> So my point is: shouldn’t sentences be equally checked whether they are from native speakers or not?

Certainly!

OK, CK gave a solution: "I was able to change this by first changing the Flag to another language and then to "unknown."

In https://en.wiki.tatoeba.org/art...nguage-request there is this section:

<< After the sentences is added, if the language was not identified as "unknown" (you will see an icon with a question mark if it is), then click on the language icon and select "other language".>>

But selecting "other language" doesn't work, see e.g. https://tatoeba.org/epo/sentences/show/7003339

Is this a known bug?

Suggestion: The green, yellow and red marks can only be seen by people who logged in. Wouldn't it be good that also occasional users are warned of bad sentences by the red marks?

I didn't know either. Thanks for reporting this! Great feature!

The automatic language recognition is back!
La aŭtomata lingvorekonilo revenis!

Merci bien, Gillux!!!

Tatoeba ne estas la taŭga loko por serĉi parolamikojn, André. Tio ĉi estas retejo por tradukantoj. Mi sugestas, ke vi iru al la retejo de Lernu, kie oni ofte vidas anoncojn pri parolado per Skype kaj similaj programoj: https://lernu.net/eo/forumo/temo/15335

Here is at least one duplicate of yesterday: #1567612

But again on strike? Or just having a break?

And now i can log in... Thanks Lisa! Which bad fairy kicked me and Guybush out? :-)

Same message: 404. This is not the chat you're looking for. Sorry. :(

tatoeba2 and tatoeba-api works though.

The Gitter site for corpus maintainers is not working anymore?

When I try to log in, I get: 404. This is not the chat you're looking for. Sorry. :(

Congrats! That's good news!!!

David, mi jam ĝentile informis vin, ke Tatoeba ne estas loko por lingvaj eksperimentoj, sed por frazoj en ordinara flua lingvaĵo. Se en propra novelo aŭ blogo vi volas ĵongli per la lingvo kaj garni vian tekston per meminventitaj novismoj, tio estas tute en ordo, sed ne en Tatoeba.

Gratulon, Krystof! Bela varma voĉo!

Very interesting material. Thanks. I saw it with Esperanto subtitles and it works great.

Mi jam avertis lin, ke liaj frazoj estos viŝitaj post 2 semajnoj, se intertempe li ne korektis lin.

> We also have 48000+ "orphan" sentences in English

But that is a different problem. Any native English speaker can adopt orphans. For those a corpus maintainer is not needed. If every native who is active on Tatoeba would adopt ten of them per day, we would solve that problem in a few months.

Done. Corrections can only be made by the author of the sentence. If the author doesn’t react, corpus maintainers can do so. One can ask them to make the correction, by adding the tag "@change" or by adding a note starting with "@[name of the corpus maintainer]", e.g. @nimfeo for French, @Pfirsichbaeumchen for German, @PaulP for Esperanto. Unfortunately for many languages there are no active maintainers, what may slow down the corrections.

Does English really have no functioning corpus maintainer? It is the most used language here on Tatoeba! There are hundreds of erroneous English sentences that have been here for years. Here is just one example:

#4027663

I've corrected a bunch of them today, but not being a native, in most cases I hesitate to touch them.