Hacker News new | ask | show | jobs
by feiri 1285 days ago
汉字 reader here! (although I'm probably only at grade 3 level)

I've never seen any web-based word games like this for Chinese, so this is really cool to see!

In my opinion, the core game-part itself feels more like a matching game than a language-learning game. Once any hints are revealed, there's usually only two remaining characters that match the current stroke pattern/positioning. This makes it pretty easy to guess the correct characters starting off with a single random click. One way you could make this more challenging is to only reveal the matching stroke elements, but not necessarily their physical position in the answer word.

While playing, I was confused by how strokes, radicals, and components got differentiated. One specific instance of confusion was that the ren2 radical (亻) doesn't get revealed when guessing any character that contains a shu4 (丨).

I've noticed that in games that include characters with shu4 and heng2 strokes, guessing any character with those two strokes often reveals a large chunk of the answer. Maybe that's a strategy that when realized, actually helps people learn? I don't know. In any case, here is a recent game I played: https://git.sr.ht/~yfxu/images/blob/master/hanzi.webp

That said, being able to see the full word at the end with the pinyin and definition would be pretty useful for a learner (I learned a new word!), but I don't know if it would be nearly as useful as flash cards would be.

At the end of the day though, I still think this is a cool game, and I bet figuring out how to hide and reveal specific components of the characters wasn't the easiest task aha, thanks for making and sharing this!

2 comments

Japanese 漢字 reader here. I don't know how Chinese traditionally splits characters in parts, but I had really weird matchings. Like in the case of 日常 as the guessed word, the character I chose first was 剪, and that matched everything except the three strokes at the top of 常, and I have no idea why.In 剪, it highlighted the horizontal stroke at the top of 前, as well as 月 and the leftmost stroke of 刂. I guess the 日 comes from 月, which seems like a stretch, but the rest is a mystery.

Overall, as someone without knowledge of Chinese characters except the ones that exist in Japanese, I found it way too easy.

Interesting concept, though.

Thanks for the feedback. I've updated the game to exclude individual strokes and single-stroke radicals from the character similarity search. I thought it would be interesting to drill all the way down to individual strokes, but it more often than not causes confusion and reveals too much of the target characters.
Hey, thanks so much for the thorough review! I'll definitely put some thought into making the game harder and less down to chance, maybe with different difficulty levels.

On ren2 vs shu4, there are some characters/components that have decompositions missing in the Make Me a Hanzi database unfortunately. I think this may be one of those cases, or it's a bug. Will look into it. Thanks again!