Wall (6,006 threads)
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.
2 hours ago
7 hours ago
13 hours ago
14 hours ago
2 days ago
4 days ago
4 days ago
4 days ago
4 days ago
Has the possibility of getting rid of language flags ever been considered, and if yes, what was the final decision (if there was one)? I believe the ISO codes will do a better job.
Obviously, inspired by the Impersonator's post about the Belarusian flag, just didn't want to rain on his parade mentioning another possibility there, as these are two separate topics.
This idea is brought up from time to time, and it's usually dismissed. Here're what Trang says:
> «Removing flags doesn't solve everything.
> It only replaces one problem with another.
> Without flags, you'll have, for sure, people
> telling you they have a harder time spotting
> sentences in their desired language and
> protesting to get the flags back. Flags are
> one of the core elements of Tatoeba's UI.
> You cannot just remove them and expect that nobody will care.
> It is a design choice and yes, it costs us, once
> in a while, very intense and even violent
> discussions. But it's a price we have agreed
> to pay in exchange of the possibility to give our
> languages a visual identity. You don't have
> to agree with this choice»
> I believe the ISO codes will do a better job.
I personally don't like ISO codes because they're in Latin script, which is not really neutral.
Thanks. So, basically, proposed and dismissed.
This is such an amazing open-source find. I am not a programmer but am very detail-oriented and would love to contribute however I can.
What kind of partnerships do you have? Have you ever approached DuoLingo, for example, to see if they'd be willing to share their sentences? They have some hilarious ones. My favorite du jour is Wakker worden, schappenmensen! (Dutch for Wake up, sheep people!)
Duolingon lauseet käännöksineen lienevät arvokas osa firmaa. Tuskin he ovat halukkaita sitä jakamaan. Muutenkin hyvin suljettu ympäristö. Jos haluaa ehdottaa yhteistyötä, kannatanee mieluummin lähestyä jotakin avoimella filosofialla toimivaa epäkaupallista toimijaa.
I agree with Thanuir. And even if Duolingo agreed to share their sentences with us, we might not be able to accept them all at once, since there's currently no established process for importing large numbers of sentences while also ensuring quality control.
On the other hand, if someone wants to use sentences from Tatoeba in their own project, there's no need to formally establish a partnership, since the Creative Commons license used by Tatoeba already allows them to do that without asking for permission.
How can I delete a wrong translation?
You can edit it to correct it. If it seems irreparable, you can leave a comment "delete please".
Hvis setningen er riktig men er ikke en god oversettelse til den den er lenka til, det er bedre å la setningen bli men fjerne lenka (eller be om noen å fjerne lenka).
I am happy to have found this site -- I tried a couple time to make something like this but don't have the technical skill. I have a few comments/questions. I suspect all have come up before, but didn't see answers in FAQ.
* is it possible to search the Wall?
* When looking at a sentence, can one see all the tags that have been applied to it? E.g. #3451959, [have you ever noticed how many near-anagrams 6 first digits in pi come up in daily life or is it just me?] I think I got there by following "Compound sentence" tag. Now, I'd like to know if it has been tagged by a French tag similar to "Compound Sentence", but I don't know/can't guess how a French speaker would tag it -- can I see all the tags applied to that sentence?
* I'm generally interested in longer sentences -- it it seems like an obvious feature to add to the "Advance Search" page (At least/At most X words). Is it not feasible?
* It seems to me like set-up with Shtooka tools is a hindrance, and the demo of how to get crystal-clear recorinds is a bit intimidating. I don't even own a microphone! So, I don't know if your goal is to have all recordings very clear -- I'd see a value in a variety of situations, like talking with other voices in the background, talking with a mouthful of food, talking with uhms, pauses, excitedly. (My interest is as a language learner, not an archivist). Have you hashed over this topic and decided a goal of Tatoeba is to have recording uniformly high quality? And would you be interested in easier to use audio-gathering tool at the expense of lower sound quality? (and probably also at the expense of attributing to a specific tatoeba user without a fair amount of programming.)
* do you solicit/accept "help pay for hosting" money contributions?
> I don't even own a microphone!
I'm using my webcam to record sentences for now.
Donations - https://tatoeba.org/eng/donate
* is it possible to search the Wall?
Not really, However, if you download the weekly export of Wall posts, you can search them offline if you know how to.
* When looking at a sentence, can one see all the tags that have been applied to it?
All tags on a given sentence are shown on the right side of the sentence's page.
* I'm generally interested in longer sentences
You can use the advanced search to find the longest sentences, though it's not possible to easily find sentences with "At least/At most X words."
Here is a pre-filled in advanced search form set up to find the French sentences with the most words that also have English translations. Just go here and input a French word or phrase, or skip inputting a French word and click the "Search" button to find the longest French sentences.
Thank you * 3
> would you be interested in easier to use audio-gathering tool at the expense of lower sound quality? (and probably also at the expense of attributing to a specific tatoeba user without a fair amount of programming.)
gillux made a proof-of-concept tool for recording audio a while ago when the question of using mobile devices for recording came up: https://tatoeba.org/eng/wall/sh...#message_33176 Was something like that what you had in mind?
What I envisioned was an independent web page, not so much to enable mobile as to get rid of the need for installing shtooka, naming files, emailing, etc. I was thinking of attracting sentences from casual contributors, including cell phone users without a PC.
Using the Web Audio API as gillux mentioned, I'd picture a simple site aping shtooka recorder. User would select the language to read in, and the web site would pick a batch of sentences that don't have audio in that language.
I haven't ever written a usable web site, so I don't really know what's involved technically. And a site that attained popularity might start to incur expenses -- but I guess that would be a good problem to have.
This is a project that does somethings like what you are suggesting.
You can listen to some of their sentences to get an idea of what the sound quality of the files are like. You can also try recording a few sentences.
Intriguing. I wasn't aware of that project.
The recording UI is something like what I had in mind for Tatoeba to gather sentences. I like the 5 sentence at a time unit.
For a hypothetical flow -- a similar site could gather 5 sentences, using a stock of tatoeba sentences without recordings, create a zip of between 1 and 5 mp3s (depending how many sentences a user skipped), and generate a URL for the user to give you to incorporate. If users have to log in to tatoeba to give you the URLs, maybe it mitigates the possibility of flooding your in queue with junk.
** Stats & Graphs **
Tatoeba Stats, Graphs & Charts have been updated:
You're welcome :-)
Do we have a map of Tatoeba-Wiki? Just to know, what do we have and what do we want to describe? And where to put or find it.
Maybe it's more or less the FAQ-main-page (when the links to the translations will work).
I would say the English main page is the map you're looking for: https://en.wiki.tatoeba.org/
Unless you are looking for a page that lists *all* the pages that exist in the wiki?
It's because I'm not sure how to going on by translating the wiki. Because I know we have for instance the FAQ translated on Transifex but not transferred(or linked) to the wiki. Are there other somewhere hidden websites? Or would it be a good idea to edit/use the empty german mainpage, at least to collect there the links to the sites already translated to german?
We have a new German voice.
Mramosch has contributed 4,390 audio files.
On August 14, this was up to 6,586 audio files.
link >"Contribute" >"Add vocabulary request" goes instead to >"Vocabulary that needs sentences"
Short Enhancement request:
The search "^Tom * * * Mary$" searches sentences with 5 words - 1 word per star.
However: "^Tom * * *" or "* * *" does not search for 1 word per star.
Make it possible to search for sentences with word counts through star numbers
I did not check if this would interfere with other search options.
Rather than give a general reason like "consistency", could you please give a broader picture as to what you're trying to achieve, and why the current behavior doesn't provide it for you? By "what you're trying to achieve", I don't just mean "I'm trying to find sentences that start with 'Tom' and contain exactly five words, without my having to specify the last one". I mean why, in the bigger picture, do you want to find sentences that start with 'Tom' and contain exactly five words, and given this reason, why is it a problem to specify the last word?
Note also that the basic syntax of searches is determined by our search engine (Manticore), not us. There are occasions when we do some preprocessing on our own side (for instance, when the search query ends with a question mark that is intended as punctuation, not a wildcard), but those are very rare.
I personally would use it to search sentences with a specific number of words. So I personally would use it as "* * * * *" to find a three-word sentence.
Why? Because in many cases searches hit too many single words ("ouch!") or two word sentences ("go away"). If their count exceeds 1000, the search result is useless.
> why is it a problem to specify the last word?
Because I there are many possible last words?
Ok, maybe something similar to
^(a*|b*|c*|d*|e*|f*|g*|h*|i*|j*|k*|l*|m*|n*|o*|p*|q*|r*|s*|t*|u*|v*|w*|x*|y*|z*) * * * (a*|b*|c*|d*|e*|f*|g*|h*|i*|j*|k*|l*|m*|n*|o*|p*|q*|r*|s*|t*|u*|v*|w*|x*|y*|z*)$
might work. But that kind of search is too heavy.
Imagine you want to study the word "go". You might search for "* * go * *" to find "usable" example sentences you can use to study with. You cannot get more than three words:
Or less than too many:
Regarding the pure word count search I had a conversation with C K before (where I thought I might solve the problem myself).
The main reason for posting again is the added reason of search syntax consitency ( I know, repeat, but I added clearer examples this time):
> This example finds English sentences that have "Tom", then two words, then "Mary", then one word, and then "John."
> "Tom * * Mary * John"
> This example finds English sentences that start with "Tom", then have three words, then end with "Mary".
> "^Tom * * * Mary$"
So why would "^Tom * * *" not work?
Rajoittamalla kielen englantiin saat myös nelisanaisia lauseita, mutta tämä ei auta paljoakaan. Jos kiellät joitakin sanoja, lauseita on vähemmän ja näet myös pitempiä. Esimerkiksi kieltämällä muutaman nimen yli puolet lauseista on nelisanaisia: https://tatoeba.org/spa/sentenc...sort=relevance
Sen sijaan satunnainen järjestys antaa paljon keskipitkiä lauseita. Ehkä siitä on hyötyä sinulle?
Thanuir: That's a workaround that might 'accidentally' work. It's nothing to rely on. Especcially, because you don't know which words to forbid at the search (remember, we're talking probably about a language the user is not familiar with)
I don't know if it is possible, but would something like the following work for you if this could be done?
In the advanced search have either both or one of these possibilities.
1. When the results are sorted by the number of words, have the option to only display sentences with X or more words. For example, 5 or more words, 6 or more words, 7 or more words, ...
2. No matter how results are sorted, have the option to skip the first so many results. For example, 500, 1000, 1500, 2000, 2500, ...
1. Would be the perfect solution for this particular problem.
2. Sounds as it works, but I think such an option is a bad idea from a UI standpoint.
Maybe allow more than 100 pages in the search results would easier to understand. One UI question here would be: how to (usefully) present (e.g.) 580 pages to the user ( https://tatoeba.org/eng/sentenc...rom=eng&to=und )
I'd guess (2) is easier to implement..?
I know it's not exactly what you want, but I quickly threw together some pages with 4-word, 5-word, 6-word, 7-word & 8-word English sentences with audio.
This is based on some programming that I'd already done, so some of the text on the pages may not be fine-tuned for this purpose.
Thank you for the work. But it looks like you were operating directly on the database?
Looking at gillux link, I guess the discussion is finishied :D
Tarjosin ehdotuksena sinun ongelmaasi, jonka tuossa esitit. Se ei ole mitenkään ideaalinen, mutta toiminee purkkaratkaisuna, kunnes jotain parempaa keksitään.
Se toimii myös sikäli yleisesti, että rajoitteiden lisääminen oleellisesti aina vähentää lauseiden määrää. Niinpä, aina jos löytää vain liian lyhyitä tai pitkiä lauseita, voi vain lisätä rajoituksen. Jos rajoitti liikaa, poistaa rajoituksen.
It looks like filtering by word count is the most straightforward way to solve your problem. This enhancement has been suggested before and the progress is tracked here: https://github.com/Tatoeba/tatoeba2/issues/1954
Edit: it was you who actually first suggested this. Just to make it clear, we didn’t forget, but we have other priorities for now. :-)
Thank you for providing the link gillux. At that time, it didn't sound like the bug was "written down" (to github) by anyone. So I guessed it was sorted out as "not important".
I should then appologize for making a fuss.
Sentence 8945813 Semua daun-daunnya sudah berguguran.
shows wrong translation in all languages.
Correct english translation is: All the leaves have fallen.
I fixed it.