that's the reason why they're limited to trusted users only, once we will implement all what you've said (and I agree that we need them), the feature will be open to everyone. But you know no time blablabla ... ^^

yep, but when I will have time, I've spend the few free times I had this weekend, to fix the duplicate script, and a bug when changing the flag of a sentence (it didn't update correctly some internal structures)

Mais il est vrai en tout cas qu'il serait intéressant de voir, surtout pour l'Anglais, le nombre de phrase réellement différentes. Il me semble que dans le même ordre idée, voir un peu la richesse réelle du corpus, CK avait établi des statistiques sur le nombre de mots différents dans le corpus, et la fréquence de certain. Ce qui n'était pas forcément très glorieux (c'est pour cela que dans les versions futures, il nous faudra trouver divers moyens pour favoriser l'ajout de nouvelles phrases, et si possible contenant du nouveau vocabulaire)

oui car ensuite j'ai fait un "select count(*) , text, lang from sentences group by text, lang having count(*) > 1" qui me montre les phrases en double, et il n'en reste plus qu'une dizaine qui ont rajouté entre temps (le script a tourné hier en soirée)
Mais le nombre de phrase anglaise a baissé, après il est vrai que beaucoup des phrases anglaises se ressemblent parfois à un mot près ou à un singulier pluriel près.

so it's why the number of sentences in some languages may have decrease

by the way overall, we have now more than 500 000 sentences ! (515 000 to be exact)
Congratulations to everyone, some months ago I would have never imagine we will reach it so fast. Next step, 1 000 000 :p

I've run the remove duplicates script

I think it's somewhat a "request for sentence"

something like "hey folks, I want to know what do English people say when they want to go the cinema and they've missed the first projection, they're at the cinema already, and want to ask when is the next one"
and a nice guy, with a blue-backgrounded avatar will answer by inputing a sentence in tatoeba :p

@aloanguyen is this what you want ?

yep good idea, I've added it to the things we will need to implement on the next version http://www.assembla.com/spaces/tatoeba2/tickets
(by the way having this ticket system open to every members can also be a good idea ^^)

I've said something boracasli, so you know what will happen when I will be back...

In fact it's a bug from the software we use, see http://tatoeba.org/fre/wall/sho...7#message_2530 :)

not difficult, too long in the current architecture, so yep if we were able to warn "real time" mod, we will be able to merge on the fly ^^, or make whatever action. You can imagine I wouldn't spend time maintenaning the mythic remove duplicate script ^^

But fortunately the database Zmoo of sthooka project has created, and upon which we will build the new tatoeba, is able to detect them on the fly. So what you've said will be possible in the new version.

http://xkcd.com/486/

unfortunately
unique sentences in Turkish : 1934

(you haven't seen me , I'm not here, thanksbye)

*Sysko runs far away with his ninja costume "hehehehe" *

Note to everyone: For some personnal reason I will be EXTREMLY busy this 2/3 following weeks, so I will need to restrict my activity on tatoeba to the mininum during this period. (basically check one time a day that nothing goes wrong)

So don't worry if you get no answer of feature request etc. during this period, it's not I don't want them, it's just I've no enough time yet. Most of them are interesting (as the idea of a more complex timeline proposed by ludoviko) and will be included in the new version (yes the mythic new version)

So be nice, don't play with flamethrower, and I will be back ASAP, you can still send PM, but also send it to Trang if you really want a quick answer ^^

Yep, because it's easier to say "this sentence contain only vocabulary / grammar pattern that someone who wants to pass LANGUAGETEST level X is supposed to know", moreover there's official list, and it's more valuable for learners, as this way they can directly filter sentences depending of which test they prepare

(for email yep I agree with you, and it was the example I've taken while chatting with Trang ;-) )

1) I agree, but the problem with "offline" edition is that, not everybody have it, and for "new words", you must have a recent edition. I say that not because somes can lie etc. (I'm sure we're between adults sharing the same goal), but for someone learning French (or other) and wanting to know if such or such word he heards is now commonly accepted as part of the French vocabulary, he will need to know where to search.
Moreover only those having the book will be able to "moderate". But I do agree most of the times these ressource are offline (for shanghainese the only real dictionnary (I'm not talking about all the commercial website pretending teach you "standard shanghainese" etc.) exist only on paper)
So I'm not against paper "authority" reference, but I would prefer, if we're hesitating between one having an online version and one without, to choose the online one.

2/3) I agree

4) Sure, I think once it will be defined, it will be easier to see if a user (and moderator) respect it.

my bad for the French audio, I'm the one who recorded them.
I'm checking them to find sentences where I've forgotten to pronounce the liaison, thanks for noticing.

crap will be all what is not recognized as french by the authority source (maybe http://www.larousse.fr/dictionnaires/francais ? it's both 'up to date' and widely recognised as a 'serious' dictionnary, but it's only a proposition), at least for words, after for grammar pattern I have no online ressource in mind yet.

for authority, I've met Trang last week, we will define (at least for French, for other languages, we will need to discuss language by language with natives) authority source, to avoid as much as possible the crap.