Jump to content
Learn Chinese in China

Journey with Cangjie

  • entries
  • comments
  • views

About this blog

On trying to learn Chinese Cangjie input method for personal computers.

Entries in this blog


Inputting 2-Unit Characters

Two-unit characters are characters that can be divided into two units based on the shape of a character. This can be a division based on left-to-right, top-to-bottom, or outer-to-inner. If you are familiar with radicals then the concept is not so alien. For left-to-right often a spacing, or lack of connecting from the left side of the character to the right side of the character, is where the the character is divided up into two units. For top-to-bottom it can be more tricky, but again if you think back to character radicals it is not a leap. For outer to inner, there is an outer shell unit that encompasses the inner unit.


Once the character is divided up into two units, the following rules apply for determining the Cangjie input code:

1) The first unit may only have up to two Cangjie symbols input. If there are more than two symbols in the unit, then only the first and last Cangjie symbols are input on the keyboard.

2) The second unit may have up to three Cangjie symbols input. Again, if there are more than three, some skipping is involved. In this case, the first, second and last Cangjie symbols are input on the keyboard. 


This may seem a bit abstruse, so let's look at some examples. Unfortunately I am on my android device right now and I don't yet have a very good Cangjie input method tool, so I'm still looking for a better way to bring up just the Cangjie symbol to show how things are built up. As such, to make sense of the following I suppose you need some familiarity with Cangjie symbols. I will try to update this post later if I can figure it out.


1) 風 / 风    


A quick visual examination yields a clear outer-to-inner relationship in both the traditional form and simplified form of the character. The outer unit is 几, which can be made up using Cangjie codes that look like 厂乙 (note these are radicals on don't reflect the exact Cangjie symbols, but I wanted to provide something to help see how the unit is broken down to Cangejie symbols. To build the unit, the input code is HN for the outer unit. The inner unit looks like 虫 with a "hat" on top. This unit requires four symbols, so we have to skip the third, with the Cangjie code being HLI. The resulting Cangjie code for the entire character is HNHLI.


For the simplified character it is similar. The first unit is HN still, but the innter unit is simplified. It turns out that this unit actually represents a Cangjie symbol, so the Cangjie code for this unit is just K. Putting it all together, the resulting code for the entire character is HNK.


2) 鍾 / 锺


Upon visual inspect, this character is a left-to-right two-unit character. The first unit is 金 or钅, and the second unit is 重. Well, 金 is actually a Cangjie symbol, so for both the traditional and simplified forms, the input code for the first unit is C, and that completes the unit. The second unit, however, has four Cangjie symbols in it, so we have to skip the third since we are only allowed up to three symbols. The unit is made of a "hat" stroke on the top, 十 just below it, then skipping 田 because it is the third symbol, and finally 土. This turns out to HJG, and thus the entire code for this character is CHJG.


3) 規 / 规


Again, this is a left-to-right two-unit character. Based on the Cangjie symbols, the left side is built with the codes QO. For the right side, the code comes out to BUU for the tradtional. For the simplified, it is BHU. Again, I wish I could input just these Cangjie symbols for reference, but this one is proving challenging for my Android input method. The final code is QOBUU for the tradtional version,a nd QOBHU for the simplified version.


Well, that is it for today, looking back at this post I am not even sure if it is useful. But oh well, this is kind of my journal too. When I am back on a real computer I will see what I can do.


Progress is Slow

I have not abandoned the project...I am still struggling with memorizing the secondary symbols. I have most of them down, but I struggle with the secondary symbols that fall under: 心,尸,and 山. Especially 心. I suppose this is because so many of them seem like a stretch (in my humble opinion). But mostly it is my own fault. I am terrible about remembering to do my Anki decks daily. Suppose part of that is because I hate flash cards to begin with. Seems like I can only remember on the weekends when my head doesn't get full with stuff from work.


I found a really great app for a smartphone (iPhone and Android) for practicing Cangjie recognition and the keyboard. It is called 五色學倉頡 by Kotech (www.koketch.hk). It has a practice mode where you can pick three different levels: easy, intermediate, and difficult. It presents you with a Chinese character that is split into components by color. It also provides a Cangjie keyboard. The goal is to try to correctly input the character without mistakes. I much prefer this to reviewing my Anki deck.


So I've set a goal to memorize those secondary symbols by the end of the month, or.....well, frankly I'll just keep going at it, I'm stubborn. But I want to try to reach this goal by the end of the month in all seriousness. Perhaps I'll try positive reinforcement--if I achieve the goal, maybe then I'll treat myself to a nice lunch?


Cangjie Status Update

The past two weeks have been slow and steady with respect to my Cangjie progress.  While it helps to work through practice exercises physically at the keyboard, I am finding that the best way to remember secondary symbols is to just use Anki and work through the memorization using a brute-force methodology.


Interestingly, I find myself double-guessing sometimes, particularly with secondary signs that fall under 尸 or 心 primary symbol keys. When I look at the secondary symbol very carefully, and then look at the primary symbol, I eventually see the reason for its categorization. It is interesting how the brain can be drawn to certain areas first, and if one doesn't look carefully, one miss the fine details.


I think in a couple more days I should be in good shape to move onto the third part of learning, which is learning how to decompose more complex characters into primary and second symbols. Perhaps this will be the hardest aspect, and I suspect it will require making lots of mistakes, checking the actual Cangjie code, and then doing it all over again. The first two steps have been finger "muscle" memory, but this third part is going to be more about training the eye.


On a more positive note, I did get a character input into a dictionary tool without knowing the pronunciation the other day. It was a great feeling, and makes all of this work more worthwhile in my mind.


One week with Cangjie Input (倉頡輸入法)

I decided to learn Cangjie Input Method recently. I often find myself in the situation where I need to look up the meaning and pronunciation of a Chinese word or character. But how can I type it into a dictionary using pinyin if I don't know the pronunciation? I usually find myself using a stylus on my tablet to draw the characters, or even using the mouse on a PC to draw the characters on screen. Worst case, I look up each character by selecting radicals. Having to draw or look-up the character slows me down so much when I'm really trying to focus on reading a passage, and I get less time to study and find I am spending a lot of time trying to input characters.

When I lived in Japan I once witnessed a typist inputing characters into an old type-setting machine. The typist was reading hand-written manuscripts and converting these manuscripts to a digital format. The typist could input any character with only a few keystrokes, it was amazing. The typist was so proficient with this system that she found using a phonetic-style IME on Microsoft Windows to be very frustrating. She complained of having to type and then select a choice from a list, where as on the type-setting machine she could look at a character and quickly get it on screen. Also, if she didn't know a difficult character, she could still type it without having to know the pronunciation for inputting the character. Of course she had to learn this typing system and I am sure that it took time to master.


Is it worth it to learn a non-phonetic based input system?  Perhaps. I learned to touch-type when I was 15 and it changed my life. It took time and work, but it has paid off in many ways and was well worth the time investment. Perhaps Cangjie may offer a similar experience?  Before committing to Cangjie, I looked into other systems, such as Wubi, but it appears that Wubi is only for simplified character sets (as I understand it).  As primarily a student of Traditional characters, I wanted to learn a system that could do Traditional characters at a minimum. I also thought about a hand-writing based input system, which would be good for remembering how to write by hand, but I still think a keyboard is more practical.


The first week was good overall. There were some surprises to begin with. I'm working mostly from a Linux laptop with the Ibus input system. Unfortunately, the Cangjie plug-in for Ibus only supports characters used in Hong Kong, and as such it is not possible to input any simplified characters. Other Cangjie implementations allow this from what I have read, such as Twinbridge (which doesn't even appear to be in active development and is Windows only). I'll have to try out Cangjie on Mac OS X next to see how that goes. While this is not a show stopper for me, it is a bit inconvenient. I'll still have to use pinyin for when I need to type a Simplified character.


Learning Cangjie will clearly take time though. I memorized the 24 primary/main signs and practised typing some characters and words that can be input with a single or combination of the primary/main signs. Note: signs are like radicals...but not the same thing. Signs are character components that make up a complete character.  From this point forward it is going to get hard though. The next step is to learn the secondary signs (80+), and then learn the rules for identifying the proper order and identification of the signs.


I hope to have more progress to report next week.

I tried out the Cangjie input system that comes with MacOS El Capitan, which can be found under the Traditional Chinese language options. I was curious if it was possible to type in simplified characters using Cangjie on MacOS. As I wrote about in my first post, IBus in Linux only supports characters used in Hong Kong, and none of the Cangjie codes for simplified characters were working with IBus. I just wanted to report the good news that things work just fine in MacOS!!
車 can be input with the symbols 十田十 (jwj), and 车 can be input with 十手 (jq).
Very cool! I should also mention that the MacOS has a nice system for figuring out the Cangjie symbols and pinyin for a given character. I have been away from the MacOS platform for about five years and just recently returned, and I don't recall this from previous versions of MacOS.  Using the "Find Input Code" option, in the IME pulldown menu, you simply paste the character(s) into the textfield, and it gives the pinyin with tone, zhuyin with tone, Cangjie code by symbol, and two other fields I am not familiar with (stroke and structural). For example, for 开车, see the attached image.

Below are some examples of using Cangjie input with only the 24 primary / main symbols. Primary symbols are the symbols that are represented with a single key press. On a Cangjie keyboard, these would be the symbols you see etched on the keys along with the ASCII character. The first column is the target character, and the second column is the combination of primary symbols used to type the character (the actual key presses are listed in parenthesis).
量 日一田土  (amwg)
或 戈口一  (irm)
車 十田十 (jwj)
王 一土 (mg)
下 一卜 (my)
味 口十木 (rjd)
某 廿一木 (tmd)
出 山山 (uu)
A character should be able to be input into a computer with Cangjie with no more than 5 key strokes. Some input codes are just one character (e.g. 金 is just 'c'), while most input codes are 3 to 4 key strokes. I am finding Cangjie is much like learning to touch type. You have to develop the muscle memory in your fingers and your mind to map the primary symbol to the proper key on the keyboard. As such, I created an Anki deck for the primary symbols and work through them both front side and back side. I also spend some time each day typing some characters so that I get the muscle memory in my fingers.
The 80+ secondary symbols are required for more complicated characters, and then there are some core rules for determining which symbols are necessary to input a character. These rules have to be mastered because there are some characters that have 6 or even 7 symbols within the character, but only at most 5 are required to input the character. The secondary symbols closely resemble radicals from a dictionary, or even some of the symbols on the Wubi input method keyboard layout. There secondary symbol shapes in some way resemble the shape / form of the primary symbol.
Tonight I am starting on the secondary symbols which will probably be time consuming and take a few weeks to really master. I found an interesting iPhone application (五色學倉頡) that allows you to practice Cangjie by showing a character with each primary/secondary symbol shown as a different color, and then you have to type on a keyboard the proper input code for the character. It is a great app to play around with when you have time to kill and nothing to do. If my Chinese reading were better I'm sure I could probably find other interesting apps for learning Cangjie.
  • Create New...