WuBi 五笔 Input Method

Write Chinese characters with a computer keyboard

Robert SIEMER

http://backsla.sh/talks

[beware, your fonts may render dramatically different]

2010-06-21

Overview

• know the basics
• first exercise together
  • understand what Wubi is about
• further introduction (25min)
• step-by-step guide to master Wubi (rest of the time)
  • simple exercises follow each step

The Wubi basics

• it is not pinyin
• for standard keyboards
• different character components spread over 25 letter keys
• for each character to write, pick the components one by one, in order of hand writing
• at the end: commit/confirm 
  • like with other Input Methods as well
  • usually “space” key commits

Exercise I (2min)

Part I — Theory

Input Methods

• pronunciation based
  • Hanyu Pīnyīn  (nǐ hǎo)
  • Standard Cantonese Pīnyīn (nei5 hou2)
  • Zhùyīn/Bopomofo (ㄚㄛㄜㄝㄞㄟ)
    
• character/structure/stroke/writing based
  • 4-corner (method used for name lists)
  • Cāngjié
  • Wǔbǐ
  • Dàyì
• combination of both

Structure based Input Methods

• many different methods exist, differing in
  • popularity
    • in Mainland China very clear: Wubi very dominant
    • e.g. Taiwan: not so clear (keyboard labels ready for two)
  • official status
    • numbers for name characters on forms/registration purposes
  • keys used
    • only numbers (0-9 or even less)
    • “letters” (that is, 26 or a little less keys)
    • “everything” (e.g. 46 keys)
  • adapted language/script variant
    • simplified/traditional
    • languages (Korean, Japanese, ...)
  • decomposition basis
    • none/indexed/numbered
    • dictionary radicals
    • writing related (stroke, components)
    • shape related (at predefined positions or writing order)
  • other characteristics
    • code length (speed, auto commit)
    • ambiguity (official use, blind typing, speed)
    • wildcards (ease of use, searching)

Wubi characteristics

• invented by 王永民 WANG Yongmin
• developed for simplified Chinese
• “normal” is Wubi86
• a revamped version, Wubi98, had no success
  • a little more suitable for simplified and traditional characters
  • same rules, but a small changes on the components
• depending on the implementation: 
  • Wubi86 may yield traditional characters as well
  • yields over 20,000 characters
  • characters combinations available as well
  • nice wildcard key Z
    • limited structure based character search capability
• high typing speeds

Advantages

• in any case: you should not learn it instead, but in addition to Pinyin
• some say: “to complicated for foreigners”, I say: “especially useful for Chinese learners”

Part II — Practice

Bihua Input Method

• is not the same, but Wubi shares some concepts with the Bihua Input Method
• most mobile phones have Bihua (apart from Pinyin)
• characters are entered stroke by stroke
• each stroke has a number form 1 - 5
• number depends on stroke type

Bihua and Wubi stroke types

Exercise II (3min)

Wubi partition: zones

• Wubi keyboard divided into five zones
• • according to Bihua stroke types
  • to ease lookup
  • first stroke of component decides which zone it belongs to
  • some exceptions exist, though...
• zone map

Exercise III (3min)

Wubi partition: keys

• every zone has five keys
• counting starts from the “middle”
• for example key “24” is: zone 2, key 4 (English keyboard: L)
• different rules exist to determine the key number of a component
• • second stroke type
  • repetition of first stroke (type)
  • similarities to other components on the key
• otherwise try
• • key number mnemonics (e.g. total stroke count)
  • key letter mnemonics (shape?)

Exercise IV (2min)

Exercise V (2min)

Exercise VI (4min)

Wubi difficulties

• unfortunately, it is hard to tell which component is available on Wubi
• • look through the keyboard layout and see what is there
  • create your own mnemonics
  • at first, remember the frequent ones
• sometimes a geometrically connected part of a character needs to be decomposed in Wubi
• • the easiest rule is: do it intuitively
  • follow the stroke order
  • minimize the number of components

Exercise VI (5min)

Code shortening

• all Wubi codes are limited to four or less keys
• characters with more components get shortened 

Code prolongation

• all characters have an official code of 3 or 4 keys
• 1 and 2 key codes are always abbreviations for frequent characters
• some “natural” codes are to short and collide with abbreviations
• writing of components often raises this problem, e.g.

Exercise VII (4min)

Code prolongation: distinguishing key

• all “natural” 2 and 3 key codes can also be extended with one extra key
• 1. determine the stroke type of the last stroke of the last component (1, 2, 3, 4, 5)
  2. analyze the geometric shape ⿰ is 1, ⿱ is 2 and others are 3

• extension only necessary in case of collision
• mostly optional
• forbidden only in case of collision...

Exercise VII (3min)

Nice to know

• Wubi is not (dictionary) radical based
• • many components are not radicals
  • some radicals really look like being composed of two, three others (and so are in Wubi)
  • 殳青足音是黾香高

• some radicals are broken up in hand writing (and Wubi)
• 亘 has radical 二

• sometimes Wubi does not follow the Mainland stroke order, or no valid stroke order at all
• Wubi: 力 = 4 5     戈 = 1 5 4 3

• you can actually enter multiple characters with one code
• 你好  对不起  不好意思  很多 

References

• Basic Stroke Types  

• 227 radicals for simplified characters

• Wikipedia
  • Input Methods (en)
  • (the English Wubi article is really bad)