6 ms·

The Forgotten History of Chinese Keyboards

127 points by mfiguiere 2y ago

dang 2y ago
See also "How the quest to type Chinese on a QWERTY keyboard created autocomplete": https://www.technologyreview.com/2024/05/27/1092876/type-chinese-computer-qwerty-keyboard/ https://www.technologyreview.com/2024/05/27/1092876/type-chi... (via https://news.ycombinator.com/item?id=40548356 https://news.ycombinator.com/item?id=40548356, but no comments there)
upon_drumhead 2y ago
Radiolab also did an excellent podcast on the topic https://radiolab.org/podcast/wubi-effect https://radiolab.org/podcast/wubi-effect
- grendelt 2y ago
  That one was so good. I was completely ignorant of the topic before that episode aired.
mjklin 2y ago
No mention of the competing MingKwai typewriter of Lin Yutang, the famous popularizer of Chinese culture to the west. Apparently his prototype suffered an embarrassing failure at an investor meeting and couldn’t get off the ground. But the idea was good. Article here: https://thereader.mitpress.mit.edu/the-uncanny-keyboard/ https://thereader.mitpress.mit.edu/the-uncanny-keyboard/
- omoikane 2y ago
  Lin Yutang had a few patents related to this typewriter, I believe this is the main one: https://patents.google.com/patent/US2613795A/ https://patents.google.com/patent/US2613795A/ The idea of searching for characters via parts is similar to how Cangjie input method selects characters from radicals. I read somewhere that Cangjie input method was indeed inspired by Ming Kwai typewriter, but I can't find the citation for it.
asdasdsddd 2y ago
> like Chairman Mao Zedong, who seemed to equate Chinese modernization with the Romanization of Chinese script One of Mao's better ideas
- alexlur 2y ago
  Thank God it didn’t happen.
  - wolfgangbabad 2y ago
    Vietnamese is relatively OK.
    - acwan93 2y ago
      Relatively. The amount of diacritics on Vietnamese surpasses European languages so text rendering becomes a challenge if a naive developer doesn't test with Vietnamese.
      - numpad0 2y ago
        Is bringing back Chu Nom script going to simplify Vietnamese support on computers by a lot? It's unintelligible to CJK users, but as far as text rendering goes, it seems just simple Kanji/Hanzi.
    - alexlur 2y ago
      Chữ Nôm is a borrowed writing system and not native to Vietnamese, which isn’t even a Sino-Tibetan language to begin with.
      - gumby 2y ago
        Latin is a borrowed writing system not native to English, German, Polish and many others which aren’t even Romance languages to begin with and must resort to di- and trigraphs plus non-Latin characters like J, V, ß, ł or å, among others (not to mention diacritics).
wolfgangbabad 2y ago
[flagged]
_lvbh 2y ago
“Safari can’t open the page because the address is invalid” How strange. More on topic: Considering how inefficient Chinese characters are in general (but especially evident in computing) as one of the few languages where characters have no direct relation to phonetics, I wonder why there hasn’t been an effort to modernize it similar to Hiragana in Japan. Well, considering how Chinese is basically Kanji, why not just adopt Japanese?
- alexlur 2y ago
  > how inefficient Chinese characters are in general (but especially evident in computing) We are not in the 90s anymore. UTF-8 has been around for 32 years now. If you’re working for a system that has no UTF-8 support, you have a much bigger problem to worry about. > characters have no direct relation to phonetics Most characters are phono-semantic where one part of the character is a phonetic hint and the other is a semantic hint. > modernize it similar to Hiragana Hiragana isn’t and wasn’t intended to replace kanji (unless you are from the fringe Kanamozikai). It serves a different grammatical purpose and is complementary to the other two. Kana is useful for an agglutinating language like Japanese, but not Chinese languages.
  - shiomiru 2y ago
    > Kana is useful for an agglutinating language like Japanese, but not Chinese languages. FWIW, the Japanese did develop a kana-based system for Taiwanese during the occupation, but it was an abomination.[1] [1]: https://en.wikipedia.org/wiki/Taiwanese_kana https://en.wikipedia.org/wiki/Taiwanese_kana
  - numpad0 2y ago
    I think one of statements with respect to CJK languages that has to be made more often is that each of the languages has own numerous dialects with dubious mutual intelligibilities, e.g. Tsugaru and Kagoshima dialects against standard Japanese. The phrase "a language is a dialect with an army" often appears in topic of Asian languages, and causing frictions between CJK non-speakers wondering about compatibilities between the three and speakers showing near vile dissents to those questions. While I understand both sides of these sentiments, the situation is not ideal for both sides. IMO, it might be weird to refer to these languages as "Beijing Tokyo Seoul" languages, but doing so occasionally(just occasionally) could create more tangible feel as to why these three seem to exist side by side so utterly disconnected against each others.
tonetegeatinst 2y ago
Asianometry has a good video on this if my memory serves me right.
geraldwhen 2y ago
I find it humorous that 鷹 was described as a difficult character in the article. It’s like 3 radicals and the character for bird. Perhaps it’s difficult to render in tiny Latin alphabet font, but if you have any Japanese or Chinese study under you, you could read and reproduce that nearly instantly on sight.
- surrTurr 2y ago
  i think he meant difficult in the sense that it consists of many strokes, not in the sense how difficult it is to remember. however, one could argue that there are many other, more complex to write, kanji than 鷹
mrweasel 2y ago
It interesting to consider that both Japan and China might have been prevent from ever being first with general purpose computers. ASCII, and other encoding schemes, only needed to make provisions for less than 200 characters, making it possible to implement with the limited storage and memory available to early computers. The shear amount of characters in some languages, like Chinese may have served as a distraction or roadblock for early computers in the those countries.
- _xnmw 2y ago
  Makes you wonder what limitations of our own language and culture are preventing us from inventing certain things?
- raincole 2y ago
  In an alternative history timeline it might be true. In our timeline I highly doubt whether it was the main reason why general purpose computers didn't happen first in China or Japan.
chai2010 2y ago
Very good article, like it. Chinese characters are a type of pictographs that have some characteristics of QR codes. In fact, there is indeed a word retrieval method called four-corner number, which quickly maps Chinese character graphics to four numbers through some simple formulas, which is especially suitable for one-way encoding and retrieval. For example, the four-corner number of "龍" is coded as 0121, and the code of "兲" is 1080 (please refer to https://github.com/chai2010/im4corner https://github.com/chai2010/im4corner). In addition, Chinese characters are actually more important as hieroglyphic shapes. For example, we have a "凹语言" (Wa-lang https://github.com/wa-lang/wa/ https://github.com/wa-lang/wa/ ) designed for WebAssembly (WASM for short, WebAssembly => WASM => Wa), in which the Chinese characters "凹" and WASM The logo is very similar, and there was even a pronunciation of "wa" in the past. After the popularization of computers, the function input method has been greatly improved, but there is still a lot of input resistance. For example, in programming, frequent switching between Chinese character names and English keywords brings a loss of input efficiency. As a programmer, I hope Chinese users can continue to pay attention to and improve these in the future.

127 points by mfiguiere 2y ago

dang 2y ago

See also "How the quest to type Chinese on a QWERTY keyboard created autocomplete": https://www.technologyreview.com/2024/05/27/1092876/type-chinese-computer-qwerty-keyboard/ https://www.technologyreview.com/2024/05/27/1092876/type-chi... (via https://news.ycombinator.com/item?id=40548356 https://news.ycombinator.com/item?id=40548356, but no comments there)

upon_drumhead 2y ago

Radiolab also did an excellent podcast on the topic https://radiolab.org/podcast/wubi-effect https://radiolab.org/podcast/wubi-effect

grendelt 2y ago
That one was so good. I was completely ignorant of the topic before that episode aired.

mjklin 2y ago

No mention of the competing MingKwai typewriter of Lin Yutang, the famous popularizer of Chinese culture to the west. Apparently his prototype suffered an embarrassing failure at an investor meeting and couldn’t get off the ground. But the idea was good. Article here: https://thereader.mitpress.mit.edu/the-uncanny-keyboard/ https://thereader.mitpress.mit.edu/the-uncanny-keyboard/

omoikane 2y ago
Lin Yutang had a few patents related to this typewriter, I believe this is the main one: https://patents.google.com/patent/US2613795A/ https://patents.google.com/patent/US2613795A/ The idea of searching for characters via parts is similar to how Cangjie input method selects characters from radicals. I read somewhere that Cangjie input method was indeed inspired by Ming Kwai typewriter, but I can't find the citation for it.

asdasdsddd 2y ago

> like Chairman Mao Zedong, who seemed to equate Chinese modernization with the Romanization of Chinese script One of Mao's better ideas

alexlur 2y ago
Thank God it didn’t happen.
- wolfgangbabad 2y ago
  Vietnamese is relatively OK.
  - acwan93 2y ago
    Relatively. The amount of diacritics on Vietnamese surpasses European languages so text rendering becomes a challenge if a naive developer doesn't test with Vietnamese.
    - numpad0 2y ago
      Is bringing back Chu Nom script going to simplify Vietnamese support on computers by a lot? It's unintelligible to CJK users, but as far as text rendering goes, it seems just simple Kanji/Hanzi.
  - alexlur 2y ago
    Chữ Nôm is a borrowed writing system and not native to Vietnamese, which isn’t even a Sino-Tibetan language to begin with.
    - gumby 2y ago
      Latin is a borrowed writing system not native to English, German, Polish and many others which aren’t even Romance languages to begin with and must resort to di- and trigraphs plus non-Latin characters like J, V, ß, ł or å, among others (not to mention diacritics).

wolfgangbabad 2y ago

[flagged]

_lvbh 2y ago

“Safari can’t open the page because the address is invalid” How strange. More on topic: Considering how inefficient Chinese characters are in general (but especially evident in computing) as one of the few languages where characters have no direct relation to phonetics, I wonder why there hasn’t been an effort to modernize it similar to Hiragana in Japan. Well, considering how Chinese is basically Kanji, why not just adopt Japanese?

alexlur 2y ago
> how inefficient Chinese characters are in general (but especially evident in computing) We are not in the 90s anymore. UTF-8 has been around for 32 years now. If you’re working for a system that has no UTF-8 support, you have a much bigger problem to worry about. > characters have no direct relation to phonetics Most characters are phono-semantic where one part of the character is a phonetic hint and the other is a semantic hint. > modernize it similar to Hiragana Hiragana isn’t and wasn’t intended to replace kanji (unless you are from the fringe Kanamozikai). It serves a different grammatical purpose and is complementary to the other two. Kana is useful for an agglutinating language like Japanese, but not Chinese languages.
- shiomiru 2y ago
  > Kana is useful for an agglutinating language like Japanese, but not Chinese languages. FWIW, the Japanese did develop a kana-based system for Taiwanese during the occupation, but it was an abomination.[1] [1]: https://en.wikipedia.org/wiki/Taiwanese_kana https://en.wikipedia.org/wiki/Taiwanese_kana
- numpad0 2y ago
  I think one of statements with respect to CJK languages that has to be made more often is that each of the languages has own numerous dialects with dubious mutual intelligibilities, e.g. Tsugaru and Kagoshima dialects against standard Japanese. The phrase "a language is a dialect with an army" often appears in topic of Asian languages, and causing frictions between CJK non-speakers wondering about compatibilities between the three and speakers showing near vile dissents to those questions. While I understand both sides of these sentiments, the situation is not ideal for both sides. IMO, it might be weird to refer to these languages as "Beijing Tokyo Seoul" languages, but doing so occasionally(just occasionally) could create more tangible feel as to why these three seem to exist side by side so utterly disconnected against each others.

tonetegeatinst 2y ago

Asianometry has a good video on this if my memory serves me right.

geraldwhen 2y ago

I find it humorous that 鷹 was described as a difficult character in the article. It’s like 3 radicals and the character for bird. Perhaps it’s difficult to render in tiny Latin alphabet font, but if you have any Japanese or Chinese study under you, you could read and reproduce that nearly instantly on sight.

surrTurr 2y ago
i think he meant difficult in the sense that it consists of many strokes, not in the sense how difficult it is to remember. however, one could argue that there are many other, more complex to write, kanji than 鷹

mrweasel 2y ago

It interesting to consider that both Japan and China might have been prevent from ever being first with general purpose computers. ASCII, and other encoding schemes, only needed to make provisions for less than 200 characters, making it possible to implement with the limited storage and memory available to early computers. The shear amount of characters in some languages, like Chinese may have served as a distraction or roadblock for early computers in the those countries.

_xnmw 2y ago
Makes you wonder what limitations of our own language and culture are preventing us from inventing certain things?
raincole 2y ago
In an alternative history timeline it might be true. In our timeline I highly doubt whether it was the main reason why general purpose computers didn't happen first in China or Japan.

chai2010 2y ago

Very good article, like it. Chinese characters are a type of pictographs that have some characteristics of QR codes. In fact, there is indeed a word retrieval method called four-corner number, which quickly maps Chinese character graphics to four numbers through some simple formulas, which is especially suitable for one-way encoding and retrieval. For example, the four-corner number of "龍" is coded as 0121, and the code of "兲" is 1080 (please refer to https://github.com/chai2010/im4corner https://github.com/chai2010/im4corner). In addition, Chinese characters are actually more important as hieroglyphic shapes. For example, we have a "凹语言" (Wa-lang https://github.com/wa-lang/wa/ https://github.com/wa-lang/wa/ ) designed for WebAssembly (WASM for short, WebAssembly => WASM => Wa), in which the Chinese characters "凹" and WASM The logo is very similar, and there was even a pronunciation of "wa" in the past. After the popularization of computers, the function input method has been greatly improved, but there is still a lot of input resistance. For example, in programming, frequent switching between Chinese character names and English keywords brings a loss of input efficiency. As a programmer, I hope Chinese users can continue to pay attention to and improve these in the future.