is correct and natural" and after to get somewhere the list of people who clicked on it (this way one after will be able to check the profile/contribution to see if the guy is reliable, and that will also permit to report abuse if someone fake to be native and start spoiling the system)
so after that will permit one to look for the sentences that the guys X, Z and Y consider as correct (except S that you don't trust)

What do you think ?

Some news:

* I've made a lot of optimization these days on the server, so it should be able to support more users (we're now at 150 000 visits/month and nearly 750 000 page views/month, which is not that bad)

* For the language detector system, i've finished to code it, in a usable version for Tatoeba, I now only need to install it on the server and to update the current tatoeba code to use it (for those who asked me to pre-test it, unfortunately providing a "user" interface will need more work than provide a simple API version usable by a software (tatoeba in that case), so I think I will directly put it in production (I've made a lot of test and we're 95% of detection with sucess)

* I will need to ugrade the version of the Debian Linux used on the server, so for this I will need to shutdown tatoeba during certainly some hours (depending on how many things will get broken during the upgrade)

* After this I will add the languages, so they will be added during this weekend.


* for the new version, I've updated the code to now use the stable version of the c++ framework, and added some little stuff in the API, so after I should be able to focus again on the new version. (as I would have solved all the problems of these days, server load, the language detection, the growing list of languages waiting to be added and the server to upgrade)

So well I'm quite happy on how things are going

Toutes nos pensées sont avec toi Eldad.

I've deleted the spam from both users