Immersion: bonuses for focusing on one document.
I would like to propose the introduction of bonuses for people who try to focus on translating large portions of a single document, instead of just picking easy sentences here and there.
1) It promotes better quality. Whenever I see a translated sentence in the middle of an untranslated paragraph, I'm almost sure that I will have to correct at least something in it.
Context is crucial, and the lack of it causes nothing but chaos (and would never seriously be done in the professional world). Sticking to one document, to one part of it, to one paragraph, at least, helps build up the understanding of the context by the translator.
Plus, I'm not sure how it is with you, but I find that I get a sense of responsibility for how the translated document turns out when I get to translate larger parts of it. Which means more quality.
2) It promotes people challenging themselves. When you don't get to choose every sentence you translate, sometimes you need to do something that isn't easy and requires looking stuff up, or checking other dictionaries, or googling phrases.
Challenge is good for learning, and taking it up should be rewarded.
I've got two ideas:
1) Giving coin bonuses if you translate a certain percentage of the article (like 20 or 40%, or maybe other milestones).
2) Giving coin bonuses for translating consecutive sentences in a document (say, 5 or 10 consecutive sentences).
Feel free to share your own thoughts and ideas in the comments.
NB: I am not against the practice of doing small sentences here and there without sticking to any particular document. It's fine for Immersion to have an 'easy' mode. But I do advocate rewarding people, who take up bigger challenges.
There is a 10 xp reward for completing an entire document. I have advocated in the past for a lingot to be attached to that as well, as people often give up on some of the larger documents, and just 10 xp often isn't enough incentive to muddle their way through.
I've never noticed it, even though I've completed entire documents. 10 is pathetic, especially if you're a tier 4+ translator and you've done a 200+ sentence document.
But it's important to promote doing more consecutive sentences, not just finishing entire documents (and I don't think it's reasonable to expect that people would often do it): even that would encourage people to understand the context of what they're translating and do better in the end.
What is the "it" you refer to when you say "I don't think it's reasonable to expect that people would often do it"? Completing a document or doing consecutive sentences?
If people complete a document, then I think it's safe to assume they would get the context. If context is important, then it is important throughout. Indeed, I've translated things where something in the middle or toward the end of the document references something in the beginning.
The problem I see is that it seems like too many people only work on a document in one sitting, and then assume they have to go onto something new the next time they go to immersion. There is a reason for the "your edits" section. It allows you to take your time working on a translation. I usually spend about a week on an article before I go onto a new one, working on a couple of paragraphs at a time.
Maybe it's just me, but I think quality is more important than quantity. A well-translated article, with the entire thing thoroughly looked over by multiple people, is a more valuable product than half a dozen articles thrown together with patchwork translations.
Yes, by 'it' I mean completing the document from start to finish. And I don't think it's reasonable to expect that the bonus for doing this concerns a significant portion of the contributors.
Of course doing the full document better explains you the right context, but encouraging doing full paragraphs or sections instead of just random sentences that happen to be easy is a good middle ground: much better than nothing.
As a general example, one of the most common mistakes I see in such 'black in the sea of blue' sentences is the lack of purely linguistic context. People simply don't choose the right meaning of a polysemous word/phrase. You don't need to have read/done the rest of the text to make the right choice: one or two previous sentences is enough.
Off the top of my head, I've seen the German word 'auch' translated as 'also' (or 'too'), even though it was an intensifier there and should have been translated as 'even' (as in 'even he managed to do it'). There were clear hints in the neighbouring sentences.
I think there is a decent bonus in place already. I got a 60 XP bonus for translating an entire (relatively short) article from Portuguese, and I'm only level 1 for Portuguese.
Well, if the tier multipliers apply, it's 480 at level 4+. Which is decent, I agree.
This is not exactly relevant to my post, though. :)
People jumping around and snagging easy sentences is being proposed as a problem. I don't see it as a problem.
Generally, these folks fall into 2 groups:
(A) less skilled translators
(B) people who are taking short-cuts for the "game" aspect of Duolingo.
Finding easy sentences is good for group (A). If group (B) finds easy sentences, that, in itself, it is not harmful to anyone.
Translating an entire document is a task that by itself generates a significant number of points. The massive number of points one gets from translating is the Grand Prize. The 10 point bonus is just a nice little cherry on top. ^_^
'People jumping around and snagging easy sentences is being proposed as a problem. I don't see it as a problem.'
I don't pose it is a problem, and I don't see that it is one either (although I am not a big fan of people blatantly 'playing the game').
I am saying that particular behaviour should be encouraged. That is, people should be encouraged to focus more on doing bigger chunks in fewer documents rather than smaller chunks in more documents. (And I am not really speaking about giving bonuses for translating entire documents, which is definitely not something people often do.)
If you want to cherry-pick easy sentences, I have no problem with that, but if you focus more on quality and take bigger challenges, you should be rewarded.
You are right. I misread you. I apologize. I appreciate that you pointed it out too.
I've read a lot of variations on the "complete document should give more bonus points" and the complaint of "people jumping around just translating easy sentences". Apparently, it has permeated my brain.
Translating is the fastest way to earn points on Duolingo. That detail inspires a lot of complaints (from people competing via leaderboard: translators will always win when putting in the same amount of time learning as the people using other features). If duolingo increases the number of points a person gets at the end of the document, it means it is the fastest way to earn points-now with even more fast.
Immersion is a very important feature of Duolingo. It is the sole reason why all of our Duolingo accounts are free. Looking for ways to improve it is not a bad thing.
But, the website needs to have some form of balance. If people feel that the time they spend doing lessons and using the Practice button "aren't worth their time", then the quality of translations in Immersion will drop.
I'm not saying staff won't decide to implement the idea of a bigger end-document bonus. But, it also makes sense if they don't. Another option could be to lower the points gained for each sentence translation, and use the difference to increase the end of Document bonus. The drawback to that would be that again, lower quality translations and less skilled translators try to take on more than they are ready for.
I hope this better explains my thought process on the matter. I'm not Staff though. I can only guess at things like any other user.
'Translating is the fastest way to earn points on Duolingo. [...] If people feel that the time they spend doing lessons and using the Practice button "aren't worth their time", then the quality of translations in Immersion will drop.'
The quality of translations will not drop with additional coins given out for translations, because on the first 1-2 tiers you don't get coins faster than by practising (I'd say, it's the opposite, with Time Practice). And if you are bad, you won't get promoted at best, or will be reported at worst.
'I'm not saying staff won't decide to implement the idea of a bigger end-document bonus. '
And as I have repeatedly said before, I am not talking about bonuses for completing the entire document. They don't even have to exist, for all I care — it is not reasonable to expect that people can get this bonus often, so it's mostly pointless with respect to the quality of translation and what language learners get out of it all.
My point is to promote translating consecutive sentences and bigger parts of documents through bonuses. So that more people see the benefit of translating an entire paragraph, or a section (which is good, see 'Why?'), as opposed to just cherry-picking the shortest and/or easiest sentence and then proceeding to the next document (which is not as good).
If you do cherry-pick, be my guest. But if you translate many consecutive sentences, paragraphs or entire sections of a document, regardless of their difficulty, this effort should be recognised and rewarded.
And to make it clearer, here's the example I've given above:
'As a general example, one of the most common mistakes I see in such 'black in the sea of blue' sentences is the lack of purely linguistic context. People simply don't choose the right meaning of a polysemous word/phrase. You don't need to have read/done the rest of the text to make the right choice: one or two previous sentences is enough.
Off the top of my head, I've seen the German word 'auch' translated as 'also' (or 'too'), even though it was an intensifier there and should have been translated as 'even' (as in 'even he managed to do it'). There were clear hints in the neighbouring sentences.'
Promoting behaviour that helps avoid such problems is good, as is exactly what I'm talking about.
One problem I encountered. I enjoyed an article and finished it. But it seems that after the first half no one read it. I presume they saw the bar was full, so they thought it wasn't worth opening. Something I don't understand, the first person to reach eg 5, how did they get to six since there was no one above them? It puzzles me. Thanks
It might be nice if the award ignored sentences of length 10 or less. I really hate voting up a "translation" of a number or a one-word label just to be complete.
Actually, what might be interesting would be if, having completed the document, you gained some sort of authority perk--something to let you override inferior translations in the doc and make it stick.
This can be problematic. The length of a sentence often does not accurately represent the difficulty of translating it.
While I know what you mean, I don't think that a sweeping restriction like this is viable.
I don't think a small sentence needs any award at all, and apparently that's what the new system does. It only awards up-votes for sentences with more than 5 characters.
I don't think the second idea would ever work simply because there is no way of judging whether the person has done a good job while translating or proofreading the whole document. Perhaps only if they consistently complete many documents and on average receive 75 % upvotes, for example.
I think we should just create a new "culture" of rewarding completion of documents. Using threads to give people lingots for working on a particular document. Like I wrote in the other thread:
It's not exactly what I am talking about. Quite frankly, I would be severely discouraged from participating in Immersion if it were built in a way as blatantly exploitative and gamified as what you describe.
There is no exploitation, people can just refuse to do that. A more important thing is to ensure that the whole document is proofread, consistent and evaluated properly. Those sentence jumpers cause mayhem because they lack context and create incorrect translations. This serves as an incentive for people to go and complete the document. It is no less gamified than receiving points for translating a couple of sentences, or rising through tiers for being up-voted a lot.
I would rather vote for someone that has proofread the document thoroughly before translating, rather than someone who is sentence hoping and wants points for doing that little work. But anyway, to each their own.