Nostre datos es liberate sub licentias de Creative Commons varie.Plus information
Si tu ama iste contento, per favor considera un donation.
Wall (1 discussion)
Consilios
Ante de poner un question, assecura te de haber legite le FAQ.
Nostre intention es mantener un atmosphere salubre pro discussiones civilisate. Per favor lege nostre regulas contra mal conducta.
frpzzd
un minuta retro
doemaar14
heri
gillux
heri
sharptoothed
heri
Babelball
heri
TATAR1
heri
LeviHighway
heri
AlanF_US
heri
LeviHighway
heri
LeviHighway
heri
-- Raw release of the transliteration tools I've coded for Sinitic languages --
I've put a quick and dirty archive here http://downloads.tatoeba.org/sinoparserd.tar.gz
It works well on the server, though it has some known limitation http://redmine.sysko.fr/projects/sinoparser/issues
I release the code under GPL V3 ( as I've been the only to coded it yet, if someone really wants me to release it under an other license, BSD?, such tell me and explain why)
The data/dictionnary files are under a cc-by-sa like license, but I need to check
Enjoy
Thank you very much! ^^
I'm sorry if I make far too many questions these days, I just want to find out why there are 8463 sentences spanning out from sentence Nº3, most of them apparently senselessly.
After pointing out a few irregularities, this net's size turned from 8463 to 5361 sentences to this week.
参ったな. ^^
See...
a) Follow only direct links. Indirect links are a known source of trouble. :) Use them ONLY if you don't have a direct link, and use them with care.
b) There are legends that in the past, Tatoeba wasn't a graph, but a table. If a sentence was added, it was added as a translation to all the sentences, and only one translation in a given language can exist. So sentences with low numbers can turn out to be ancient remnants of that ancient legendary antidiluvial Tatoeba; and they may be linked not really well.
That legend could indeed be very true!
Perhaps that explains why I sometimes find sentences linked with no history whatsoever of it in the log : o
Anyways, it's not really wrong that we find hundreds of sentences in the same subgraph, but it's just inconvenient sometimes in terms of computation when you work with Tatoeba as a database or graph. That's why I'm not going to insist much with the details, I'll just make some suggestions.
I think that a few corrections here and there might indeed dramatically simplify the problem.
[АНГ] PLEASE, everyone, NEVER translate something using automatic translators!
You think they work OK? They may be OK for some languages and EXTREMEMLY bad for others!
Neither on Tatoeba, nor on Lauchpad! ESPECIALLY at Launchpad!
I've just seen the Belarusian translations Guybrush88 has added to Lauchpad:
https://translations.launchpad....erson=guybrush
I don't understand WHY HE DOES IT? "Прысуды у" may be correct translation for "Sentences in" in a court, but not on Tatoeba!
I don't understand WHY DO PEOPLE DO IT? You want to see Tatoeba translated into Belarusin so much? Then please bug those who can do it, instead of hindering them!
P.S. Using Lauchpad's suggestions is also NOT a good idea. Context is important. What is more, there are several standards for Belarusian, and the translation must be CONSISTENT.
[БЕЛ] КАЛІ ЛАСКА, прашу вас, НІКОЛІ не перакладайце нічога аўтаматычнымі перакладчыкамі!
Кам здаецца, што яны добра перакладаюць? Магчыма, на нейкія мовы і добра — але на нейкія ЖУДАСНА дрэнна!
Ні на Татоэбе, ні на Lauchpad'зе! АСАБЛІВА на Launchpad'зе!
Я толькі што паглядзеў, якія беларускія пераклады Guybrush88 дадаў на Lauchpad:
https://translations.launchpad....erson=guybrush
Я не разумею, НАВОШТА ГЭТА РАБІЦЬ? "Прысуды у" можа быць правільным перакладам «Sentences in» у судзе, але не на Татоэбе!
Я не разумею, НАВОШТА ЛЮДЗІ ТАКІМ ЗАЙМАЮЦЦА? Вам так хочацца пабачыць Татоэбу па-беларуску? Тады, калі ласка, папрасіце тых, хто можа перакласці, а не замінайце ім!
P.S. Выкарыстоўваць прапановы перакладаў на Lauchpad'зе — таксама дрэнная ідэя. Кантэкст важлівы. Акрамя таго, для беларускай існуе некалькі стандартаў, і пераклад павінен быць АДНАСТАЙНЫ.
>They may be OK for some languages
Which ones, I'm curious? ...not any that I know so far, in any case!
I don't want to speak for Demetrius, but I think he added that more than to avoid the "but it works with language X" argument, rather than in a strong belief that machine translation give good results.
> Which ones, I'm curious? ...not any that
> I know so far, in any case!
Well, not that you've said that, I can't think of any case except Chinese Traditional into Simplified...
Why do you allow exact duplicate sentences?
In the current version, there is no inline-deduplication. This will come in the future.
Meanwhile, a batch deduplication procedure that merges duplicates and their attributes (comments, tags, translations) is run from time to time (I wish it was more often).
It is not such a problem, in the end.
*** listoj por lingvoj kiu jam estas kreita ***
[epo] La listoj: Asturian language, Greek, Ido, Interlingue, Klingon, Malayalam, Siciliano, Thai, Turkish, lojban...ne plu utilas ĉar tiuj lingvoj estis kreitaj delonge. Bonvolu forigi ilin, ĉar la listo de listoj jam estas tre longa.
[fra] Les listes : Asturian language, Greek, Ido, Interlingue, Klingon, Malayalam, Siciliano, Thai, Turkish, lojban...ne sont plus utiles puisque ces langues ont été créées depuis longtemps. Veuillez les supprimer car la liste des listes est déjà fort longue.
[eng] The lists: Asturian language, Greek, Ido, Interlingue, Klingon, Malayalam, Siciliano, Thai, Turkish, lojban...are no longer necessary since these languages have been created a long time ago. Please delete them, as the list of lists is already very long.
why deleted? but Turkish shouldn't be deleted
Before replying to a post, you should first learn at least one of the languages it is written in, in order to understand it properly.
because I'm native Turkish.
now I understand! just the lists, not languages ;)
if turkish should be deleted, georgian and estonian are candidate to delete.
Mi volus aldoni la saami language(s)ISO 639-2 and 639-3
sed intertempe, vi povas jam aldoni viajn frazojn kun la "ne-konata" flageto kaj fari "Saami" liston en kiu vi aldonas tiujn frazojn.
Bonvenon al "saami" ĉe Tatoeba ! (kiu estas la nomo el la saami en esperanto?)
.-.-.
Dankon, sed mi ankorau ne scias kiel fari por aldoni la samean tekston .. Sámegielat leat Eurohpá eamigielat ja nuortamearasuoma lagamus fuolkkit.
Vi povas rekte aldoni frazon ĉi tie http://tatoeba.org/epo/sentences/add (bonvolu elekti "Alia lingvo" en la lingvolisto.
Aŭ vi povas, ekde iu frazo ke vi volas traduki, premi la tradukbotono super la frazo kiun vi volas traduki (la plej malrekta botono kun japana signo).
Traduku nur frazojn en lingvoj kiujn vi bone komprenas.
La frazo en la lingvo kiun vi tradukas devas ĉiam esti super la aliaj. Se alia lingvo estas super, vi devas antaŭe alklaki la frazo en la lingvo kiun vi volas traduki, do ĝi moviĝas super la aliaj.
Kiam vi havas aldonita vian frazon, vi povas alklaki ĝin (bonvolu alklaki la verdan sageto en la malrekta) kaj elekti liston por vian frazon kun la botono en la rekta, super la frazo.
Por krei liston, vidu ĉi tie en la rekta http://tatoeba.org/epo/sentences_lists/index ("Krei novan liston")
Vi povas rekte aldoni frazon ĉi tie http://tatoeba.org/epo/sentences/add (bonvolu elekti "Alia lingvo" en la lingvolisto.
Sed en la lingvolisto ne estas "Alia lingvo" ! (en aliaj lingvolistoj estas ja plej supre, sed en tiu listo ne estas, ghi komencas kun "Afrikansa")
Elektu iun ajn kaj vi povas sxangxi gxin poste se vi alklaki sur la flageto...
Vi devos atendi, ĉar la administristo, kiu aldonas novajn lingvojn, forestas pro personaj kialoj.
The following sentences are identical apart from the kind of apostrophe used:
http://tatoeba.org/eng/sentences/show/3447
http://tatoeba.org/eng/sentences/show/373830
Would it be possible for the script to merge those kinds of duplicates at some point?
[not needed anymore- removed by CK]
-1
straight quotes look awful in most languages! But +1 to the duplicates script ignoring them. For the apostrophe, there is a *right* one in french.
Actually there are even two rigth ones http://fr.wikipedia.org/wiki/Ap...typographie%29
Nevertheless wikipedia says that "Unicode recommande l’utilisation du guillemet-apostrophe (U+2019) pour représenter l’apostrophe" (Unicode recommends the usage of the guillemet-apostrophe (U+2019) for representing the apostrophe). I think we can stick to this.
La seule apostrophe acceptable en français est en effet la U+2019, malheureusement, les claviers de la plupart des ordinateurs ne la proposent pas en standard.
Je trouve que Tatoeba devrait automatiquement transformer toutes les autres apostrophes en U+2019 à la saisie...
Oui, mais je suis contre les automatismes qui changent les phrases. Il fallait que la page propose des changements qu'on peut facilement appliquer. Peut-être déjà quand on compose la phrase. Peut-être je peux améliorer mon userscript (Tatoeba Symbol insert helper) pour faire ca.
ça serait une bonne idée !
+1
same for
sentence Nº 403068
sentence Nº 688010
or
sentence Nº 14970
sentence Nº 745948
: )
Where is greece language?
Greek?
You'll find it under "Modern Greek".
Mi trovis ion por korekti, sed mi ne scias, kiel. En la ĉeĥa frazo Jak se maš devas esti Jak se máš. Super "a" devas esti streketo, kiu indikas longan vokalon, do longan silabon.
je ne vois pas dans la liste des langues le sango, langue nationale de la république centrafricaine.
.-.-.
Il est a noter que TRANG étant absente pour des raisons personnelles, pour l'instant l'ajout de nouvelles langues est gelée, cela pourra donc prendre quelques semaines, voir mois, avant d'apparaitre.
Désolé pour la gêne
Ah, les excès du taylorisme ! Pourquoi ne peux-tu pas en ajouter toi-même ? C'est si compliqué ?
C'est vraiment dommage tout de même, parce qu'il y a plusieurs langues, comme le sorbe, qui attendent depuis longtemps...
Si je peux aider à quelque chose...
car oui la procédure est loin d'être simple, même si je suis entièrement d'accord que cela devrait l'être, jamais eu le temps de refondre cette partie là
Et à présent le temps que je consacre à tatoeba se résume à 30 minutes à une heure vers 4h du matin, seul moment de la journée ou j'ai du temps libre. Donc non désolé comme dit précédemment ces derniers mois je fais le maximum que je peux faire chaque jour. Je dors environ 2h par nuit, donc passer 30 minutes à essayer de retrouver la procédure exacte pour ajouter une langue n'est pas envisageable, surtout quand au même moment il y a les demandes diverses des utilisateurs, la puérilité de bora etc.
et malheureusement l'ajout de langue demande plusieurs manipulations directement sur la base de données et le moteur de recherche, donc à part en ayant un accès ssh au serveur et en connaissant comment on tripatouille sphinx ...
Je suis en vacances à la fin de la semaine pour 2 semaines, j'essaierai de boucler le maximum de chose pendant ces 14 jours.
N'oublie pas de dormir !
>le sorbe attendent depuis longtemps...
Oui, c'est cela, il est grand temps.
.-.-.