From the discussion so far, I would put a couple of theses:
Initally, the app was run by language lovers with a warm feel and well-meaning. Besides wanting to cash in, which is quite legitimate in my opinion.
The app has been taken over by sofware (and potential business) geeks which do not care for the actual use in learning, but for “nice code” and reducing costs (easier maintenance).
This leads to changing things which do not need to be changed (from a user perspective) - but reduce cost.
I presume none (or few) of the develpers used the app to learn Japanese. Actuall, this is like a car mechanic not beeing able to drive. Whereas this can still be, of course, at the end of the day, it impeedes knowing what actually is wanted. So if the code is nice, why should it be useful?
The team obviously does not give anything on the users complaints. But why should it? From the lifetime-sentenced users, no money can be derived anyway. So if you have the bucks, why care? And then, the app is still popular. And there is no froum in which one can unabigously warn from choosing it. So, at the end of the day, why care. (Such a thing as #wantitled ironically puts, is not so far from reality. Perhaps not literally, but in the sense). My proof (as Yumari-1 put it):
Highly used and appreciated feature gets removed.
Users broadly complain.
Management sells them the usual corpo-speak “evaluating options/looking for alternatives etc”.
Users continue to complain and some find a workaround to bring the feature back themselves (rev summary on dashboard script).
Workaround gets axed.-.
[add.] and this workaround was basically necessary because the features got removed.
besides that, there has not been any earnest communication in changing the criticiesed items. So they are just sitting it out.
I fully side with Yumari-1 in the analysis s above, but also, in that we are to kind to accept all of this.
But that’s to be expected, seeing as they’ve essentially ignored everyone. I don’t think it’s that people are getting used to not having a summary page, it’s just everyone’s learning that WK doesn’t care.
It really does seem like things have gone downhill fast in respects to the dev side of WK. I could be completely misremembering, but I think I saw a message on the update thread posted by a staff member addressing userscripts being broken, claiming that that was the function of the API. And then what happens just a few weeks later? Access to a massive endpoint removed. And I understand that server costs are a thing, and there was probably a large influx of users making calls after the summaries were reviewed, but Wanikani absolutely needs much clearer and concise communication between the devs and the community. And transparency - some changes are promised but seemingly never touched on again. I love wanikani and will continue using it, but they need to take a serious look at current operations and reevaluate how all of this is approached.
I guess we all have opinions but the website feels much better now. The update even got me to use the mobile version sometimes. The summary page wasn’t even that useful, you already now what you failed or succeeded at when doing the review, and the Apprentice/Guru/etc ratio tells you if you made progress when back at the main page.
Never used scripts. I understand it’s frustrating coming to rely on some assists/cheats or whatever they do, but they really aren’t necessary for using the website and learning the words and characters.
Compared to the majority of websites I experience daily, this one is shockingly functional.
I’m also enjoying the improved native experience on mobile, but I understand that it’s immaterial to folks who use the site in a heavily modded way or via third party app.
I agree they could do a better job of keeping people informed of long-term planned changes. But in the case of the recent database issue, there’s only so much you can do when scripts start to impact site health. I’d rather have a website that works versus extras like streaks.
I do think script users tend to assume that every relies on scripts, just because they do. So I’m not sure I buy into the notion that the userbase overall is unhappy, just because the minority of people who post to the forum seem to be.
Sorry if im out of the loop, but did wanikani change ownership, is that what you are saying? Or are you just talking about some old codes they wrote before hirings profesionnals?
To be fair, getting all reviews would be slow no matter whether it’s accessed internally or by script authors via the API. Fixing that requires changes to the database which are harder to do if you don’t want to suffer downtime.
Funny thing is: bunpro did not have a summary page and added one, meanwhile wanikani took it away.
I do like not getting timeouts though, so at least something good came with the bad.
I suspect that most of the load on that review API was for heatmap-style functionality which would be trivial and efficient to implement as a first party (and a pretty standard widget in most SRS apps in my experience).
The API is already pretty aggressively throttled, you can only do one request per second.
I think the problem is that WK’s API doesn’t offer any consolidated review data, so if you want to do heatmaps and the like you have to fetch every single review ever made. On top of that I suspect that the table(s) holding the review data are not well optimized for that use case, and the query to get all reviews for a single user requires loading and filtering a huge amount of data.
I honestly think that this could be optimized fairly easily, if only by just archiving review data for inactive users (to reduce the size of the “hot” tables). Of course it’s easy to say from the outside, it might be more complicated in practice…
I believe heatmap can be a double-edged sword where in some cases it can motivate people and in other cases (such as when you lose your streak), it can also de-motivate. But yeah, native heatmap implementation (perhaps something optional) would be a lot easier to optimise.
Yeah I don’t really care about heatmaps and streaks personally, although I can see how others might find them motivating.
One thing that I do miss however is getting a sense of how my review count is increasing so that I can adjust my SRS targets and not burn out by starting too many new items at once.
The removal of the review API fell at an awkward moment for me because I just started burning items a couple of weeks ago, but now I have no simple way to estimate how I’m doing with those burn review sessions and how much of a load it’s adding to my routine (and how much it’s going to keep adding as I make mistakes and reset some entries to lower levels).
Some people go really overboard with gamification. You can feel that for some the game becomes the objective rather than actually learning the language.
That’s how you end up with the stereotypical Duolingo user who has a 1000day streak in Spanish but still can’t speak the language decently.
