Okay, so we just hang out with a bad system because we'll never be able to make a perfect one.

I know each country has its own varities, but most of the time the differences are not as problematic as differences among countries, or between Valencian and Catalan, where the political discussion plays a role. Tags could be used for these smaller differences, but if you feel like the Spanish spoken in a specific region differs much from that spoken in the rest of the country, I don't see why not adding a specific marker. It's a flexible thing. While the ISO code and the current classification remain the same, the addition of a new tag-like flag apart from the countries could be dictated by requests and discussions on the Wall.

A system will never please everyone, but does the actual system? It's not about eliminating every single problem; it's about trying to reduce the ones we can. Looks like a step forward. This is more about making Tatoeba inclusive to the rest of the world than to make it linguistically perfect.

Since this would be an optional resource for ordinary users, I wonder why one would oppose it so much.

:: isnamar's suggestion for language representation ::

Dear community,

The use of country flags has always been a debated issue on Tatoeba. I'm not getting into details on why using ONE flag is problematic; instead, I'm going to introduce and illustrate isnamar's idea to overcome this problem.

He suggested implementing an option to add flags just like tags. I'll call these flags tag-like flags. The standard icon (i.e., the one shown in the homepage and over comments) would still be fixed. It could be, say, the flag of the country with the most speakers, the 3-letter ISO, or even the current icon itself. The difference would reside in the information contained in the sentence itself.

Say a Mexican user writes a sentence in Spanish. The standard icon would be provided by Tatoeba, just like it is now, to indicate that the sentence is in Spanish, doesn't matter from where. Then this user would be able to add, just like a tag, the flag of Mexico, which would be displayed at the right of the sentence, somewhere above the tags, meaning that that sentence could be said by someone from Mexico. Then an Argentinian member could also add the flag of their country if they feel like that sentence could be written by or said by them.

I've created a preview of how a sentence could look like with this extra data. You can check it here:
http://postimg.org/image/bqxpy9hy7/

Adding tag-like flags one by one could be tough, but we could implement something to automatically add the tag-like flag of a native speaker, so he doesn't have to select his most used flag manually, one by one.

Here are some reasons why I support isnamar's idea:
1. Every country flag would be covered. People from Portugal, Brazil, Angola, Mozambique etc (all Portuguese-speaking countries) would all be able to find their flag and add it if they feel like that sentence could be said there. They wouldn't feel misrepresented in the website.
2. Regional differences would be better taken into account. A sentence having the tag-like flag from Brazil but not from Portugal would indicate that that sentence could be said in Brazil (it wouldn't indicate that it couldn't be said in Portugual or other Portuguese-speaking countries though).
3. It doesn't really have to be countries only. This resource could be used to tell Valencian from Catalan, for example. It could also be used to indicate that a sentence is the same in both dialects, just by adding both flags.
4. This wouldn't damage the corpus at all, since it's just extra information that will be added to the sentence (just like tags are). Such a system wouldn't violate ISO 639-3 standard.
5. This would provide us very interesting data. When clicking on the Cuban flag, for example, one could easily get Spanish sentences that a Cuban could say.
6. You really don't have to spend time adding tag-like flags if you don't want to. You can just keep doing what you've always been doing, using the standard icon only without any tag-like flags. Other users who care about flag-tagging sentences could do it.
7. Yeah, Tatoeba would look more beautiful, I think.

I'm pretty sure isnamar's suggestion deserves to be discussed.

I see them now. :-)

Thanks.
Is there a reason why the Portuguese strings haven't been updated? Could you please include this in the next update?

> I just don't understand why it should be something different

It shouldn't, really! I just thought it would be easier to classify and organize content in a different way, since that new area wouldn't be vulnerable to classic mistakes from new users. I also thought it would be easier for ordinary users to export sentences from there.

But it doesn't have to be like that. As long as people like you continue making suggestions and people like Trang continue coding them, we'll have a better website day after day.

About the horchata and fartons, it was just a joke, you know.

I'm done with this thread.

> We already have an option 'Owned by a self-identified native

You're right. I had forgotten this.

It wouldn't be a perfect Tatoeba. But, by containing sentences filtered and proofread by several native speakers, it would be more suitable for people who, like CK, are trying to make some use of our sentences. Of course another possibility would be to improve our own rating system as well.

P.S.: if you ever come to Valencia, don't forget to try horchata and fartons, combined. The word "perfection" will acquire a new meaning. ;-)

I like this idea. Although it won't prevent users from encountering bad sentences, it's a good place to start.

My greatest discontent with Tatoeba is that it's primarily concerned with quantity, not quality. The welcoming message reads: "We are collecting sentences and their translations. You can help us by translating or adding new sentences.".

A day will come when we'll have to discuss what to do with these millions of sentences.

> I wonder if we need to take on the responsibility of taking on yet another dead language

You would have to tell Trang to change her "In Tatoeba, all languages are equal" slogan first.

So we share the same idea of collecting these natural, correct sentences together. The only difference is that I thought of a second place for that, preventing them to be displayed among unnatural, incorrect ones.

I totally oppose your suggestion of making sentences unchangeable. Every user should be able to change their sentences freely.

Besides, just in case you misunderstand my use of the verb "to split", what I proposed was to make a mirror reflection of Tatoeba containing only the "trustworthy" database. As things are for now, the only way you can do this is creating your own lists. But we don't need to get into this discussion again. The thread was closed long ago.

It's impossible to have an error-free corpus UNLESS we decide to split Tatoeba into two sections, one for general contributions, and other for filtered, double- or triple-checked sentences. This was proposed by me a couple of years ago, but I admit it's not a simple thing to do (and maybe Trang never wanted to do it either - it's her website after all).

If I were to act like you're saying, never would I have used Wikipedia again. :-)

We're all humans prone to error.

> Unfortunately, we have some members who treat tatoeba.org like a playground

Let's try not to be rude. One might scare people away with such comments.

It's undoubtedly true that translating sentences into your native language is the best way to contribute, but it's not the only way. Even though many sentences written by non-natives are bad, some are very good and might perfectly fit the project after reviewed by native speakers. We've got our rating system. Let's use it.

In the case of ancient languages like Latin, Ancient Greek and Sumerian, I think their presence should be praised, not questioned. That some users write bad sentences (and there's no way you can tell that unless you do study these languages) doesn't mean all sentences are bad.

+1

I'm also working on a better translation for the Portuguese version. There were lots of untranslated strings and the ones who had already been translated were not uniform or/and contained too many mistakes. I'll be done by tomorrow or the day after tomorrow, so it's good to know Tatoeba will be updated soon.

It's working great now. Thank you very much!

After the update, Tatoeba is so slow in my computer that I can't use it anymore. In Firefox, the interface looks OK, but I can barely scroll down and up. In Edge, the icons look weird and the buttons are not functional at all.

+1

Indeed. Good job.

Clear enough. Thank you.