I understand your comment about wanting the next post to be about the actual solution, but even so, after all this time… is there any progress to report on?
3 Likes
It would indeed be nice to get an update on this - long enough has passed that surely you must have a very approximate timeline at least? Or even if not it would be great to get some reassurance that it’s actively being worked on and not just abandoned on the backburner…
6 Likes
“Lily. After all this time?”
4 Likes
I guess that’s a no then. Got it…
5 Likes
It’s been 85 days now and still the best we have is “it’s in the works” with no further details? As harsh as this is going to sound (and I don’t want it to sound harsh, but I realize that it indeed does), this really shouldn’t be that hard of a fix. WaniKani is not the first app in the world to try and organize hundreds of millions of records. It can be done, and it doesn’t take 85 days of silence to find a solution.
I understand that this may not be the company’s #1 priority, seeing as it only affects 3rd-party developers and their userbases, but I would like to think that those 3rd-party applications are important and valuable to the company? And ignoring the issue for 85 days (or at the very least, appearing to ignore it and not giving us any reason to believe otherwise) is not necessarily the best way to show that you care and value the effort of the hundreds of people who have worked to make WaniKani what it is today. (And worked for free, mind you.)
Technical difficulties happen, and they do indeed suck. But this is the type of problem that should be solved in under a month, or if it’s especially daunting, 2 months. But 3 months have passed and we don’t even have a timeline for when or even if it’s going to be addressed. I don’t think that’s right. These people have spent money, time, and energy to make your application better, and you just pulled the rug right out from under them and told them to just deal with it while you… Do nothing to make it easier on them? It’s one thing to not give them any alternatives or workarounds, but it’s entirely worse to just outright stop responding to them and basically say “you won’t hear from me until some undisclosed time in the future which I won’t talk about because I can’t even guarantee that that time will ever even come.”
I think solving this problem should be very high priority now. Putting it on the back burner 3 months ago was iffy, but acceptable. But leaving it on the back burner now that 3 months have passed with no resolution? It’s definitely time to step in and solve this problem now, and quickly. People are (clearly) waiting, and they’ve been waiting long enough to at least get some idea for when the waiting will end.
I love WaniKani (and Tofugu as a whole) a lot, and, for now, I’m more than happy to spend my money on this incredible app. But I really don’t like to see the steps that have been made over the last few months. It feels like we’re receding, and it feels like the technical side of the app has been on a steady decline with seemingly no plan to put it back on course. Please don’t allow the app to slip into complete disfunction, and absolutely please do not leave your loving and passionate fanbase in the dark. All these people in this thread complaining about the changes - these are people who care about your app. These people are invaluable. They need to be your first priority. If you lose these people, you lose what makes your app so great. Don’t lose them.
20 Likes
Or they could provide the functionality as a first party feature and remove the need for these third parties. Basic review stats are a pretty standard feature for most SRS products in my experience.
I’ve used my share of language learning websites and apps to study various languages, WK is the first time I need to install third party scripts or apps to get a decent review environment. It’s pretty wild. It’s not a cheap service either. I pay less for bunpro and I feel like they put so much more work into it. I’ve reported bugs or missing features and they had a fix deployed in a couple of days! And they have first party smartphone apps!
Meanwhile WK has actually become worse since I’ve stayed using it… Thank jeebus for flaming/smouldering durtles.
11 Likes
I have now entered the phase where I actively recommend friends to not get into WK.
15 Likes
The review_statistics endpoint still exists, and you can get information about overall stats per item that way (like item stage, correct and incorrect counts, and it can be filtered by date to get a minimum estimate of reviews in a day) but yeah I’ve been mostly fine with the changes because I know programming is pain but this is taking too long.
4 Likes
I wasn’t aware of that, all I know is that the third party tools I used to get that info have all been broken since the API has been disabled. I didn’t attempt to write my own stat tool, although now I will consider it…
I agree completely. For this, the kana toggle and the review summary, I don’t get why the dev cycle is so long. I suspect that the devs are just busy elsewhere.
1 Like
A lot of functionality that these sites used could probably very easily be recovered by just using review_statistics (and other appropriate endpoints, but I can’t speak for all sites). I used WK History until the endpoint got disabled and about a third of the information could easily be fully recovered, with about half being recoverable by estimation from review_statistics. It’s kinda silly that 3rd party devs aren’t making use of less intensive endpoints (less load on WK, faster to load the information, and easier to look through).
1 Like
I noticed that many third party extensions and websites seem to use the same library to interface with WK (because I see the same dialog pop up with the progress bars while loading data from WK on all of them). I wonder if it’s just that that library is all-or-nothing with the data it fetches.
WK Open Framework it is. And it’s not really an all or nothing, as some userscripts don’t need all information (some might just need assignments, or review_statistics to work, like WK Ultimate Timeline).
1 Like
For me what I miss the most is the wanikani heatmap 
I’d be happy with a first party version of it as well, its a pretty common feature across lots of other study tools I use.
8 Likes
What I miss most is the freedom it gave me to make third-party apps. I was going to make an awesome one where it automatically compiled and exported a timelapse video of your entire WaniKani history in a format similar to the WaniKani Wallpaper Generator. I spent like an hour reading through the docs and planning out exactly how I’d do it (and making sure it was possible with the information the API provided), and then as soon as I started to actually write the code I realized it’s no longer possible because the entire endpoint was disabled. Very disappointed with that one. Would have turned out super cool.
Maybe another time if we could ever get some news on this… I’d still love to give that a go.
9 Likes
Dam that does sound sweet. I’m optimistic they’ll turn the data back on at some point. Hopefully it comes back and you can build it. Yeah I’ve been so impressed with the 3rd party eco system around wanikani in general, so its a bummer to have a fair amount of access cut off.
1 Like
I hate to sound like a broken record on this, but this is honestly getting absurd. WaniKani is so very eager to take away features at the drop of a hat and then promise, promise, promise, but with absolutely no follow through. The last message on here mentioned that the next one would be a solution or plan - does that mean no plan has been made at all? If that’s the case, what’s “still working on it” supposed to mean? Realistically, how long should a change like this actually take? Even the API docs haven’t been updated to reflect any changes yet despite that being mentioned back in April.
The part that’s so confusing about it is the complete lack of communication, despite communication being promised. At this point, if something’s hindering the progress, just let us know with an actual update on the thread or something similar. That absolutely doesn’t mean I and others would stop complaining if there was one, but it would at least feel like the team is actually trying to do something instead of completely ignoring the problem and shoving it under the rug - which is exactly what it feels like. I’ve honestly never seen anything similar on another platform that involves them fully disabling a feature, promising a fix, and then leaving for months.
And it’s not like there isn’t work being done at all on WaniKani - the Recent Mistakes panel proves that. So what makes this one so special? Is it not high priority enough? If there’s a bigger project in the works, clearer transparency with that would also be greatly appreciated. But it doesn’t seem that way. It would benefit both the customers and WK if they actually listened to the community, and responded to and discussed issues brought up, instead of adopting this seemingly locked-in development. Bunpro is a phenomenal example in that way, having proper beta programs available, taking in and genuinely considering feedback, and implementing community suggestions. All while being cheaper than WaniKani.
I really don’t mean to come off as overly-harsh here, although I’m definitely being critical. I’m writing this because I truly love WaniKani as a service, and I want to see it continue to thrive. I’m definitely gonna keep on using it, but all of the recent changes over the past year or so have me seriously concerned for the future of it.
17 Likes
It stands to reason that, while the silence is damning for WK, anything they could possibly say would be even more damning. It would either very transparently be a lie / empty promises, or it would be truthful and make most people jump ship.
I believe that WaniKani is essentially a dying product. Whatever resources they’re able to allocate for development (and by now we know that development is on an hourly, as-needed basis) isn’t able to fix all the tech issues they have and their latest attempt to generate momentum by introducing Kana-only vocab seems to have backfired and is now on hold. Probably, competitors are now eating WK’s lunch and WK is unable to keep up. WK might not shut down tomorrow, but the momentum might just be gone at this stage. Even Koichi seems to have left - maybe he sold the company or is otherwise disinterested but there has been no word from him in recent memory.
And if WK should ever shut down, don’t expect transparency - the death of EtoEto was only mentioned in a tweet.
This is all just theorising, I might be wildly off base. But rumours and speculation are what happens when communication is essentially radio silence or deliberately misleading.
11 Likes
If I may ask, what WaniKani alternatives/competitors currently exist? I am not aware of any that even come close to competing, let alone “eating WaniKani’s lunch.” WaniKani still seems to be a best-in-class product, but I’d be happy to be shown that I’m wrong.
2 Likes
I’m also really curious… What is standing in the way of simply partitioning the reviews table by user??? I mean, yeah, sure, it’s a hard, time-consuming, and risky thing to do… But it’s not that hard or time-consuming or risky. 4 months is PLENTY of time to create, test, and execute a plan.
The application never needs to access the review data of more than one user at a time, making this table the perfect candidate for partitioning by user. You could have hundreds of users to a partition, and that alone will completely and permanently solve the problem. Why has this not been done already? If you can’t come up with a better solution than this, then just do this. Stop kicking it down the road and making everybody wait for nothing.
We heard the excuse of “we need to do it without downtime.” Do you really?? I mean, I don’t think anybody would be complaining if WaniKani went offline for ~10 hours one single time, and if you’re smart about it you could even narrow it down to like 3 hours.
You can do the migration and partitioning on a non-prod database (obviously) while the site is still running, and then once you have everything tested and working as expected, put the site offline for a couple hours while you update the non-prod database with the new data that came in while you were working on it, then simply put the non-prod database into prod and put the site back online. You’ll obviously have backups of everything through this whole process, and if anything goes wrong you can simply revert right back to how it was before and try again in a few days. Nobody’s going to complain about 3 hours of downtime - even if it took you like 5 tries in a row to get everything perfect.
You could even have both databases running in unison - the old, original database keeping track of everything reliably to ensure you don’t accidentally lose anything or get tables mixed up (e.g. the assignments table getting out-of-sync if you need to reset the reviews table), while having the new database running right alongside it. Leave it for a few weeks, check frequently to make sure all the data is perfectly matching, and then when you’re confident in the new system, shut off the old one. You might need to run a second server if the one server couldn’t handle the weight of both databases at once, but that’s more than doable and the extra cost would be well worth it.
I’m not saying it’s easy, but I am saying it doesn’t take 4 months. You should be able to fix this in the span of a couple weeks.
If what I said is somehow not possible, then please do explain it to me and educate me. I might be completely off-base here and what I’ve described is not possible or would somehow not solve the issue. If that’s the case, I would like to know, and I would like an explanation as for why that is. But as of now, this looks like a classic case of “trust me it’s impossible” when really you just don’t want to spend the necessary resources (time, energy, money, risk) to make it happen.
6 Likes
JPDB.io is the main one I know of, and its free, just doesnt have the fancy interface.
kamesame.com is superior and free, though could use some improvements, but the developer has a busy personal life.
bunpro.jp
Maru mori
Anki, pain to set up maybe, but endless decks to use and can install mods and UI tweaks, though I dont know much about that as I havent tried myself.
Additionally there is a member on these very forums creating an alternative from scratch, and recently posted a screenshot. Time estimate to completion yet unknown.
1 Like