Our data is released under various Creative Commons licenses.More information
If you love this content, please consider a donation.
Wall (7,124 threads)
Tips
Before asking a question, make sure to read the FAQ.
We aim to maintain a healthy atmosphere for civilized discussions. Please read our rules against bad behavior.
sharptoothed
10 days ago
sharptoothed
10 days ago
TATAR1
10 days ago
AlanF_US
11 days ago
sharptoothed
12 days ago
Shanaz
15 days ago
Qaztat
15 days ago
TATAR1
15 days ago
Tartar
15 days ago
menaud
18 days ago
Note: I'm really sorry I didn't /don't answer to your emails private messages/e-mails these last/following days, but I'm currently in a rush with my professional life.
I hope that will be back to normal soon.
[fra] désolée si ces derniers temps je n'ai pas répondu/reponds pas (ou avec beaucoup de retatd) à vos messages privés, mais occupant un nouveau poste à haute responsabilité, le temps me manque.
Cela devrait cependant revenir à la normal bientot (le temps de tout mettre en place)
Ouais, ça va pour cette fois, mais que ça ne se reproduise plus...
I forgive you.
je veux aprendre l esperanto
Why can't one choose Azeri or Tajik from the list of languages. What's the procedure for including new languages?
http://tatoeba.org/fre/faq#new-language
Welcome to Tatoeba, Azeri and Tajik!
and Welcome to Tatoeba, Turkmen!
http://tatoeba.org/eng/tags/sho...with_tag/Azeri
There are 13 Azeri sentences which were added by someone of questionable background in the language. If you could check them to see if they are correct, it would be very much appreciated.
All the 13 sentences are in correct Azeri, some of them are rather formal variants of more widely used ones but still.
Thank you very much.
can i learn the writng and pronuncation of every language?
and even the writi*ng
I have some new Blog entries have a look --> http://chinesemandarinlearner.blogspot.com/
Doesn anybody know what software is used to turn Chinese sentences into pinyin? Why isn't pinyin and furigana available to download together with Chinese and Japanese sentences?
I'm working on a Chinese and Japanese online dictionary and this would be extremely useful. Thanks.
As said in one of the topic below, for the pinyin, it's a software I've made myself for the need of Tatoeba, it's independant from Tatoeba itself and so can be used separately. Though I plan to make it opensource, I still didn't find time to put it on github or any public face, and anyway the code has for the moment no documentation at all.
Romanization in general (not only for Chinese and Japanese, but also Shanghainese, Cantonese etc.) are not available to download because they are generated on the fly (I know that's not "efficient", but it's due to some legacy code)
In the future it will be possible to have it along with the sentence itself, but not yet.
Ok, I'm gonna write it myself. I forgot that's why I started working on that dictionary, to solve this kind of problems. Thanks anyway.
Great job with Tatoeba, I'm glad the Tanaka Corpus gets updated. I link to your pages within every example sentence on Tangorin. Keep up the good work.
actually if you feel like compiling it yourself, I can mail you an archive with the source of software i've coded, it would be a pity to reinvent the wheel especially as in the end i will make it opensource too, it's simply I have a really hard time these days to find free time (and as tatoeba brings me no money, I unfortunately can manage it only in my freetime)
That would be awesome. bober@tangorin.com
Can you tell me if you're script takes under consideration the changes in tones that depend on word's position in a sentence? Or is it just pulling readings from CEDICT and replacing words?
I know how it feels working on something that doesn't contribute to your income. Although I have one small ad on Tangorin it just recently started bringing enough revenue to pay for hosting.
What is the rule here for using macrons in Latin?
I think it would be alright to use diacritical marks, if you wanted to.
I'm going to. It will be beneficial to the corpus I believe.
I've never learned it like that at school in austria, but i feel that it would be okay if it's easily removable. I guess for the latin its only adding macrons on top of the normal letters, so i guess a "remove all macrons" script should be easily feasable.
I don't use.
http://tatoeba.org/eng/sentence...lexmarcelo/lat
Diacritical marks don't exist in Latin, they were introduced to ease pronunciation. I don't really recommend.
That's why I think they would be helpful here, because this site is intended for beginners.
[not needed anymore- removed by CK]
Yes you are right. But we are constructing a multi-lingual sentence dictionary, and Latin dictionaries contain these macrons.
However, I did realise a problem with bloating of the corpus that could exist.
If I add "Salvē." and someone else adds, "Salve.", we have two duplicates that won't be automatically joined.
> Yes you are right. But we are constructing a multi-lingual sentence dictionary, and Latin dictionaries contain these macrons.
Same goes for other languages like Russian, where dictionaries usually mark the accent (because it's quite important for pronunciation there), but no one writes it that way.
And this could be said about Arabic too, where at least many dictionaries indicate the vowels. Dictionaries of Serbocroatian and Baltic dictionaries also often contain the tonal accents, but no one ever writes them out in texts. I think these are comparable cases.
Also: I'd prefer to see macrons in Latin too, but that makes it a little difficult for some people to add; plus, many people who do know Latin ignore the lenghts altogether and for many words wouldn't know where they have long or short vowels if not in the "important" endings.
Yes, it would be nice if there were some automatic system like for Chinese.
Hmm... so how about Arabic? Beginners can't read without the short vowels. We would have to include them, too.
بِنْت = بنت
That wouldn't be easy...
We could do it in Latin and not in Arabic.
I think it should be optional. The reason? Many people know Latin, but cannot pronounce it, so these people wouldn't be able to participate.
I agree. Although these people should try it and they could learn how to pronounce it.
Is everyone allowed to record sentences in
- Latin?
- Esperanto?
- Toki Pona?
It would be interesting to have audio for these languages.
As for Latin, one should be careful. There are dozens of different "national" pronunciations, like the Italians pronouncing Cicero as /ˈʧiʧero/ and funny things like that. Germans say /ˈʦɪʦɐʁoː/, French people /siseˈʀo/, I think.
If we were to record extinct languages like Latin, I think we should first agree on the kind of pronunciation. Best thing by far in my opinion is the reconstructed, classical pronunciation, where the name was /ˈkɪkɛroː/ or /ˈkikeroː/ (I'd have to read up on the vowel qualities again).
I have a whole book on that topic, so I could help with questions, but I think Wikipedia is quite helpful there, too.
I agree. Classical Latin was used during the Golden Age of Latin literature (that's what I use). Although many people don't like this reconstructed pronunciation, it would be the best choice.
The problem with reconstructed pronounciation is that it differes from scholar to scholar. For example, is _sonus medius_ in l*bet, opt*mus, lacr*ma, max*mus [ɨ], [ʉ], or [y]?
It is unlikely that we will ever know this. From our sources we just know it’s _medius_...
(I tend to think it’s unlikely to be [y] since it was rarely spelled with y. But I may be biased towards [ɨ] since it exists in my mother tongue. ^^")
Well, some description of an ancient scholar about the sounds of Latin isn't the only source we have. Personal, I believe that [ɪ] might be meant, but I don't know it either.
When I was in High School, in Latin class we pronounced it /ˈkɪkɛroː/ or /ˈkikeroː/ (I'm not good at IPA), I don't know where you get that French people would pronounced it /siseˈʀo/, maybe some French who have never learnt latin?
Oh?
Well, I vaguely remembered a recording back in Encarta Encyclopedia, where "Cicero" was pronounced by a German, Italian and French person. I remember the French speaker was a woman and pronounced it like "Sisséreaut" (or so). ;)
But maybe that's not standard and maybe it only wanted to show how different people WOULD pronounce the word "Cicero" when the saw it.
Actually nowdays we write it "Cicéron", which sound more or less like you said, but I'm pretty sure in Latin class it is pronounced "correctly". But yep I get what you wanted to explain :)
For other "without native" languages, they certainly have an IPA, for them, as long as it pronounced "correctly" even with a local accent that's okay, but I think for this it would be more "judged" by other speakers
Now *that* makes sense! While repeating the *.wav file in my head, I remember it had a nasal sound at the end, which I found so weird and misplaced that I ignored it. Yeah, that file there definitely had Cicéron.
So yeah, Esperanto, Lojban, Toki Pona and also Klingon could be pronounced by anyone capable of pronouncing the words correctly. Users should also be strict and complain when someone's accent makes him pronounce different consonants (as I, for instance, sometimes pronounce Esperanto "sed" like "zet" due to my German accent, when I don't pay attention). I agree. ;)
I could pronounce Klingon quite well... but my voice isn't deep enough, haha. :D
Seriously, it would be amazing to have recordings for Klingon! I think you should record some! ;)
I'll send sysko some Latin recordings, so you tell me what you think. If you like them, I can record more.
At our Latin class we pronounced it /'tsɨtsero/, using a Mediaeval pronounciation. :P
J'ai appris à prononcer /siseˈʀo/ dans mes cours de latin en France. Je pense que c'est une question de génération...à confirmer...
I believe so. You'd have to talk to TRANG about it though.
Are there any options for differentiating between Mandarin Chinese in simplified or traditional characters? I know that all of them are currently being translated under the PRC flag, which is labelled Simplified Chinese, regardless of what character set is being used. It's not too big of an issue, but for the sake of accuracy it seems strange to have a text in traditional characters labelled as simplified.
Regarding inclusion of languages, I was wondering about the absence of Hokkien while Teochew is present in the list. I'm not trying to put it down or anything, but the two are pretty much at the same level of taxonomy, both being part of Southern Min with Hokkien having almost 30 million more speakers.
And, one final question: how are pronunciation guides for languages like Mandarin, Cantonese, and Japanese added? Are they generated automatically or must they be added manually somehow?
The fact that a Mandarin Chinese sentence is labelled "Mandarin Chinese" (the fact we use the PRC flag is an other issue), the script used doesn't change the fact it's Mandarin. If you pay attention, there's an icon precising the script used
http://flags.tatoeba.org/img/si...ed_chinese.png
or that one
http://flags.tatoeba.org/img/tr...al_chinese.png
We accept both script, because people in Taiwan or Hongkong when writting mandarin will used traditional script, books written before the simplification reform in mainland China are in traditional etc.
For Hokkien/Teochew, actually the problem is the same as with Shanghainese and Suzhouese.
We're using for the moment the international standard ISO 639-3 alpha 3 to put the "limit" between languages (this way it's easier to have an objective point of view, because as you may know, for a lot of Chinese, Cantonese is only a "dialect" of Mandarin, which is linguistically speaking not correct, and it's the case with a lot of languages around the world). According to it, Teochew and Hokkien are both under the code "nan" for Min Nan language http://www.ethnologue.com/show_...e.asp?code=nan , so yep actually it would have been more accurate to name it "Min Nan" in Tatoeba, we were not aware of that when someone proposed us to add Teochew around a year and a half ago. But for the moment it's the only "dialect" of Min Nan we have, a user proposed us to start to Hokkien some months ago, but I was quite busy, and we were talking about this problem, and somehow we lost contact.
In a more general in tatoeba we add a language only when we have somebody with a sufficient level (either a native or someone with pretty strong knowledge in that language, to avoid the "I find this language cool, and discover 5 sentences on a website"). Because as we have no limit on the number of languages we can support (the goal is to have them all), but as there's more than 6000 languages, it would really clutter the interface if we were to add them all know, we prefer to add them when we start to have sentences in it. So for Hokkien, if you're a native speaker or know some, or have strong knowledge in it, we can add it, but it would be a bit special as we have already teotchew with the same ISO code, so maybe I can just rename it "Min Nan" and sentences in Teochew dialect of Min Nan will be tagged so, same for Hokkien, soon a new ISO standard will be released with this time a code for each dialect of each languages, at that time we will be able to make something smarter to be able to not only tell the language but also the dialect it's part of.
For the pronunciation, they are automatically added and for the moment they can't be edited (but in the future it will be possible to do so)
Thanks for the clarification. I wouldn't object to having Min Nan as a language with Teochew and Hokkien included underneath, but I guess it's not my decision to make. I am a native speaker, but I'm only partially proficient in the language so there's no need to make any special moves for me. At the moment, however, it seems as though there are only 6 sentences in Teochew, so... I mean obviously it's still up to you.
How different are Teochew and Hokkien? Are they mutally intellegible?
They're similar enough to both be classified as part of Southern Min, but they are not mutually intelligible.
I'm not sure if it has already been discussed, but what is the community's opinion on being able to create tags more freely? I mean, for many languages, if we search for a verb, sometimes we are not able to find any matching sentences due to verb conjugation. For example: imagine I'm confused about the use of te expression "to see red" and I write "see red" on the search bar, but the only sentence avaiable is "she SAW red". Then I won't be able to find it. Especially if I don't know that the past of "see" is "saw". Wouldn't it be useful to be able to tag the infinitive form of verbs into our sentences? What does everybody think about it?
actually it's more about improving the search engine capacity to recognize infinitive form of verbs, because right now an advanced user can create any tags (though normal users can't) without any restriction, rather than improving tagging capacity, otherwise we will would finish with hundrends of thousand tags (number of languages we have * number of verbs in each language) which that will uselessly clutter the tags list.
Hm, I get it know. Well, it's just that I was actually thinking of a whole new tagging system, different from the one we use today. Instead of having fixed tags that we select from a list, it would work pretty much like a website search engine. Whenever we create a website, we can create tags for them which are included in the search of any search engine. It would pretty much work like that. So, imagine I've got a sentence, like:
> He woke me up
I could include the tages: "wake up; woke up; wake someone up", for example. Whenever the search engine was used, it would look up for the searched term on the sentences as well as on the hidden tags created by the users at the moment of adding up that sentence.
Of course there would have to be a huge effort into changing the search engine. I just though it would be more helpful.
Certainly, it's up to the community, of course. I'm just trying to learn everyone's thoughts on the matter. :)
Cela impliquerait de faire confiance à tous les utilisateurs (y compris les petits plaisantins et les saboteurs) dans leur capacité à étiqueter leurs phrases de manière pertinente...
Ou alors il faudrait, comme pour le contenu des phrases, pouvoir débattre des étiquettes personnelles, ce qui créerait un système parallèle de discussion...
Personnellement, je ne crois pas à la sagesse des foules. Je crois même plutôt à la bêtise crasse des foules, que l'Histoire des hommes nous enseigne abondamment. Les foules ont inventé le Stalinisme, le Nazisme, Le Maoïsme, Les Khmers Rouges, Le génocide arménien, au Rwanda, et plus récemment on a pu voir la foule se marcher dessus lors d'une simple parade en Allemagne (comme ça arrive fréquemment dans tous les stades et lieux publics...)
Coluche, un humoriste français, disait qu'à plus de 2, on est une bande de cons. Je ne suis pas loin de partager cet avis.
Je me suis déjà souvent exprimé sur ce sujet mais, au risque de me répéter, je vais le faire encore une fois :
De même que la mauvaise monnaie chasse la bonne, la mauvaise information chasse la bonne, et c'est vrai aussi des phrases et de leurs traductions.
Donc si on passe assez de temps avec suffisamment de personnes, sachant que les personnes éclairées ou expertes sont une minorité, elles seront toujours débordées par l'ignorance, l'incompétence et la bêtise.
À terme, je prédis même que le nombre de phrases et de traductions erronées sur Tatoeba dépassera le nombre de phrases et de traductions justes. C'est peut-être déjà le cas...
Mais, au moins, grâce aux débats (parfois houleux) la plupart des phrases erronées arrivent plus ou moins, tôt ou tard (plutôt tard) à être corrigées.
Pour les étiquettes personnelles, je serais beaucoup plus dubitatif. Les gens s'y accrocheraient à tout prix et l'indexation deviendrait complètement foireuse.
Tous les sites d'étiquettes « libres » que je connais terminent tous en un ramassis d'étiquettes mal rédigées, pleines de fautes d'orthographes, redondantes, non administrées et au final inexploitables, surtout en plusieurs langues...
Un exemple pour vous en convaincre : http://fr.forvo.com/tags/
Vous pouvez constater que même la liste actuelle des étiquettes Tatoeba, pourtant réservée aux contributeurs confirmés, et plusieurs fois expurgée, est déjà assez bordélique http://tatoeba.org/fre/tags/view_all et la raison en est que, à part pour les étiquettes utilisées pour l'administration des phrases et de leurs corrections, personne ne s'accorde sur ce que devrait être les étiquettes et à quoi elles devraient servir (il n'y a d'ailleurs pas de débat sur ce sujet, car chacun considère le statu quo selon sa propre interprétation.)
I think that the tag system works pretty well, but tags should be classified. The corpus quality is good in my opinion, though moderators should probably be a bit more active in applying corrections (in French at least).
Here are some of the links for moderators:
http://tatoeba.org/eng/tags/for...e_spelling/eng
http://tatoeba.org/fre/tags/for...rs/@change/eng
http://tatoeba.org/eng/tags/for...ge_grammar/eng
Overall, I agree with sacredceltic that opening the tag system to everyone or having "personal tags" would be a mess.
Il y a actuellement des étiquettes parfaitement inutiles, parce qu'elles n'apportent pas d'information générale aux contributeurs, mais sont utilisées comme codes privés par certains, et qui devraient plutôt faire l'objet de listes...
Le principe des étiquettes, c'est qu'elles sont visibles par tous, et donc leur lecture doit être utile à tous, autrement l'espace public est encombré et ça introduit de la confusion, d'autant qu'elles sont dans une seule langue que tout le monde ne comprend déjà pas forcément...
Toute étiquette qui n'est pas destinée à informer tout le monde et qui ne le fait pas suffisamment clairement (abréviations, mots intraduisibles ou difficilement compréhensibles) devrait donc être retirée et son usage actuel devrait être traité autrement.
oui dans le nouveau système de tag (j'ai plus en tete s'il est deja utilisable sur tato.sysko.fr(en tout cas je l'ai codé)), ne pourront être ajouté que des tags déjà existants, (il sera bien évidemment possible de me demander (à terme il y a aura un formulaire pour cela, pour automatiser la chose de mon coté))
cela évitera les tags inutiles, les tags redondants, permettra de plus facilement les traduire etc.
beaucoup d'autres (dont ce que proposent l'utilisateur au dessus) passeront sur des champs "métas" qui fonctionneront sur un principe de clé valeur comme suit
clé valeur
auteur Victor Hugo
forme "brut" je avoir mangé un pomme
grammaire S V C
(ce ne sont que des exemples, pas la peine de discuter de leur contenu propre )
L'essentiel, c'est que ça n'encombre pas l'espace visuel. L'intérêt d'une étiquette, c'est bien sûr d'attirer l'attention sur quelque chose d'important (par exemple le fait qu'il s'agit d'une citation ou d'un proverbe, ou d'une rime, et qu'il faut donc en tenir compte pour traduire).
Mais trop d'information tue l'information. S'il y a donc trop d'étiquettes, l'œil n'est plus attiré par l'essentiel et ça entraîne de la confusion chez les nouveaux arrivants qui ne comprennent pas très bien à quoi ça sert...
Does anyone know if there is a site like Tatoeba but for words rather than sentences? I've seen wiktionary, but it really is not as user friendly as Tatoeba's interface and so I've had trouble getting people involved. It would be great if Tatoeba could have a sister site for word to word (dictionary type) translations as well - especially because the data can be so easily accessed through downloading .csv files.
If you mean, something which the same as tatoeba on
1 - multi-language with not a single "source" language
2 - possibly to interlink, in order to permit to people to easily create a A to C dictionnary if there's already a A to B and B to C dictionnary by validating translation
3 - Free as in free speech (bab.la is not no?), which permit to reuse it in other project and not keep it captive of one website
4 - code source of the plateform also free to permit contributors to either help to improve it, or to "fork" it, if one day the admin turns crazy
No there's no such a plateform yet (actually even only point 1 2 and 3), the much closer, if you think the "open culture" aspect is important is the wiktionnary, with all the drawbacks we know it have
So actually last year while I was trying to find back a code architecture for the new version of tatoeba, I've made a test on creating a project like you described http://redmine.sysko.fr/project...ict/repository (don't pay attention to the name "shanghainesedict", it's because at first I only wanted to make a shanghainedictionnary website, and soon it turns out that it was actually able to host the same features as tatoeba for a dictionnary)
Of course it will not be a simple "copy paste" of tatoeba code with only a change in the content, but for the moment I'm focusing on the new version of tatoeba itself, so this new project is not going to appear before some months.
Totally understand something like that would take some time. I do think it would be a great resource. I will take a look at the code and see if I can get it to work on my server - I'm not a programmer, but perhaps I can give some feedback that would be helpful.
I work with English and Khmer, so that is my main motivation (I have a mediawiki site as a dictionary right now, but it isn't user friendly and it is limited to English and Khmer, rather than any language http://dictionary.sbbic.org) - there aren't many resources out there in Khmer, so I have been looking around for a long time for something that could take a large dictionary but make it open (as in anyone can download it and use it for whatever they want) as well as collaborative and user friendly (unlike wiktionary). Your site is as close as I have found anywhere.
So thanks for your hard work - it is great to see a site like this thriving.
-Nathan
>Of course it will not be a simple "copy paste" of tatoeba
>code with only a change in the content
It won't? The Tatoeba software is Open Source and available for download, isn't it? You just need to install an instance of it and put in words instead of whole sentences. Or am I missing something?
I like using http://bab.la
Even though it's a normal dictionnary (that focuses on words rather then sentences), it shows them in several different contextes so you can choose the best option for you.
Cool, never seen bab.la, but it doesn't look opensource (meaning the content cannot be downloaded, unless I am mistaken). But it does look like a good resource.
Actually I just found this: https://www.assembla.com/spaces...Wr3Bj7ab7jnrAJ
Is this the source of the backend of Tatoeba? If we hosted a dictionary site using the Tatoeba backend, would that be an acceptable thing (unless of course there is already one that exists)?
Thanks,
Nathan