Can vocab just reject readings put in the meaning field instead of counting them wrong?

WaniKani Feedback
#46

In the midst of all this yelling about the moral character of people who do or don’t want this feature, can I just state the obvious here?

ANY knowledge testing system is going to have some (seemingly arbitrary) rules built into it, and as a result any such system will also be testing your ability to follow the rules (as well as your actual knowledge).

Teachers (human or computer) can’t read minds[1]. We can only monitor your learning based on what you say/write and only if you give us useable input[2]. I’ve seen so many students flub exams because they didn’t read the instructions, didn’t answer what was being asked, and when they say “but I KNEW it!” all I can say is “but you didn’t SHOW me you knew it!”

So yes if it’s doable program-wise, I’d love to see WK give an extra warning wiggle (it already does that if you put kana in the meaning field or romaji in the reading field).

But let’s not accuse people of being sloppy personalities or robot sycophants cuz at the end of the day there will always be some element of arbitrariness that the learner will have to pay attention to and follow. That’s the price we pay for not being telepathic robots[3].

[1] This is a shocking and close-kept secret! I may get into big trouble for telling you about it!

[2] Especially for computers. Human teachers can at least try to figure things out from context when students start writing about the “Japanese Candy Pictographic System” or whatever.

[3] Most of us, anyway. I mean, I don’t want to make assumptions about all of you.

7 Likes
#47

Honestly? Just get it. You will not regret it. You will eventually just zone out again, and again, no matter how well you think you are paying attention, and it’s just way too annoying to get it wrong for no reason.

Even if Vocabulary doesn’t stop you from leveling up, you don’t necessarily want 1000 reviews because you can’t get simple stuff out of apprenticeship.

1 Like
#48

What I don’t understand is why some people feel the need to give the OP unsolicited comments about his study methods. This is a feature request thread and he requested a feature, any other comment on how the OP studies or the supposed faults in his study methods are not only rude, but simply off topic.

6 Likes
#49

I guess I’ll reply. I was going to let it be, but apparently my reply offended people so much that it got removed (probably a sign that I’m onto something). I hope you weren’t one of the people who flagged it.

I think my comment needs to be taken in context. It is a reply to someone who makes a great argument for why the feature listed would be beneficial for actually learning to read.

Many of the replies here (and also the ones I got when I made a similar argument elsewhere) are of the form: “just learn to see the colors” or “you got it wrong so it should be marked wrong, suck it up and get better,” “be more careful and proofread,” etc.

(Obviously, those aren’t exact quotes, but they capture the sentiment).

The community is basically defending the status quo with these comments, and they are telling the person to get better at WaniKani rather than think about whether the feature would actually benefit someone who is merely using WaniKani to get better at reading.

I see no reason why people should have been offended at my comment. By pointing out this difference in mindset, I was able to make sense of why people were responding the way they were. Several people liked my comment, so others clearly thought it made sense.

If anyone who flagged my comment sees this, please consider why you were so offended.

5 Likes
#50

I was not, and I’m surprised it was flagged enough to get hidden. What you said was a little too pointed, which is why I responded as directly as I did in return. But I don’t think it should have been flagged.

They did make a great argument. That doesn’t necessarily mean the feature is worth implementing, depending on what percentage of users might actually benefit from it. But I do think it made a great distinction between formal testing and learning processes.

Keep in mind that many users who have expressed confusion between the reading and meaning questions had never actually noticed the color cues. In those cases, (some of) the responses you’re complaining about are actually helpful. Even some people when told to be more careful basically respond with “sigh, I guess you’re right”. The problem of course is that when there are dozens of voices saying the same thing, it can come off as dismissive (or even attacking) instead of helpful.

2 Likes
#51

I completely agree with every single word you said, and I find extremely sad that many people flagged your comment as ‘offensive’, but it also proves you were and are absolutely right. P

2 Likes
#52

Being flagged proves nothing one way or the other. If you say bigoted statements that would also get flagged. That doesn’t make the bigoted statements true. (That obviously wasn’t the case here, but your assertion has no basis.)

7 Likes
#53

My response telling someone they’re on a high horse was also flagged

4 Likes
#54

Several posts from this thread:

Though it should be noted that

3 Likes
#55

Wow, the discussion in this thread is around 700% more heated than necessary.

I personally don’t remember ever having this problem - when I started with WaniKani, I didn’t know anything and did my reviews really slow and carful. By the time I managed to go faster, I was already trained to differentiate reading/meaning questions quickly.

However, I also don’t think that counting an answer that is a correct reading, when asked for a meaning, as wrong has any benefits for learning Kanji. Additionally, I was interested in learning how to write WaniKani userscripts, so here is my attempt at implementing the requested behavior (input reading when asked for meaning → shake + try again): The Meaning of 六 is roku¹

The script is mostly based on Close But No Cigar by @Ethan. It should work in lessons and reviews, but I only tested it with a few reviews.

General Script Installation Instructions

Known conflict:
If you’re also using WaniKani Mistake Delay, make sure that “The Meaning of 六 is roku” has a lower number in Tampermonkey (higher up in the list when sorted by number).
¹ I already know that tomorrow I will regret choosing such a silly name, but now, at 3:45 am, it seems like a good idea

EDIT: Wait - did I misinterpret the excessive flagging in this thread? Was it all part of “The Competition”? :stuck_out_tongue:

6 Likes
Is this new?
#56

If anyone (with scripting experience) is interested in adding a feature to [Double-Check] to catch reading-instead-of-meaning errors, I’d be happy to merge the changes and post an update. I would implement the changes myself if I had time.

The check for reading-instead-of-meaning (or visa versa) would go in the 'first_submit' section of the script. I think you’d just need to call the following line twice:

answer = old_answer_checker(qtype, $("#user-response").val());

And just change the qtype (i.e. question type) in between the two calls, and see if the answer comes back as correct after changing the qtype. Then add a new message to the set_answer_state() function to tell the user what happened.

3 Likes
#57

Lots of useful replies here. The way I fixed the problem for myself for me is this:
image

Courtesy of

3 Likes
#58

I have added this feature to the Double-Check script, but how should I send you the code?

I guess I just post it here and delete it after you tell me that you have accepted/rejected it.

deleted

In the end I did not use the old_answer_checker(), because it changes the displayed answer to the attempted kana-fication of the input.

Also, I have only implemented one way: reading instead of meaning. The other direction would probably be more involved. I don’t know if the original input is stored anywhere before wanakana converts it to kana. And the converted input cannot be converted back to the latin alphabet (was く created by “ku” or “cu”?).

I’m wondering if for meaning-instead-reading leniency it would be reasonable to convert every accepted meaning to kana and compare the input against those…

2 Likes
#59

I can still see it. You’re talking about this one, right? :arrow_down:

Not relevant to the OP but that comment really stuck in my head. I’m trying to relearn a lot of content after a long break so my apprentice pile has exploded and as a result reviews get out of control really fast. Every so often I find myself strategizing about how many items I need to upgrade to Guru or what script I can use to pin-point the items that have the greatest “weight” on the review count so I can get a manageable load and then I go “whoa, hold on, you’re not here to become an expert WaniKani strategist! It’s not about optimizing the pink number versus the purple number, it’s about learning the darn things so you can read Japanese.” So thanks for the reality check. :+1:

6 Likes
#60

I have this saved. It will be a while before I can merge it, though.

Thanks for looking into it, by the way! I had implemented something similar in my Self-Study Quiz, but hadn’t thought of putting it in Double-Check.

2 Likes
#61

I don’t know why you had to delete it. All scripts that are installed with Tampermonkey/Violentmonkey etc. are readable by anyone installing them anyway. Why the need for secrecy?

#62

I just thought that the box with the code takes too much space, and the code is irrelevant to this topic. It would not be secret anyway because it is still visible in the edit history of my post.

#63

If I’m reading your Self-Study Quiz code correctly, it seems that you’ve used for meaning-instead-reading mistakes the same method that I came up with.

I’ve experimented with it today, but I was not sure if I should even post this, because it only works if the meaning that the user has written is an exact match. So if the meaning is “eyebrow” and the user has written “えいぇbろws” (“eyebrows”), then it fails to recognize the reading/meaning mixup (okay, in this specific example it is not that much of a problem, because the latin characters in the response result in a shake anyway).

Here is this new version with reading-instead-meaning as well as meaning-instead-reading:

deleted

Sorry for spamming code boxes into this topic ^^

2 Likes
[Userscript]: Double-Check (Version 2.x)
#64

I’ve updated my local copy, and will review when I get time.

3 Likes
#65

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.