Some radicals are appearing as Chinese characters for me

It’s only happened for 2 so far… yet I still have no idea why? Anybody have any ideas?


2 Likes

Looks like a font issue. Couldn’t reproduce for either of them, though :frowning: .

1 Like

Yeah it’s really weird… Like 99% of everything else displays correctly, I have Japanese fonts & IME installed on my PC so I have no clue what to do :sweat_smile:

2 Likes

Is this perhaps what’s happening to me as well…? The same radical looks differently on the different pages:


I noticed it this morning, and I had been debating if to post about it… :eyes:

1 Like

Huh, looks like that on my work computer too. I don’t think there are any fonts/IMEs installed.

There was another thread recently about this. If the material in question doesn’t explicitly mark which version of the character to use, then it’s up your browser to pick one to be displayed. Chinese is the default when it has nothing to go on. Installing Japanese as one of your browser languages should make them default to the Kanji variant.

For demonstration, the character on the left is unspecified and the one on the right is explicitly Japanese.

2 Likes

Hmm, I checked my chrome settings and I only have English and Japanese as my languages. I even set Japanese to be displayed as the default langauge and the problem didn’t go away. What’s weird is that with the unspecified and specified examples that you provided, both appear as the Japanese version for me. :thinking:

Doh. This is even simpler. They are Chinese characters. They don’t exist as Kanji or official radicals. There’s no Japanese version, not that I found anyway. For streamlining purposes, WaniKani calls anything it wants to use in mnemonics a “radical” and has an item and lesson for each one. Most of them are just Kanji, so when the page displays it, it’s just a text character for that Kanji. A few are “real” radicals which also have text characters.
Some don’t actually exist as real text anywhere so WaniKani actually needs to use image files instead. This is why the gun and leaf radicals are actually PNG files. The Keisei Semantic-Phonetic script can’t even display them for that reason.
These ones do exist as Chinese text characters, but they’re not used in Japanese. I searched a Kanji database and it said they’re not in the records. (Actual radicals are in the records.) Instead of bothering to mock up a PNG imitation, WaniKani just uses the Chinese character.

TLDR: They appear as Chinese characters because they are Chinese characters.

13 Likes

Same. My work pc has goofy looking kanji. Though i do have an IDE installed there too so iunno.

Ah I see. Makes sense! Thanks for the explanation :slight_smile:

Why do they look like that though? They have the same font as everything else on my usual PC.

I looked into this for a while, and while I don’t really know what I’m doing, here’s some kind of interesting findings and speculation:

WaniKani / Radical / Guard for me on Chrome:
image

WaniKani / Radical / Guard for me on Firefox:
image

Here’s an experiment with HTML on the forum:

河豚 脱臼 兑
河豚 脱臼 兑

for me on Firefox:
image

for me on Chrome:
image

image

豚 has a non-font variant across languages (this is the unihan thing described here), but it seems like 兑 on chrome stands out and gets a different font specifically when trying to be displayed as Japanese.

Also!
If I use developer tools to tweak the language to “zh” instead of"jp" in the radicals page in Chrome, I get the Firefox version.
But if I do the same in Firefox, nothing changes.

So, my conclusion is that because as @Rushian_Asset pointed out, 兑 isn’t used standalone in Japanese, BUT Wanikani specifies “ja” as the language for their radicals (since they teach Japanese after all), in at least Chrome, that causes it to stand out.
Other font differences are NOT noticeable because the language mismatch is not present.

My guess (completely speculating) is that that’s because in Chrome, because of a quirk or a bug, the fallback for “oops this is definitely a Chinese character but I’m supposed to be outputting Japanese” is “okay output it in the default Chinese font,” which in Chrome is more… serify, and thus noticeable.
Whereas in Firefox, either the default is “okay output it in the specified font anyway” OR… the Firefox default Chinese font is just like, the same as the ones used on these pages or something.

I’d be super curious if someone knows better or if that lines up with a more knowledgeable person’s take!
I’d also be very curious to see any existing developer discussion about this very specific edge case.

if (IF) I’m right, Wanikani could fix this extremely minor display issue by specifying the language for radicals affected as Chinese, or replacing them with static images as someone said other radicals already are. But it seems not really worth it to me, and honestly like it could (arguably) maybe be on Chrome to fix.
(and again I’m not at all a web developer and could easily be wrong or have weird options somewhere I forgot about)

(I was also logged in on the Chrome examples and not logged in on the Firefox examples, but I sure hope that has nothing to do with it)

P.S.: @Brave-foot That “saw” radical issue is something different I think, since that one appears to use a png, rather than an actual character. What page is the first image from? I don’t recognize it.

3 Likes

The first image is from the dashboard page (I’m on level 15 right now) :grinning_face_with_smiling_eyes:

I think I can tell which is which.
The PNG ones GLOW bright white, while the others (most of them) are slightly gray (can be caused by being burned or cause of scripts, but they glow!)
Another set, you can spot the two glowy ones:

Ah I see, I think you might be seeing a very minor bug then!

I played around with Chrome’s developer tools a bit:
The image on the left is that area using the character: 巩
The image on the right is that area with the image: https://cdn.wanikani.com/subjects/images/241-saw-small.png
image

I think probably what happened is they made an image that looks more like how that character looks as a component of other characters, because the standalone character itself looks a bit different, and in that particular location they did not use that image and used the character instead - either accidentally or maybe because of some constraint.

@Toyger , what screen is that with the gold? I thought maybe it was from not making png’s for slightly gray variants, but my radicals progress pages all look slightly gray just fine, I’m just not sure how. It doesn’t seem to be using a separate slightly gray image file, so maybe it’s styling them somehow. Anyway, perhaps that styling is not quite right on the page you’re looking at. Is it a dark mode script?

2 Likes

I assume it might be the GoldenBurns script making it show better =)
The gold color is Golden Burn, but I also use a dark mode script. So one or the other, or possibly combination, makes a noticeable difference (and at a glance it seems to be the “not a real kanji” ones, which is kinda fun!)

1 Like