This is gorgeous, and it works beautifully on a touchscreen. Incredibly valuable tool for learning mandarin, I bet. It even works with traditional characters!
How did you do this??? Was there a database with stroke orders somewhere?
Love this app. It would be wonderful if this could be included in a spaced repition app somehow!
>> Incredibly valuable tool for learning mandarin, I bet.
If someone were actually trying to Mandadin, Prof Victor Mair at Penn tells people to mostly ignore the characters for the first few years.
Victor Mair, at Penn, remarks (in an e-mail I quote with permission): “It’s a tragedy that so many young Americans spend years stuffing their heads with hundreds of Chinese characters, gaining no usable proficiency, and then forgetting them all by the time they’re 25.” Mair thinks that “if the Chinese would wake up and permit pinyin to function as part of a genuine digraphia, then I would say it might make sense for maybe 2 percent of the population to learn up to third-year level of Mandarin—strictly romanized, mind you. But there are exceedingly few teachers who are enlightened enough to teach it that way.”
> It’s a tragedy that so many young Americans spend years stuffing their heads with hundreds of Chinese characters, gaining no usable proficiency, and then forgetting them all by the time they’re 25.
There are similar complaints about all kinds of languages, even when they're written in Latin script. (The proverbial "high school French" comes to mind.) I think Mair is mistaken about the characters being the culprit. Most people taking a language class are simply not invested enough to actually practice, and then they forget everything they did learn once they stop using the language.
While it's true that it would be unnecessary to learn the characters if Chinese were usually written in Pinyin, that's unfortunately not the case and so you'd better memorize them if you want to learn the language without remaining an analphabet.
It's also not like you need to know all that many characters right off the bat. The most common 300 make up about 80% of every text; you can get to that level in a month if you learn 10 new characters per day. Most of the effort is in learning words formed by combining those characters, just like learning any other language.
Korea's hangul is pretty neat. As the linguist Geoffrey Pullum says, English has the spelling system that positively stinks, but the Chinese/Japanese writing system is a whole nother level.
Learning to recognize the characters is super important. It unlocks so many fun native resources to use as study material. Memorizing how to write characters is a waste for most learners who aren't really invested in it.
That aside, as it currently stands, yes, if someone were to learn Chinese/Japanese long term, yes, learning to recognize a few hundred characters helps in understanding a lot of native materials.
I'd say learning how to write those characters is mostly a waste for anyone other than historians, although kids are forced to learn it in school (I was one of them).
> I’d say learning how to write those characters is mostly a waste for anyone other than historians
You are totally wrong here: this is a useful skill for anyone living there because, like it or not, this is how people are actually writing their languages. Also have a look at any scientific paper written in Korean and you will see how comical it is: every conceptual word is follow by its writing in Chinese characters in parenthesis because there is so many homophones for all the loan words coming from Chinese that using a phonemic script makes the text un-understandable.
You must have misunderstood, when I said "write", I meant handwriting, with stroke order and all. There's no need to learn to write this way, at all, for almost anyone.
Oh I see. But then, handwriting taken appart, stroke order does matter and is a feature for more easy remembering. One actually only have to rote learn 500 characters to get a feeling of the system and then handwriting get optional. Anyway, how would you recommend to learn the characters instead?
Like Prof Victor Mair says, if someone is a beginner, I don't recommend learning characters at all for the first 2-3 years.
(If you are a Japanese learner, learning kana wouldn't hurt, it takes a few weeks at most anyways vs years for kanji - and it's not like you "master" kanji after those years, as any native speaker would be able to demonstrate).
2-3 years is an awfully long time to spend learning a language without being able to read, though. And if you plan to learn it later anyway, you might as well start with it from the beginning.
Yeah I should've said "I don't recommend learning how to write the characters by hand in the first 2-3 years, or ever".
As gibolt said, it's much easier to just learn to recognize characters, without being able to write them by hand.
If you learn Japanese for, say, 2 years, even without consciously trying to learn the characters (e.g. by using rikai-chan/kun) when reading articles, you'll encounter the same words over and over so you'll be able to recognize them, even if you can't write them by hand. And that's good enough.
Most native Japanese speakers cannot write words like 咀嚼 or 憂鬱 by hand, but many more can pronounce them just fine.
Pr Mair have his opinion on the matter, I have mine. And while he his famous in his field, given I have formal education in Japanese at the MA level, some self taught Chinese proficiency (mostly reading) and linguistic knowledge (mostly phonology) about both these languages and some of their older varieties, I would not blindly follow his advice that I think is misleading.
That doesn't mean Chinese education is perfect right now, and I also think that traditional characters, while seeming more complex can be better for beginners as the system has a more regular internal consistency. And it's easier to learn simplified when knowing traditional than the reverse.
Yeah, I'd recommend actually learning the language (vocab, pronunciation etc), rather than trying to learn the characters and trying to improve the retention of them.
With Chinese, I felt like sunk a lot of time into memorizing characters for the illusion of progress without actually learning how to engage with the language.
On the other hand, with Korean, I spent some time up front learning to type, and that paid dividends as I was able to text with people and read maps before long.
Maybe if I had learned characters in Chinese not with the goal of being a master calligrapher, but with the goal of communicating with other people using today's tools, they would have had similar benefits.
I learned Chinese in 2001 and after that I haven't had much practice speaking it but I think that knowing the characters has actually kept me from forgetting it completely. It's nice to have a conversational partner but lacking that, practicing characters and reading is a way to practice the language without having a partner. It takes some effort but if you live somewhere where you have Chinese characters around (e.g. LA San Gabriel Valley, or many Chinese restaurants) or use some Chinese social media (e.g. WeChat) you usually have some examples that you can plug into Google Translate (by swiping the character shape--which isn't strict about stroke order), then plug that into mdbg.net for decent dictionary type translations and concordances. It is a good way to spend time waiting for Chinese food at a restaurant. If you share your activity with your table it's not as anti-social as run-of-the-mill playing with your phone.
Fun fact about stroke orders, many people regard Chinese ideograph as some sort of bitmap, but if you watch calligraphy artists[1] you will discover they are not only some straight lines, but the the speed and acceleration also varies when drawing different part of the line. It's an important factor for distinguish writing styles.
So far we have character as bitmaps, as vector graphics, and vector graphics with stroke order, but we've yet to see a db about the composition speed of the strokes.
Also tilt and pressure. But I didn't notice rotation in the sections I skimmed.
With XR coming to unshackle visual output from fixed flat small screens, the question arises: how might we unshackle input? Say one can type while holding a single chopstick. What might one do with 7DOF fingers and brush in 2D and 3D? If a laptop keyboard was also a touch surface, and the space above it were tracked? What does an IDE look like for typing and painting code, where the "paint" is alive? So I'm interested in stylus and brush movement vocabularies, and gesture recognition.
animCJK is great, but it only covers the most commonly used Kanji and Hanzi. For example, it doesn't have 鱚. If you need a database that also has hyōgai kanji, KanjiVG [1] is a better choice. (鱚 can be seen here [2]).
There's also the Kanimaji project [3] to animate files in the KanjiVG format, which could be used to get results similar to the Hanzi Writer project. If you use it to produce GIFs rather than animated SVGs, the resulting files are somewhat large though, so I wrote my own (ugly) renderer [4] to get better compression.
34 comments
[ 2.5 ms ] story [ 73.4 ms ] threadHow did you do this??? Was there a database with stroke orders somewhere?
Love this app. It would be wonderful if this could be included in a spaced repition app somehow!
Thanks for sharing this.
If someone were actually trying to Mandadin, Prof Victor Mair at Penn tells people to mostly ignore the characters for the first few years.
Victor Mair, at Penn, remarks (in an e-mail I quote with permission): “It’s a tragedy that so many young Americans spend years stuffing their heads with hundreds of Chinese characters, gaining no usable proficiency, and then forgetting them all by the time they’re 25.” Mair thinks that “if the Chinese would wake up and permit pinyin to function as part of a genuine digraphia, then I would say it might make sense for maybe 2 percent of the population to learn up to third-year level of Mandarin—strictly romanized, mind you. But there are exceedingly few teachers who are enlightened enough to teach it that way.”
https://www.chronicle.com/blogs/linguafranca/2012/03/02/more...
There are similar complaints about all kinds of languages, even when they're written in Latin script. (The proverbial "high school French" comes to mind.) I think Mair is mistaken about the characters being the culprit. Most people taking a language class are simply not invested enough to actually practice, and then they forget everything they did learn once they stop using the language.
While it's true that it would be unnecessary to learn the characters if Chinese were usually written in Pinyin, that's unfortunately not the case and so you'd better memorize them if you want to learn the language without remaining an analphabet.
It's also not like you need to know all that many characters right off the bat. The most common 300 make up about 80% of every text; you can get to that level in a month if you learn 10 new characters per day. Most of the effort is in learning words formed by combining those characters, just like learning any other language.
Chinese and Japanese are definitely held back by their awful script + many people who confuse language with script.
http://thepenngazette.com/characters-in-search-of-writers/
That aside, as it currently stands, yes, if someone were to learn Chinese/Japanese long term, yes, learning to recognize a few hundred characters helps in understanding a lot of native materials.
I'd say learning how to write those characters is mostly a waste for anyone other than historians, although kids are forced to learn it in school (I was one of them).
You are totally wrong here: this is a useful skill for anyone living there because, like it or not, this is how people are actually writing their languages. Also have a look at any scientific paper written in Korean and you will see how comical it is: every conceptual word is follow by its writing in Chinese characters in parenthesis because there is so many homophones for all the loan words coming from Chinese that using a phonemic script makes the text un-understandable.
(If you are a Japanese learner, learning kana wouldn't hurt, it takes a few weeks at most anyways vs years for kanji - and it's not like you "master" kanji after those years, as any native speaker would be able to demonstrate).
As gibolt said, it's much easier to just learn to recognize characters, without being able to write them by hand.
If you learn Japanese for, say, 2 years, even without consciously trying to learn the characters (e.g. by using rikai-chan/kun) when reading articles, you'll encounter the same words over and over so you'll be able to recognize them, even if you can't write them by hand. And that's good enough.
Most native Japanese speakers cannot write words like 咀嚼 or 憂鬱 by hand, but many more can pronounce them just fine.
As a person who learned and teaches it, don't skip the characters.
That doesn't mean Chinese education is perfect right now, and I also think that traditional characters, while seeming more complex can be better for beginners as the system has a more regular internal consistency. And it's easier to learn simplified when knowing traditional than the reverse.
They each have a specific order which makes the radicals and similarities and relationships between different characters clear.
On the other hand, with Korean, I spent some time up front learning to type, and that paid dividends as I was able to text with people and read maps before long.
Maybe if I had learned characters in Chinese not with the goal of being a master calligrapher, but with the goal of communicating with other people using today's tools, they would have had similar benefits.
So far we have character as bitmaps, as vector graphics, and vector graphics with stroke order, but we've yet to see a db about the composition speed of the strokes.
[1] https://youtu.be/8-99YB4iou0?t=45 as an example. You can find plenty.
Making pixel area fill rate constant is probably another decent approximation.
Also tilt and pressure. But I didn't notice rotation in the sections I skimmed.
With XR coming to unshackle visual output from fixed flat small screens, the question arises: how might we unshackle input? Say one can type while holding a single chopstick. What might one do with 7DOF fingers and brush in 2D and 3D? If a laptop keyboard was also a touch surface, and the space above it were tracked? What does an IDE look like for typing and painting code, where the "paint" is alive? So I'm interested in stylus and brush movement vocabularies, and gesture recognition.
Like 必.
[0]: https://github.com/parsimonhi/animCJK
[1]: http://gooo.free.fr/animCJK/official/
There's also the Kanimaji project [3] to animate files in the KanjiVG format, which could be used to get results similar to the Hanzi Writer project. If you use it to produce GIFs rather than animated SVGs, the resulting files are somewhat large though, so I wrote my own (ugly) renderer [4] to get better compression.
[1] https://github.com/KanjiVG/kanjivg
[2] https://kanjivg.tagaini.net/viewer.html?kanji=%E9%B1%9A
[3] https://github.com/aehlke/kanimaji
[4] https://github.com/Yorwba/kanjivg-gif
So I think this tool is useful for both.