Wall (6,960 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.
morbrorper
5 hours ago
marafon
5 days ago
CK
5 days ago
sharptoothed
11 days ago
Cangarejo
11 days ago
Cangarejo
14 days ago
Thanuir
14 days ago
ondo
15 days ago
ddnktr
15 days ago
ondo
15 days ago
Duplicate sentences
I'm asking everyone to stop posting comments about duplicates and to stop deleting them.
I will set up a page where you can enter 2 ids and it will merge the sentences if they are "mergeable". So please be patient, stop posting comments to say it's a duplicate, stop deleting duplicates.
I know it's frustrating for many of you to know that there are so many duplicate sentences in Tatoeba but trying to delete duplicates manually will not solve the problem.
We do have a script that would take care of deleting duplicates, except it has some small bug that sysko did not have time to figure out, so we stopped running it.
To give you an idea, here's the number of duplicates for the 5 top languages in Tatoeba, on 20/03/2011 (that is today), at 1AM, France time.
eng => 676
jpn => 29
epo => 3057
fra => 653
deu => 1135
The main cause: Tatoeba cannot display any further than indirect translations. So what happens is that:
1) Someone adds a sentence in French.
=> FRA
2) Someone translates FRA into English.
=> FRA-ENG
3) Someone translates ENG into German.
=> FRA-ENG-DEU
4) Someone translates DEU into Esperanto.
=> FRA-ENG-DEU-EPO1
5) Someone translates FRA into Esperanto. They couldn't see EPO1 because it's too far away from FRA and is not displayed for performance issues. It would be very slow if we tried to display all the translation chain so we stop at indirect translations. And the result is that we have such chains:
=> EPO2-FRA-ENG-DEU-EPO1 (where EPO2 = EPO1)
It also happens that a user adds a "new" sentence that already exists, but this case is not as common.
Things should get better in the next version of Tatoeba, that some of you may have heard about. The next version will be able to display the whole translation chain without taking forever. It will be possible to do things like "get French sentences that do not have Esperanto translations at all, not even in-in-indirect ones".
Will we add creoles as their own distinct languages or as specially tagged sentences within their own standard languages?
ISO 639-3 rules ^^
(so, surely as distinct languages if they have ISO codes, and I think that we will analyse case per case for languages that don't have ISO codes)
[eng] I've uploaded the audio of BraveSentry, 1300 German sentences
[fra] Je viens de mettre les audios de BraveSentry, 1300 phrases en allemand
[esp] Sysko ha subido los audios de BraveSentry, 1300 frases en alemán
#266486, "She doesn't drink enough breast milk," has just been untagged as "not for safe search." This raises a question: might we need a strict safe search, or shall we assume our readership is familiar and comfortable with the prime objective of boobs?
#notforsafesearch
A couple more things to say about future sign language support:
1. Livemocha uses a Flash interface to record audio submissions. Would a Flash interface to record video submissions violate a free software license or anything? If so, would prospective contributors submit video clips to Sysko the way we submit audio files to him now?
2. I don't know about other sign languages, but American Sign Language sometimes has much shorter sentences than equivalents in other languages. People concerned with the completeness of sentences on Tatoeba will want to keep this in mind.
3. We'll probably want to write a tutorial for video submissions once we beta-test and perfect the system.
...
Je pense qu'elles sont en ordre chronologique d'ajout ^^
** Tatoeba Linker User Script **
In order to make it easier for trusted users and moderators to link sentences which do not appear in the indirect translations, I have created this user script:
http://userscripts.org/scripts/show/99149
It puts a box below each sentence that allows you to enter the ID of another sentence to link.
It should work in Chrome and Firefox (at least with my minimal testing). For Firefox you'll need to install GreaseMonkey:
https://addons.mozilla.org/en-U.../greasemonkey/
When you install my script, you'll probably be asked if you trust me. To use it, you'll have to. :P
The source code is pretty simple but relies pretty heavily on the current structure of the Tatoeba HTML code. So I'll try to update it if Tatoeba has a big update that breaks it.
This now makes my old bookmarklet completely unnecessary:
http://tatoeba.org/wall/show_message/4507
unless you're one of those pesky IE users. ;-)
Hey Zifre! Very nice script!
If you could add
// @include http://tatoeba.org/*
as @match does not work for greasemonkey. Greasmonkey then makes the script "work" on all pages. It seems to be okay to have both @include and @match. I will add @match to my script too ;)
My school's stupid filter prevents me from getting to the script. :-(
I'll try to fix it when I get home, but I might not have time until tomorrow.
Okay, I've added that. Sorry that it took so long.
Also, I now recommend that this only be used in Firefox, not Chrome. For reasons I haven't quite figured out, it interferes with the normal linking and unlinking buttons in Chrome.
Hello all Tatoeba Users!
I've recently posted a feature request and have now temporarily "solved" it using a so called USERSCRIPT:
If you are using FIREFOX with GREASEMONKEY-addon you can install it right away from this url:
http://home.pages.at/k8ag/tatoe...nguage.user.js
(Some other Browsers such as Opera and Safari also can handle Userscripts. Please search the web (with duckduckgo.com) for further information on "userscript" and your browsers name)
I designed my script to enhance the RANDOM SENTENCE function of Tatoeba. What it does:
*************
It randomly selects a language for previous, next or random sentence out of a predifined list.
*************
So let's say you speak English, German and Esperanto as I do, then you make a little list like this ['eng', 'deu', 'epo'] and the random function will give you only sentences in these languages. For further information see inside the script or feel free to ask here.
I hope you enjoy it :)
Jakov
hey....i installed this, but don´t know where exactly to put the list of languages i want.
also: does this work for "show several random sentences" too?
It should be line 32 to 37 in the script code. It should work on the other pages too if you add after this line
// @include http://tatoeba.org/*/sentences/show/*
these lines:
// @include http://tatoeba.org/*/sentences/...ndom_sentences
// @include http://tatoeba.org/*/activities...late_sentences
// @include http://tatoeba.org/*/activities/adopt_sentences/
// @include //http://tatoeba.org/*/sentences/show_all_in/
and after this line $('#randomLangChoiceInBrowse').val(usethisone).change();
these:
$('#SentenceInto').val(usethisone).change();
$('#filterLanguageSelect').val(usethisone).change();
There are some bugs i cant handle until now! For instance you cannot link sentences anymore when useing the script, so i dont really recomend it for now :(
I will try to get it working soon!
We've hit 4000 members!
¡Ahora hay 4000 miembros!
今は、 メンバーが 4000人いますよ!
Quand la prochaine mise à jour de l'interface aura-t-elle lieu ?
À l'attention des francophones, j'ai enfin traduit dans Launchpad pratiquement tout ce qui restait encore en anglais dans les menus, etc...pour autant que ça soit traductible...
je repose la question, car j'ai effectué toutes les traductions qui manquaient en français, mais c'est toujours en anglais...
Oui désolé pour le délai, en fait il faut savoir que cela de notre coté se fait en trois étapes,
on génére d'abord le fichier contenant les parties du sites à traduire
on mets le tout sur le site de launchpad
Les gens en suite traduisent
Launchpad tout les un ou deux jours génère le fichier de traduction que l'on peut ensuite remettre sur le serveur.
Donc en fait souvent le décalage qu'il peut y avoir vient du délai de launchpad, après c'est un moindre mal en comparaison des avantages qu'offre la plateforme.
Si certaines parties restes malgré tout en anglais il se peut parfois que nous ayons oublié de marquer certains passages comme étant à traduire. Où parfois Trang choisit délibérément de ne pas rendre certains passages traduisibles car ils ne sont pas finalisés etc.
Il reste encore toute la partie FAQ à mettre dans launchpad
> Quand la prochaine mise à jour de l'interface aura-t-elle
> lieu ?
Weekend prochain, ou avant si j'ai le temps.