menu
Tatoeba
language
Register Log in
language English
menu
Tatoeba

chevron_right Register

chevron_right Log in

Browse

chevron_right Show random sentence

chevron_right Browse by language

chevron_right Browse by list

chevron_right Browse by tag

chevron_right Browse audio

Community

chevron_right Wall

chevron_right List of all members

chevron_right Languages of members

chevron_right Native speakers

search
clear
swap_horiz
search
AlanF_US AlanF_US March 10, 2019 March 10, 2019 at 5:02:55 PM UTC link Permalink

Recently ([1]), I suggested an experiment where a native speaker of language X who is learning language Y could try leaving comments on selected sentences in language X. These comments would contain proposed translations into language Y, together with a request that a native Y speaker submit corrected versions of those sentences as translations. It's been working very well for me, thanks to @odexed, who has been kind enough to respond to my requests. Not only can I increase the chance of getting translations for the sentences that are useful to me, I am also able to compare my proposed translations against a native speaker's. This gives me practice and lets me see where I can improve. odexed also likes contributing sentences that are guaranteed to help someone. So it's more fun for everyone involved. :)

[1] https://tatoeba.org/eng/wall/sh...#message_31468

{{vm.hiddenReplies[31493] ? 'expand_more' : 'expand_less'}} hide replies show replies
soliloquist soliloquist March 10, 2019 March 10, 2019 at 8:16:19 PM UTC link Permalink

This method may work well with some languages and between friends, but I think English is a special case and this would likely create a bottleneck as it's the 'language Y' for most users.

Another and bigger problem is the difficulty of finding such comments, especially after some time passes. People would leave comments under sentences in different languages. One request would be under, say, a Turkish sentence, and another would be under a Hungarian sentence. One request would be for an English translation while the other would be for German. Categorizing and finding them would be difficult as they begin to pile up.

If there was a search function within comments, it would be easier to find them using some keywords or tags.

{{vm.hiddenReplies[31494] ? 'expand_more' : 'expand_less'}} hide replies show replies
AlanF_US AlanF_US March 11, 2019 March 11, 2019 at 11:50:23 AM UTC link Permalink

> This method may work well with some languages and between friends

Exactly. An approach doesn't need to work for everybody in order to be useful to somebody. Tatoeba is a social medium -- let's be social! You can cultivate relationships with native speakers of the languages you want to translate into. If you don't already have one, you can look up who has been contributing recently in your language Y and has your language X in their profile, and send a private message asking them whether they'd be interested in translating into your language. Then, if they're willing, you can flag that person with an @ sign in your comments. If they have e-mail notifications enabled, they'll see your comments without having to search for them (though I agree that the ability to search within comments would be helpful for various purposes).

When you say that you think English would create a bottleneck, do you mean that you think that explicit demands for people to translate into English would overwhelm the number of native speakers in English? I'm not sure at all that that would be the case, but there's nothing to lose by trying the experiment.

{{vm.hiddenReplies[31497] ? 'expand_more' : 'expand_less'}} hide replies show replies
soliloquist soliloquist March 11, 2019 March 11, 2019 at 5:23:27 PM UTC link Permalink

Of course, as long as we mention some users on comments for translation requests, they likely get notifications and respond. However, mentioning specific users might reduce the chance of other users' noticing and participation. I thought you were referring to a more general concept addressing the whole community, like the @NNC tag.

I'm not against this btw. It would work just fine. But a broader concept as I mentioned on the previous message (keywords/tags to be used on translation request comments + a search function on comments) could be more effective in the long-term.

> When you say that you think English would create a bottleneck, do you mean that you think that explicit demands for people to translate into English would overwhelm the number of native speakers in English?

Yes, that was my point.

CK CK March 12, 2019, edited November 1, 2019 March 12, 2019 at 1:04:28 AM UTC, edited November 1, 2019 at 3:13:56 AM UTC link Permalink

[not needed anymore- removed by CK]

{{vm.hiddenReplies[31500] ? 'expand_more' : 'expand_less'}} hide replies show replies
soliloquist soliloquist March 12, 2019 March 12, 2019 at 6:20:57 PM UTC link Permalink

I guess you're collecting sentences with such comments in a list. Otherwise it would be difficult to check them. I would rather prefer non-native speakers to create their Turkish sentences and then leave comments saying @NNC. It's possible to filter comments by languages so it would be easier for native speakers to notice and respond. Leaving comments for Turkish translations on non-Turkish sentences may go unnoticed as they move to back pages of the comment feed (which happens pretty fast).

{{vm.hiddenReplies[31501] ? 'expand_more' : 'expand_less'}} hide replies show replies
Guybrush88 Guybrush88 March 13, 2019 March 13, 2019 at 5:57:20 AM UTC link Permalink

along with this, what about creating sentences, unadopting them and then letting native speakers adopting them? There's already a long list of orphan sentences in many languages ready for adoption

{{vm.hiddenReplies[31502] ? 'expand_more' : 'expand_less'}} hide replies show replies
Thanuir Thanuir March 13, 2019 March 13, 2019 at 8:13:29 AM UTC link Permalink

Are native speakers adopting them? I adopted all but two Finnish sentences, and I am uncertain about the correctness of those two, but there are about 50 thousand unadopted English sentences, while only 1000 French ones; checking some other languages, this seems to depend mostly on whether anyone has taken the time to go through and adopt the sentences recently.

Maybe some English natives would like to start checking and adopting the English sentences and others should do the same in their native languages, if there is a large amount of sentences without owners. It would make this approach a lot more encouraging.

{{vm.hiddenReplies[31503] ? 'expand_more' : 'expand_less'}} hide replies show replies
Guybrush88 Guybrush88 March 13, 2019 March 13, 2019 at 9:44:11 AM UTC link Permalink

that would be a way also to proofread them and, therefore, improve the quality of the corpus

{{vm.hiddenReplies[31504] ? 'expand_more' : 'expand_less'}} hide replies show replies
AlanF_US AlanF_US March 13, 2019 March 13, 2019 at 12:45:48 PM UTC link Permalink

Personally speaking, adopting sentences is one of my least favorite ways to contribute to the site, and the number of unadopted sentences in a language is one of my least favorite measures of unfinished business. I prefer tasks in which progress can be measured easily, especially when I feel like they are very useful to at least one person in particular. For instance, I find it productive to go through sentences that are tagged "@needs native check" (=@NNC) because I can do something with most of them:

- leave a comment suggesting a particular change, delete the @NNC tag, and add the @change tag
- mark it OK and delete the @NNC tag
- delete the sentence
- mark the sentence not OK (for instance, by using the rating feature) and delete the @NNC tag

It's easy to measure progress because I am taking sentences out of the @NNC bin as I go along. I also feel as though the authors of the sentences have a chance to learn from my comments.

But adopting sentences is different. Often, I come across a sentence that I don't want to adopt, but there's no way for me to mark that decision so that I can skip that sentence in the future. Sometimes my reason for not wanting to adopt the sentence is that the sentence is not necessarily wrong but cannot easily be transformed into something a native speaker would say. Sometimes it's because it expresses a sentiment that I don't share. Sometimes it's because there are a whole bunch of sentences that vary in only a small way (for instance, a pronoun), and I don't see the value in taking the time to adopt all of them.

Furthermore, sentences end up in the unadopted bucket for a variety of reasons. Sometimes it's because the owner leaves after contributing a bunch of bad sentences, while sometimes it's because the owner wrote decent sentences in a non-native language, but wanted a native speaker to take ownership. There's no easy way to determine, at first glance, which sentences fall into which category. With @NNC sentences, it's usually obvious why someone gave them the tag.

Also, the number of unadopted sentences in English is larger by orders of magnitude than the number of sentences tagged @NNC. Going through @NNC sentences takes time, but is something I can manage. Going through the unadopted sentences in English would take a large number of people AND a large amount of time -- on the order of years. I've never been convinced that it's worth it, especially considering the other ways to contribute to the corpus.

However, the situation changes when we talk about arrangements between specific people. If person X agrees to unadopt their non-native sentences with the understanding that person Y will adopt them, then you don't have the "undifferentiated bucket of sentences" problem. I think this points to the same basic attitude that you can see with my discussion with soliloquist above: I favor approaches that involve one-on-one communication, even though I can see the value of infrastructure that allows people to act in ways that are less firmly coupled to specific individuals.

Aiji Aiji March 13, 2019 March 13, 2019 at 2:05:44 PM UTC link Permalink

There are 1000 orphan French sentences left because I adopted several thousand of them already. I don't have time recently to check them so the counter does not go down.

The system of writing translations and then unadopt them have several flaws, most of them explained by Alan. But for me, the main flaw is that unpractical sentences are added to the corpus and it becomes very hard to deal with them. Typically sentences that nobody would ever pronounce (or write) but that are correct grammatically (and correct translations). Also, sentences that have plenty of translations although incorrect are a pain :) (but that is due to the French corpus history mainly ^^)

soliloquist soliloquist March 13, 2019 March 13, 2019 at 5:00:24 PM UTC link Permalink

Sure, it would work, too. It's just that when an unowned sentence is adopted and edited by a native speaker, they may not find it necessary to leave a comment about the correction or improvement they made (and they're right about that). However, if the creator of the sentence wants to be involved and informed, they may prefer not to unown the sentence. Personally, I wouldn't encourage non-native speakers to unown their Turkish sentences as long as they're cooperative with suggestions, but I respect other policies preferred by CMs of other languages. If I want to add an Italian sentence, I'll keep your preference in mind. :-)

Ricardo14 Ricardo14 March 10, 2019 March 10, 2019 at 11:24:06 PM UTC link Permalink

I'd like to help you too. Please ping me if you want a sentence translated into Portuguese :)

{{vm.hiddenReplies[31495] ? 'expand_more' : 'expand_less'}} hide replies show replies
AlanF_US AlanF_US March 11, 2019 March 11, 2019 at 11:51:51 AM UTC link Permalink

Thank you, Ricardo. I can only concentrate on one language at a time (currently Russian), but if I ever go back to focusing on Portuguese, I'll definitely keep your offer in mind. :)