Too few words for so many lessons
There are currently eight to twelve times as many lessons as there were before
I think it's time to add new words to the higher crown levels.
Otherwise the achieved vocabulary would not be proportionate to the time spent to complete a course (now eight times bigger)
Yeah I was really confused when the email announcing the crowns update kept saying "More to learn!" When, no, not really. It's the same lessons that were there before you just have to repeat the same lessons over and over more to show progress and you can't pick and choose which part of each lesson you want to work on, you have to do them all. It's not "more to learn" it's just "requires more work"
Getting an active command over the vocabulary when only passive (comprehension) was realistically possible before in many instances would seem to provide a plausible basis for the "more to learn" claim. Also, sentences that were previously "retired" for being missed too often have returned.
I think the Crowns system is independent of the lessons and vocabulary in a given language, so being switched to crowns has no impact on the amount of vocabulary available in whatever language you're working on.
I haven't gotten the new French 3.0 tree yet, but the reports are that it has additional words as well as additional skills. That probably means that other courses also need to be revised to offer additional vocabulary. A revised course is also likely to have new skills in it.
Why is the course 8 times bigger? Cause of the crowns?
Don’t the crowns just more or less keep track of how often you have reviewed a skill? Not increasing the course.
Unless you are ocd and must be level 5. Can’t you just complete everything to crown 1 or 2 and just review as normal?
Like if I gave you a book that is 200 pages and then later I gave you a sheet so you can record how many times you read a page, it doesn’t increase the number of pages in the book.
The problem with this approach: the lower levels are much easier with little translation now, at least from my observation and from what I've read. If it was all translation to the target language, I would agree with you. But right now, just using level 1 and 2 is not such a great idea.
And not having more words at the higher levels, that is a bit of a wasted potential.
The book in your example isn't supposed to be used like this. In Duolingo, learners are encouraged to spend a certain amount of time rereading each "page". There is no OCD in this (really, americans should stop overusing psychiatrical terms and labeling everyone and everything), there is trust towards a language learning resource and also the need of many people to finish things before leaving them. Many people see completing the tree absolutely similarily to completing a real coursebook instead of leaving it in the middle.
you can't add words to higher crowns for two reasons.
logistically, it would take forever to do. months or years of actual work and still more months (hope not years) for the inescapable A/B testing.
structurally, as you go forward (down) in the tree, new vocabulary is shown by being combined into sentences with old vocabulary. but there is no telling what crown level the prior skills will be at. i suspect we will have users who will leave the beginning of the tree at whatever crowns the conversion gave them, as low as level 1, which is also what placement tests seem to result in. we could only ever rely on level 1 vocabulary of the previous skills when composing sentences in the rest of the tree. or put differently, the extra 4 levels of vocab in each skill would be used only in that one skill, never reused later on.
hardly seems worth it.
you can't add words to higher crowns for two reasons
And a third (technical) one: words are added to a skill itself.
@Mereade , things roughly works like that:
- volunteers (not staff) add vocabulary to a given skill (during course's creation) and associate sentences to vocabulary (at any time)
- crown-system (=staff's work) decides which sentences are used for crown-level 0+, which for crown-level 1+, which for crown-level 2+, ..., which for crown-level 5 and which are never used.
So crown-system controls sentences, not words.
Quote: crown-system (=staff's work) decides which sentences are used for crown-level 0+, ... and which are never used.
So crown-system controls sentences, not words.
Why should be a system alone in total control (auto rating sentences because of average error rate) of what more complicated and longer sentences are deactivated, even if a user has passed crown level 5?
Do course admins / contributors at least have an overview what those sentences are, so they can be corrected, changed or manually flagged "very difficult" (>above crown level 5).
Well, I know the answer from a different thread were a mod (you?) replied to me:
The Incubator does not provide (yet) a manual default sentence difficulty mapping for crown levels L0-L5!
How often are sentences changed, enhanced in the course database to reset the error interval?
Does the RESET mean that those more difficult sentences could be re-displayed to crown levels L0-L3?
Why can a user not FORCE the system (on the web portal) to display "VERY DIFFICULT" for a tree which 2-3++ years have been spend on?
An "average error rate" does not necessarily mean that those sentences are wrong....they simply could be quite difficult and e.g too long to be used in "timed practice".
- Answering this, I think, rhetorical question: I don't know. Only staff has the back-the-scene reasons for their decisions.
- I can only answer for contributors. General overview, no --- but us contribs have created our own scripts (on the incubator) to have some kind of basic partial overview --- but asked for years. And have shared with Duo that discarding the sentences in the (few) cases where the high failure was due to an error in the exercise was not only inefficient but also counter-productive as we don't have an overview pointing us to those hence with users not having them, we're even less likely to be aware they have a problem.
Information available once a given sentence, yes.
Flag them (about anything), no.
- Wasn't me. ;)
- AFAIK, each time a modification in their list of accepted answers is performed.
Otherwise, still AFAIK, for those discarded, the system keeps "showing them back from time to time" to re-check if it's still hard or not. It could change as hints could have been changed, T could have been changed/created, etc. I mean for reasons "external" to purely the exercises and its accepted answers.
- I'm not sure but I think so.
- I don't know. Only staff has the back-the-scene reasons for their decisions.
Like many (like you, I think to remember), I've for long suggested/been in favour of letting users having a "easy-normal-hard" general control in their settings.
- Yes, being highly failed can be for plenty of reasons (and in most cases it's not because they are wrong, nor because they miss anything) among which simply that they introduce a concept that is hard for example because it doesn't exist in the source language (for example French partitive articles). In such cases, it's absolutely normal that they are failed, and they shouldn't be discarded (even in crown-level 0), otherwise, how to teach that concept? I mean, if the skill is about teaching French partitive, what's the sense of discarding (all) sentences with it?
I agree that it is more work. But I do not agree with the second argument. When someone starts an exercise down the tree, the server has the student's identity and therefore knows the student's skill level for every prior (and non-prior) skill.
If the server is programmed to dynamically create the sentences, then it is able to provide exercises that match the student's set of skill levels.
Words taughed in higher skill levels may have associated pointers that refer to one or more lower level alternatives. In this way, one can construct an example phrase with a level 5-word, that is presented with a level 3-word instead if the student's skill level for the relevant skill is 3.
The "server" does not create the sentences. People do, which is the substance of svrsheque's second point.
Back to the comment about taking 8 times longer to finish a course, I haven't done the maths but I suspect that getting an entire tree to crown level 5 will take you far beyond level 25 in terms of XP. So maybe the XP levels need looking at too?
For the maths: a tree with 250 lessons (= "sections" inside skills) would bring you exactly to 30 000 XPs, hence XP-level 25, if you bring all skills to crown-level 5.
And most trees have more than 250 lessons, some more than twice that number.
Given that they seem to have eliminated the notification and lingot grant upon reaching a new XP level and the XP levels are now deeply buried on the app (at least the iOS one), I think it's more likely the system will simply be eliminated than expanded. I don't like this, but the writing seems very much on the wall.
I have that feeling too - although that would also eliminate the leaderboards and any ability to compete with other users so it would strike me as a particularly backwards decision. It's almost the last way you can interact with other learners outside of the forums now. After all, there is no point in adding people as friends if there are no XP to track and compete with...there is no other way of interacting with them.
People who don't visit the forums (which, I think, is the vast, vast majority of users) probably don't really follow anybody they don't know personally. They have competition with others through clubs, which apparently also allow communication.