Chapter 2 – The Keyboard

The transcription of language, the conversion of verbal to written, has always been an artistic endeavor. It’s difficult to take brush or pen in hand and use it to apply ink to paper without thinking about how it looks, even if it’s only to worry over the qualities of the handwriting. We wonder what those who eventually attempt to read it will think. We know they will consciously or unconsciously judge us by this attempt to communicate. Some will pack the page struggling to get in all the information they can and some will inscribe defensive runes, such as happy faces or hearts, to protect them from judgment. Some even take defiant pride in the illegibility of their scrawl.

The keyboard destroyed much of this part of our language by forcing an input standard on us that limits our choices to those that can be expressed by the press of a key. The product is uniform and although this has been a boon for the cause of communication, perhaps something was lost that we are only now attempting to reclaim. Email and now especially texting have sparked the creative abuse of language that causes purists to squirm. I suppose I must follow proper paragraph form and give supporting examples, but I know that you probably already know some of them. “lol” for laugh out loud, “brb” for be right back, or “:)” for a smiley face, are ligatures or perhaps compound glyphs that are shared nowadays by cultures throughout the world. There are online dictionaries filled with other glyphs like them. Like all glyphs, they depend on shared knowledge for mutual comprehension.

It is the artistic nature of writing that makes it difficult to force through a keyboard. We are limited by the number of our fingers and the reach of our arms to a finite number of choices. Remember the complexity of Arabic? Not just its cursive nature, but the compound symbols as well. These must be expressed in terms of keystrokes, the pressing of a single key. The keyboard maps keystrokes to code values, just as those code values are mapped to code points. Every glyph that appears on the page must in some way be mapped to a set of keystrokes.

QWERTY

The entire planet uses more or less the same keyboard, differences being limited to the positioning of the and a few control keys. We can thank cheap PC’s for this. They spread so quickly that all other standards were swept away. It’s the mapping that’s important here. How do we use the limited set of keys we have available to express the hundreds of thousands of glyphs required by the languages of world? How did and will the cultures Figure 1: The Sholes and Glidden of the world force their written languages into a handful of keys . spanning the width of a pair of hands?

The keyboard we use today was invented by Christopher Latham Sholes, with assistance from Samuel W. Soule and Carlos S. Glidden, in 1873. It was the first commercially successful typewriter, later being bought and produced by the arms manufacturers E. Remington and Sons. It featured the standard four row “QWERTY” layout that we know today, although at first it only had upper case letters. Its initial reception was stumbling and barely profitable. Typed documents were considered harsh on the eyes and typed letters impolite. But it was useful where precise communication was required, and so the typewriter gradually worked its way into offices and then finally homes as workers brought the skill back from their jobs.

The QWERTY layout, and others like it that we use today, was dictated by the proximity of the type hammers to each other. When two hammers in close proximity are pressed one after the other they tended to collide and jam. Sholes did extensive statistical analysis on character order within words and sentences and ordered the keyboard to avoid this. The typewriter was an American invention, but it was quickly taken up by the rest of the western world. German Olympia and Swiss Hermes were much treasured and sought after. Although they standardized on the QWERTY form of keyboard, they were instead QWERTZ (German), AZERTY (French), or ЙЦУКЕН (Russian), each following the same statistical design principles.

Sholes’ methodical approach to design used 0.75 inch or 19mm spacing as opposed to the 0.9 inch or 23.5mm spacing used in piano keyboards. His design has proved to be so successful that we still use it today on standard

“full-sized” keyboards. There are of course other sizes of keyboards, mostly Figure 2: Jammed typewriter smaller. Many laptop computers use smaller keyboards due to the limitations hammers. of the size of the device and many phones use “thumb-boards”, tiny keyboards for text entry. He placed keys for commonly used character pairs on opposite sides of the keyboard in order to speed to up typing and ease work. This zigzag mapping principle has been used by language after language as they’ve mapped their keyboards.

So there was a lot of keyboard precedent to work with when extended ASCII and then code pages helped Europe into the computer age. The familiar typewriter keyboard patterns were just copied onto the computer keyboard. The path from typewriter, to teletype, to PC is obvious. But at the dawn of the computer revolution many parts of the world were still wrestling with the feat of fitting their written languages into the typewriter. The situation for them had become critical because modern civilization, if they wanted to participate, was not waiting for them.

Mechanics and Terms

We must first understand the thing itself before we can understand the problems with making use of it. There are a number of important terms used in describing aspects of keyboard use. The act of pressing and releasing a key is called a “Keystroke”. Keys generally have more than one value, even with typewriters, accessible through a variety of “modifier keys”. In English keyboards they are Shift, Control, and Alt, but other language’s keyboards can have many more. In the QWERTY keyboard Shift, Control, and Alt are called “Shift” keys, which when held down during and through the following keystroke changes its value. For both typewriters and computers, the most common shift function is to change key values from lower case to upper case. Another type of is a “Dead” key. Dead keys, when pressed and then released, alter the next keystroke. Shift-lock is a . A third type of modifier is a “Compose” key. Compose keys are an extended type of dead key that allows the entry of a series of keystrokes that will eventually be combined into a single code value.

Compose Key Sequence Composition Value ‘a á oc © ae œ

Although most of the world uses the standard QUERTY physical layout keyboard, there are often different shift keys to allow for the character mappings required by the language. The only differences between different nationality’s keyboards are the symbols embedded in the tops of the keys. The keyboards underneath do not have character values encoded within them. Instead they transmit a scan code for each keystroke. The scan codes are translated by the operating system into the correct value. This is why keyboards can switch language mappings on demand despite what they have written on the tops of the keys, a feature offered by most modern operating systems. There are two other terms commonly used in references to keyboards, key press and key release. Key press is the act of pressing down a key, followed by a key release. Each of these acts produces a separate scan code and can thus be tied to a different code value. They are often used in conjunction with mouse events, such as selecting groups of objects on the screen. All of these opportunities to generate different scan codes and scan code combinations are used by different cultures to squeeze the most out of their keyboards.

Standards

There isn’t a single standard for keyboards, rather there are many. But there are two dominant physical keyboard standards that are commonly in use throughout the globe, ISO1 and ANSI2. ISO stands for the International Organization for Standardization, based in Geneva Switzerland, founded in 1947. It a private company that sets worldwide standards, negotiates treaties, and advises governments in setting

1 The International Organization for Standards. http://www.iso.org/ 2 The American National Standards Institute. http://www.ansi.org/ local standards. ANSI stands for the American National Standards Institute, another private standards organization based in the United States. ASCII is an example of an ANSI standard. The difference between these keyboard standards may seem small when they are compared, but to touch typists those differences are very important. First, ISO standard compliant keyboards reduced the size of the left to make room for an extra key. Internationally, in most keyboards it’s used for greater than and less than. Another important difference is that ISO puts the tilde “~” on the right, next to the . ANSI compliant keyboards have the tilde in the upper left corner, a feature that is very favored UNIX users who use Figure 3: ISO (left) and ANSI keyboards the key quite a bit.

You might think that that extra key next to the shift on the ISO keyboard would make a very handy additional shift key and that’s probably what ISO was thinking when they designed the layout, especially in cultures with large character sets to support. Unfortunately no one has been able to take advantage of it. Computer makers pretty much used whichever keyboard was coming off the docks at the time. Since software companies could not count on its existence, keyboard mappings tended to conform to the ANSI key layout, ignoring the additional key in the ISO layout.

So most cultures keyboards use the ANSI layout, but who decides what the key mapping will be for a particular language? Generally it’s a government with dominant control over the culture that sets the official standard. Many have national standards offices, although most will turn to private international standards organizations because of their reputation and standing. A keyboard mapping is not just a national standard. It’s a world standard that will extend beyond any Figure 4: Sebeolsik Final and Dubeolsik keyboards one nation’s borders.

Sometimes though, lesser standards can take hold when the official standard does not meet user’s needs. In Korea for instance, the official standard is Dubeolsik (두벌식), set by the South Korean government in 1969. However there are several other standards: Sebeolsik 390 (세벌식 390), Sebeolsik Final (세벌식 최종), and Sebeolsik Noshift. These emerged in the 1990’s. The Sebeolsik standards support the English QWERTY layout as well as Korean. DubeolSik is pure Korean based on Korean typewriters. The flexibility of the computer combined with the need to use English mixed with Korean allowed new unofficial standards to form in competition with the government standard.

Unfortunately standards are only meaningful if the software companies are willing to support them. Microsoft IME natively supports only Dubeolsik. Anything else requires the installation of additional third party software. The support of major computer manufacturers is very important in setting a standard, which is why the use of internationally recognized standards organization such as ISO and ANSI who can better gain their cooperation are important.

Outside the Alphabet

It should be clear now that alphabetic languages are handled in a fairly straight forward manner, mapping characters to keys, common keys to the strongest fingers, common key pairs split to each side of the keyboard, shift keys altering key values to less common mappings, and perhaps a compose key to group keystrokes into composition characters or ligatures such as in Arabic.

But then there’s something we’ve missed. All the other characters, the ones not mapped to keys. Every language has those – all those Wingdings and Dingbats. You can use compose to create ©, but there is no compose sequence to produce “♥”. Unfortunately there is little alternative in these cases to picking the characters manually. Most word processors have an Insert Symbol function to allow us to access all the possible characters available. Many fonts come with more symbols than can be mapped to the keyboard, or in the case of fonts like the Latin fonts,3 which support several related languages, glyphs that are only mapped to keys or key sequences in certain culture settings. They all have code values inside the computer and once you get them in, the computer will buffer them just like any other character. You just can’t get them in using the keyboard directly.

In addition to the selection provided by your language’s glyph set, most word processors will provide you with an extended selection of “dingbats”. Dingbats are considered to be “printer’s ornament” characters. One of earliest digital dingbat fonts was Zapf Dingbat, created by Hermann Zapf for Linotype in 1979. It was first distributed for PC as an Apple postscript font ITC Zapf Dingbats. Since then we also have Wingdings and Webdings. These fonts are known to the word processor and are probably appended to the document’s current font in the insert symbol function. As well, has a special area set aside for Dingbats, code points U+2700 to U+27BF which may be added as well if there are fonts installed to support them.

Dingbats are non-alphabetic characters that have no rhyme or reason. They have no reliable meaning. There is no way to order or identify them other than to pick them as you see them from a list. Hence, there is no logical way to map them to the keyboard. Pictographic languages can be a little bit like this although luckily, with them we have a little more to work with.

Ancient Egyptian

3 Chapter 3 Ancient Egyptian seems like it’s not far away from Zapf Dingbats. Unicode includes over 7100 Ancient Egyptian pictographic4 symbols. Like dingbats, there is no immediate sensible way that we can directly map them to a standard keyboard, nor has anyone yet officially tried. Unlike dingbats, these symbols do have distinct meanings which can provide a handles with which to manipulate them. In order to input any set of characters through the keyboard we are going to have to find a way to map keystrokes to code values, an order that can be used to translate the mechanical key press to a glyph.

Unicode is almost certainly supported by your computer, so all you need to have in order to write Ancient Egyptian is the font that maps to that section of Unicode. It’s called Aegyptus5 and once you have it downloaded and installed, with some patience you can write letters to your friends in Ancient Egyptian. There are three variations of Egyptian hieroglyphs6, Old Egyptian, Middle Egyptian, and Late Egyptian. Although Egypt itself today speaks Arabic, Egyptian is still spoken in the form of Coptic which is the liturgical language of the Coptic Church. However, Coptic today no longer uses hieroglyphs but instead a modified Greek alphabet. The spoken words though are still pretty much the same.

In addition to the three forms, Ancient Egyptian also had a cursive form for hand written documents. Luckily for this example, it’s not yet supported yet by Unicode. Often Ancient Egyptian is described as being ideographic or logographic7 but this is not entirely true. It is instead partly phonographic8, with pictures representing sounds. The pictures themselves were chosen for the way the word for the thing sounded. They would spell the word by sound using phonographic symbols and then they would sometimes add a picture of the thing itself just to be sure of the meaning. For example:

= cat (miw)

m i w

The above symbols give you the word as spoken as well as a logogram of the thing itself.

These glyphs were picked carefully one at a time from the insert symbol window on my word processor, which can be a tedious task. However, if you try this yourself you may notice that there is some logical grouping of the symbols provided by Unicode that can help you to find the ones you’re looking for. They are grouped by picture type, animals, tools, plants, each object in a visual grouping wherever possible. The grouping of ideographic symbols by visual clues is just one way of ordering them, making it possible to find the appropriate symbol.

4 Ideograph or Ideogram is a picture that represents an idea or concept. 5 http://users.teilar.gr/~g1951d/ 6 Hieroglyphs, characters formed by graphical figures. 7 Logograms, pictures that represent words as opposed to ideas or sounds 8 Phonograms, symbols that represent sounds

Another way to approach the problem finding ideographic symbols is to relate symbols to some other intermediate symbol set that we do have easy access to, such as the Latin alphabet in the example for “cat” above – “miw”i. That still doesn’t help us with the cat symbol or with the sheer number of other symbols that have no phonetic link, such as the cartouches that stand for people’s names. There aren’t enough letters in the Latin alphabet to map to all the Egyptian symbols. But then we don’t need all of them all of the time do we? Some are commonly used and some aren’t. Ancient Egyptian has what is probably the first alphabet ever created:

Vulture Flowering Flowering Arm Quail Leg Stool Viper Reed Reeds Chick

Owl Water Mouth Reed Twisted Sieve Vulture Door Folded Shelter Flax Bolt Cloth

Pool Hill Basket Jar Loaf Tether Hand Snake Slope

One Ten 100 1,000 10,000 100,000 1,000,000

With these basic symbols, we can write just about any Egyptian word by sounding it out phonetically and this symbol set will easily fit within the limits of the keyboard. This alone though would not be writing as it was traditionally done, but rather instead only a kind of pidgin Egyptian. Pidgin Egyptian might be enough if the computer itself could be taught to understand it. It could then supply the character in question. Miw could be a composition character set for cat.

Another approach might be to use a search function that allowed us to divide up the remaining characters into categories, perhaps by a physical characteristic of the glyph such as picture type as in Unicode, or even by the way the words sound when spoken, narrowing the search until we arrived at our desired character. Then perhaps we might have an input method that we could fit into a modern computer keyboard.

This type of algorithm is called a “divide and conquer” algorithm and is applicable in the real world as well as in computers. You divide a seemly insoluble problem up into smaller and smaller pieces until you arrive at sub-problems small enough that you can solve them. Then you recombine the pieces into a total solution. But you need some measure to use in breaking things apart. We could consider the visual ordering that comes with Unicode. Using it we might have four or even more keys based on visual type. For simplicity let’s say: animal, vegetable, mineral, and name/cartouche. We pick animal and in response the computer presents us with a set of choices: wings, legs, or fins? We pick legs and get the choice: 2 or 4? We pick 2 and get: standing or sitting? On and on until we have a final selection of five or six glyphs to choose from that hopefully contain the glyph we’re looking for. This may seem torturous but with carefully tailored selection criteria we could get to most glyphs in five or six selections, each selection being a single keystroke, which isn’t much different than the average number of keystrokes in the average English word.

Chinese

So you don’t necessarily have to use single keystrokes to invoke a particular glyph. You can use the software in the computer as an aid, trying to find natural orders within the language to break the character set apart so you can look up required symbols. Chinese, like Egyptian, is also a mix of phonetic and ideographic symbols and presents many of the same problems. These are problems which the Chinese have been struggling with for thousands of years. Chinese is spoken by over one fifth of the world’s population with the Mandarin dialect being the most spoken language in the world. There are 13 major dialects9, some of which are actually considered to be sub-languages and are unintelligible to other Chinese speakers. The Kangxi Dictionary 10 recognizes over 47,000 characters, with one or two new ones being created each week. How did they fit all this into the QWERTY keyboard?

Unlike America and Europe, China doesn’t have a long history of typewriter use to draw on when designing their keyboards. Building a Chinese typewriter was a difficult problem that was constantly set back by war Figure 5: A revolving table like type case for individual and political upheaval. The evolution of Chinese typing movable-type characters arranged primarily by a took a very different path from ours and the first rhyming scheme. problem they had was in finding a way to order their symbols.

Moveable type was invented in China around 1040 CE, the first being made out of carved wood. With so many characters in their language, the filing of type when not in use became an immediate problem. Some type sets numbered as high as 60,000 pieces11. The first ordering of the type pieces so they could be found when needed was accomplished by a system of rhyming:

9 Mandarin, Wu, Cantonese, Min, Hakka, and Gan are the top 5. 10 Kangxi Dictionary, the standard dictionary for Chinese in the 18th and 19th centuries. 11 The “Records of Jingde County”, a 100 copy edition contains more than 60,000 characters. “For each character there were several types, and for certain common characters there were twenty or more types each, in order to be prepared for the repetition of characters on the same page. When the characters were not in use he had them arranged with paper labels, one label for each rhyme-group, and kept them in wooden cases.”ii

Rhyming is a method, one of three, used in the ordering of Chinese characters in Chinese dictionaries. The first surviving “rhyme dictionary” is the Qieyun12, published 601 CE. In rhyme or rime dictionaries, characters are first divided into four groups according to “tone names”: level, rising, dark, and low. These categories are divided into sub-groups according to their rhymes or tones. These dictionaries where mostly intended for the composition of poetry, but they worked as traditional dictionaries as well.

Another system of ordering, used in what were probably the first dictionaries ever created, was first published starting in the 3rd century BCE. These were specialized dictionaries that ordered words on particular subjects, such as plants. They ordered words into semantic categories, such as grasses, brush, or trees, much like our plant recognition books do today. In this case, it was necessary to have some knowledge of the subject being discussed in order to understand the semantic divisions being used in order to reach the explanation of the symbol in question.

Radicals and Strokes

The last type, appearing around the 2nd century CE, used a graphically organized system of “radicals” and “strokes.” Radicals in English are the roots of words, after the prefixes and suffixes are removed. In Chinese, symbols are often composed of lesser elements. For instance 采 cǎi ‘to pick, pluck’ is composed of two elements, hand 爫 (zhǎo or zhuǎ) and tree 木 (mù). Neither of these characters is an actual root in the English sense, but many Chinese characters break down into simpler terms. Because

12 Cutting (spelling) Rhymes Figure 6: CJK strokes defined in Unicode of their limitations of space, the width of the character, and the limitations of design elements, the brush strokes, Chinese characters can also be characterized by design elements as well. For instance 亅 jué or juě in 了 liǎo. Jué is not a real word, but a design element, a visual root. Together these basic words and design elements comprise a set of characters collectively considered to be radicals in Chinese. They are used as section headers in dictionaries and generally number about 200. They are the first division of the Chinese character set.

Strokes are just what they sound like, brush strokes. There are a set number of brush movements that can be used in the creation of a Chinese character. In mathematical terms, you could think of these as eigenvalues, the most basic common components of a set of objects that can be used to create all of the objects in the set. Just about all Chinese characters are composed of these basic brush strokes. Stoke type and stroke count, were and still are used to further subdivide the dictionary. Combined with radicals they form an important system of collation that is used as the basis for several systems of character input.

Four Corners 四角號碼檢字法

An important method of organizing Chinese came about in the 1920’s, The Four Corner Method. It was developed to help telegraph operators look up codes for characters in their code tables. A number is used 1 Horizontal stroke to “encode” the shape of each of the four corners of the Virticle or diagonal stroke Chinese symbol in order to produce a four digit number, 3 Dot stroke which can then be used as an index into the telegraph 4 Crossed strokes encoding table. 5 Three or more intersecting strokes 6 A box This system has been adapted for many uses including 7 A corner stroke dictionaries, library catalogs, telephone books, and recently 8 八 or 八 inverted in computers as a method of selection from limited datasets 9 小 or 小 inverted or sets of characters. It isn’t precise enough to be used for 0 Empty corner the language as a whole, but is useful in returning a quick Figure 7: Four corners stroke encoding. subset of characters from a limited list that can then be chosen from to arrive at the desired input.

Wubi 五筆字型輸入

The Wubizixing Input Method, or Wubi for short, is a method of keyboard input that has some similarity to four corner in that it encodes visual elements of the character into four codes, letters instead of digits. Instead of four fixed points, encoding progresses in the order in which strokes are applied when writing the character. If there are more than four strokes, then it uses the first three and the last. Strokes are grouped on the keyboard

Figure 8: The Wubi . into five sections, based on stroke shape or direction.

Wubi is a popular method for touch typists as it doesn’t require selection from lists on the screen, and if you can remember how a character is supposed to be written then it can be very fast. Unfortunately, thanks to computer assistance, many Chinese are forgetting how to brush characters and there are too many strokes to fit on a cell phone keypad. Another similar system is Boshiamy, which uses about 300 radicals.

Cangjie 倉頡輸入法

The most popular input method for personal computers, because it’s found installed natively in most Chinese operating system versions, is Cangjie. It was first developed for the Apple II system. It uses elements similar to the Chinese dictionary system, but reduced and tailored to fit common keyboards.

Cangjie starts with a set of 24 non- standard radicals mapped to the normal Latin alphabetic keys. X is mapped to the difficult character key. To help the user remember the radicals, they are divided into four arbitrary groups: Figure 9: The Cangjie keyboard. philosophical, stroke, body parts, and character shapes groups. Radicals are selected with the keyboard and assembled on the screen in an order defined by a series of rules such as: left to right, top to bottom, or outside to inside. There are some characters that do not fit the decomposition rules and punctuation marks that must be input using fixed codes. There is another less popular system Dayi, which uses 46 radicals rather than 24.

The system is difficult to learn, but once mastered can usually produce typing speeds of 25 characters per minute. Experts can produce as high as 60. It has the drawback though of requiring a QWERTY keyboard which means it cannot be used with cell phones. There is a more automated Figure 10: Cangjie radicals in version called Simplified Cangjie, which only requires the user to enter the order as they are combined. first and last radical code. The user is presented with a list of possible characters to choose from. Although easier to use, it is difficult to reach high typing rates.

Bopomofo 注音符號

Bopomofo is very popular input method used in Taiwan and like Cangjie, it often comes standard on Chinese versions of operating systems. It is one of two popular phonetically based alphabets, the other being Pinyin, which can be used to describe Chinese characters by how they sound when spoken in Mandarin. It uses 37 characters and 4 tone marks, a just small enough character set that it can be used on telephone keypads.

Bopmofo was developed in the 1920’s by the Republic of China as an aid for showing pronunciation in dictionaries, text books for children, and books for foreigners. The Bopomofo character set is based on syllable onsets and rimes rather than consonants and vowels and is similar to the system used in early rime dictionaries.

Pinyin 拼音輸入法

Pinyin is a Romanization system developed by the People’s Republic of China in the early 1950’s for use in computers and phones and is the most common text input system in Mainland China. It replaced older Romanization systems such as Wade-Giles and superseded Figure 11: Bopomofo phone keypad. Bopomofo on the mainland. It’s function and purpose is similar to Bopomofo in it use in books and street signs, except that it uses Latin characters instead of stroke based characters.

Pinyin data entry is generally highly automated. Each character entered presents the user with a list of most common words that begin with the sounds entered. As more data is entered, the lists become more specific. This feature makes it very popular on cell phones and PDA’s as it reduces typing. It helps too that Pinyin is ubiquitous on just about every device and operating system produced throughout the world.

Conclusion

Languages with an alphabet were at an advantage in the development and dissemination of typewriter technology, allowing the transition to computer keyboards to be fairly straight forward. Phonetic and ideographic languages however were at a disadvantage and were caught at a disadvantage when computer technology arrived. It was this technology though that allowed them to overcome their disadvantage through the development of new input technologies such as Pinyin and Wubi.

These new systems of input are based on divide and conquer solutions, built around common elements within the languages themselves, such as visual elements in their writing or phonetic elements in their speech. They sometimes use an intermediate form of writing such as Bopomofo or Pinyin that requires fewer total numbers of keys, used in combination in compound expressions, to identify single glyphs. With these systems they can reach input rates that are higher than their traditional Latin based counterparts.

We as people have developed a wide range of input strategies, providing a wide range of precedent to work with, as new languages are incorporated into the World computing community.

Figure 1: Smith, Clarence Charles (1922). The Expert Typist. New York: The Macmillan Company, p. 4 Melville, Arthur (1923). "The Machine Gun of Commerce" The Rotarian (Rotary International) 23, p. 18

Figure 2: Audriusa. Typebars, showing the guiding mechanism till the paper. Wikicommons. http://en.wikipedia.org/wiki/File:Typebars.jpg

Figure 3: bitguru. Image is a modified version of an image posted on Bitguru Blog. http://bitguru.wordpress.com/2008/05/23/ansi-vs-iso-keyboards/

Figure 4: A modified version of “Keyboard layout, Hangul” combined with “Keyboard layout of Sebolsik Hangul”. Wikicommons. http://en.wikipedia.org/wiki/Keyboard_layout.

Figure 5: Wang Zhen (1313 CE.). The Chinese agricultural book, “Nong Shu.” Wikicommons. http://en.wikipedia.org/wiki/File:Chinese_movable_type_1313-ce.png

Figure 6: Unicode Consortium. Modified from http://www.unicode.org/charts/PDF/U31C0.pdf

Figure 8: Wubi keyboard.

Figure 9: minghong. Keyboard Layout Middle North America. Wikicommons. http://en.wikipedia.org/wiki/File:Keyboard_layout_Cangjie.png

Figure 10: Samwingkit. Coding of "倉頡輸入法" (i.e. Cangjie method) in traditional Chinese characters. Wikicommons. 倉頡輸入法_拆碼.jpg

Figure 11: Callek6. ZhuYin keypad for a mobile GSM phone. Wikicommons. http://en.wikipedia.org/wiki/File:Bopomofo.jpg

i Omniglot. Ancient Egyptian Scripts. http://www.omniglot.com/writing/egyptian.htm ii Tsien, Tsuen-Hsuin (1985). "part one, vol.5". in Joseph Needham, Science and Civilisation in China,. Paper and Printing. Cambridge: Cambridge University Press. Omniglot, Pŭtōnghuà (Mandarin). http://www.omniglot.com/writing/mandarin.htm