Teaching and Reference Material on Japanese Kanji in SVG

Stroke Order, Animated Drawing of Characters, Kanji Components and their Relationships.

Keywords: kanji, path animation, computer-aided language learning, dictionaries

Julien Quint
National Institute of Informatics


Dr. Julien Quint is currently a visiting researcher at the National Institute of Informatics in Tokyo, Japan, under a postdoc fellowship from the Japanese Society for the Promotion of Science. He completed his PhD in Computational Linguistics at the Université Joseph Fourier in Grenoble, France in 2002, where he worked on a language-independent formalism for text segmentation based on weighted finite-state techniques. He is currently interested in applications of this formalism to the Japanese language.

Ulrich Apel
National Institute of Informatics


Dr Ulrich Apel is currently a visiting researcher ath the National Institute of Informatics in Tokyo, Japan, under a postdoc fellowship from the Japanese Society for the Promotion of Science. He studied in Munich, Germany, and Osaka before completing his PhD in Japanology at Munich University in 2002. Since 1998, he is working on a Japanese-German dictionary database; it contains about 215,000 records and 90,000 headwords (see http://bunmei7.hus.osaka-u.ac.jp:591/WaDokuJT/). For a comprehensive description of the Japanese language, better data on Sino-Japanese characters (kanji) is needed, on which he is working now with the help of SVG.


Learning to read and to write Japanese is a difficult venture for Japanese and foreigners alike. While Japanese study these culture techniques in a period of many years at school, foreigners who mostly try to learn them beside job or other classes would rely on good teaching material and easy means to look up Japanese characters. Most kanji dictionaries assume that their users already have a lot of information on the character--its stroke count, its radical or pronunciation. This project tries to put together data for easier look up of Japanese characters and better teaching material. We use SVG for a structured description of kanji and their graphical characteristics; then describe applications of such a description.

Table of Contents

1. Introduction
2. An Introduction to Kanji
     2.1 Kanji as the Main Obstacle for Learning Japanese
     2.2 History of Kanji and their Main Building Principles
     2.3 Problem Description
3. Representing Kanji in XML and SVG
     3.1 Kanji Strokes, Stroke Groups and the Kanji Project
     3.2 Possible Applications
4. Three Example Applications
     4.1 Searching for Kanji by Strokes
     4.2 Automatic Animation of Kanji Strokes
     4.3 Highlighting of Kanji Components
5. Prospects of the Project and Future Work: Animation Details and Variations in Stroke Weight
6. Conclusion

1. Introduction

The complex writing system of the Japanese language, notably its reliance on a very large set of characters borrowed from China (the so-called "Chinese characters", or kanji, 漢字, in Japanese), makes the life of the language learner very difficult, and can even sometimes cause difficulty to native speakers. The many shortcomings of currently available teaching material make the situation even worse.

Kanji are complex graphical entities consisting of possibly dozens of strokes. In order to write such characters legibly, it is necessary to draw the strokes in the correct direction (as this affects the shape of the stroke) and the correct order (as this affects the overall shape of the character). Conversely, to be able to look up an unknown kanji in a dictionary, it is necessary to be able to recognize the exact number and shapes of the strokes.

Shortcomings of paper resources for learning and reference material on kanji can be addressed by more flexible electronic tools. However, if electronic resources are now widely available (in the form of electronic dictionaries for computers, mobile phones, PDAs, or as handheld standalone device; or services on the Web for glossing, transliteration or low-quality automatic translation), they do not take into account the very graphical nature of kanji and treat them as atomic elements, much like letters of the roman alphabet.

Describing the structure of kanji in more details leads to applications that would be difficult to create using the sort of resources described above; these applications would include tools to learn how to write kanji or how to look them up in a dictionary.

An obvious format for such a description is XML; and as we are dealing with graphical entities, SVG is perfectly suited for the visual part of the description (as a companion to a more abstract description of the character). A kanji is composed of strokes organized into components and sub-components; this maps perfectly with SVG elements such as paths (for strokes) and groups (for components). Moreover, SVG is not limited to static content; and animating the strokes of a character is a very clear way to show how to write it.

We show how the structure and shape of kanji can be described in SVG, and what sorts of application this description permits. Of particular interest is the automatic animation of strokes for learning how to write a character correctly.

2. An Introduction to Kanji

2.1 Kanji as the Main Obstacle for Learning Japanese

The modern Japanese writing system is considered as one of the most complex in the world. To describe it Japanese often call it kanji kanamajiri bun (漢字仮名交じり文).

This means that Japanese normally uses the originally Chinese characters—kanji—and kana (hiragana and katakana), which are cursive or shortened versions of former kanji and which now represent only the sounds of syllables. Hiragana are used for example for grammatical particles and function words; katakana are used mainly for non-Chinese loan words and names. Each set of kana consists of 48 different characters, which can be combined with diacritical marks.

It is possible to write Japanese totally in kana, but because of the large number of homophones in Japanese, this would lead to misunderstandings. During a lecture, Japanese professors write kanji on the black board to clarify what they are talking about, and Japanese television makes extensive use of subtitles. Nowadays, latin characters and roman and arabic numerals are used along with kanji and kana.

In Japanese schools, about 1,000 kanji are taught with their correct stroke order at primary school (these are the kyôiku kanji, 教育漢字). When pupils leave compulsory school, they should know about 2,000 of them (the jôyô kanji, 常用漢字). It is said that educated Japanese people know about 3,000 kanji.

The introduction of electronic writing systems has had two contradictory effects on the usage of kanji. The ability to write kanji by hand has worsened, because now handwriting is only one way to write. On the other hand, it has become very easy to write even very difficult and exotic kanji provided that one can key in their pronunciation. Today, it is mostly a passive use of kanji that is needed.

Kanji are difficult even for Japanese people. After a long stay outside of Japan, the kanji ability of Japanese people decreases. If a Japanese is asked to write a difficult kanji, she is not sure about the correct stroke order. This uncertainty then shows in the script.

For foreigners learning Japanese, it is very difficult to learn, memorize, read and write kanji. The subject is of course difficult, but to make things worse, the teaching material on kanji in western languages has often too little information on how kanji are actually written. This concerns for example stroke direction, possible variations of character forms or of their components.

2.2 History of Kanji and their Main Building Principles

There exist kanji dictionaries with tens of thousands of characters, but the number of kanji which were in actual use in Japan didn't exceed 5,000 or 6,000. The kanji set of older computer fonts contains about 6,400 characters. There are additional character sets like JIS X 0122 or JIS X 0213, which contain several more thousands kanji.

Chinese characters were developed in the second millennium before Christ. Kanji were introduced in Japan in the fourth century AD. Together with kanji, many Chinese loan words came to Japan with their original pronunciation adaptated to the Japanese phonetical system. Sometimes, the same kanji was introduced several times in different compounds, from different eras or different regions in China, with different pronunciations. Furthermore, in Japan, kanji are very often used with their meaning only, and the pronunciation of the corresponding Japanese word is assigned to them. That means that the same kanji may be pronounced differently according to its usage.

Kanji are often called ideographs, meaning that the characters refer to an idea rather than to their pronunciation. Actually, for the majority of them, this is not the case. A typology of Chinese characters from the Later Han Dynasty distinguishes six categories of kanji (rikusho, 六書). The first four consider the way the kanji are built:

  1. Pictographs (shokei, 象形) are pictures of the thing that they designate (e.g. 日 "sun", 月 "moon", 木 "tree", 鳥 "bird", etc.)
  2. Diagrammatic characters (shiji, 指事) represent their meaning symbolically (e.g. 一 "one", 二 "two", 上 "above", 下 "below", etc.)
  3. Characters with combined meanings (kaii, 会意) combine simple characters to indicate a meaning: for example the combination of 日 and 月 ("sun" and "moon") gives 明 "bright".
  4. Phonetic characters (keisei, 形声) have one element that suggests the general area of meaning and another that indicates the sound; for instance, the character ume 梅 contains the kanji 木 "tree" to indicate the meaning, and the kanji mai, 毎 "every", for the sound. This is the biggest group of kanji and seems to be still productive in China and Taiwan for characters used in proper names.

The last two categories deal with (5) extensions of the original meanings (tenchu, 転注; for example, the meaning of 悪 "bad" was extended to mean also "hate"); and (6) characters which are used with a borrowed meaning (kasha, 仮借; for example, the character depicting a scorpion 萬 is used for the meaning of 10,000).

There are also a few kanji which were invented in Japan, often called kokuji, 国字, such as tôge 峠 "mountain pass"; sakaki, 榊 "Cleyera Japonica"; tsuji, 辻, "crossroad"; and hata or hatake, 畑, "plowed field". Such characters normally have no Sino-Japanese reading.

Developments following WWII lead to a separation of kanji forms in Japan, China, Taiwan and Korea. Japan and mainland China invented simplified forms of kanji (often different forms for different kanji), whereas Taiwan and Korea still use the traditional kanji forms (in Korea, the use of kanji is nowadays quite limited). For example, in Japan and China, gaku or manabu, 學, became 学; ten, 點, became 点. In Japan, ben is now written 弁 instead of 瓣 or 辯.

In Japan, the Ministry of Education (Monbushô) selected a number of around 2,000 kanji for official and general public use. In 1958, it published instructions for the stroke order and kanji forms that should be taught at school for about 900 kanji. Traditionally, there used to be a certain freedom in stroke order and kanji forms, but the Ministry of Education selected one "official" form. Although the Monbushô tried to generalize and use rules for stroke order, they still seem quite arbitrary. Calligraphers also stress that the proposals of the Monbushô don't use the more common character forms or stroke orders [Emori, 2003] .

2.3 Problem Description

Although nowadays kanji are very often written on a computer, a typewriter (wapuro) or a mobile phone, it is still important to be able to write them by hand. In Japan, for example, letters written by hand are considered more personal than printed ones, and people with a nice handwriting are considered smart and well educated. For learners of Japanese, writing kanji by hand is still one of the best ways to memorize them.

To recognize Japanese written by hand, or inscriptions and computer fonts which use characters that were written by hand as models, it is important to be able to recognize the original strokes, because in handwriting strokes are often joined up and hard to recognize as individual strokes. In small fonts sizes and on the computer screen, strokes are often hard to distinguish. As kanji lexica use the stroke count as one criterion to index their contents, it becomes more difficult to find a certain kanji if one isn't able to count its strokes.

Most lexica give no information on the stroke order, and if they do, only for a small number of characters. Some dictionaries with stroke order don't use the kanji forms and stroke order considered as the standard by the Monbushô. Normal kanji lexica have little information on variations of the components of the kanji. Ordinary paper lexica are of little help if one only recognizes parts of a kanji. One has to recognize the whole kanji to be able to look it up. These might be further reasons for the bad handwriting of many western learners of Japanese, and their overall difficulty of learning and using kanji.

3. Representing Kanji in XML and SVG

3.1 Kanji Strokes, Stroke Groups and the Kanji Project

The project tries to put together data that will lead to better teaching material, and easier (and, if needed, more complex) ways to look up kanji. Furthermore, it will allow a number of new ways to display informations on kanji. This is achieved by considering kanji as hierarchical, graphical entities, and representing data in XML and (for the visual part) SVG.

As a very abstract and basic way to analyze characters, one could use a graphetic approach which would lead to the recognition of graphemes. Graphemes are defined as the smallest units that allow to distinguish meaning. For example, in the case of kanji, this could be stroke length (as in 土 vs. 士, or 末 vs. 未), angle of the stroke (as in 三 vs.vs. 彡, and even, if kana are also considered, ミ), stroke direction (as in 干 vs. 千), or the ending of strokes (as in 干 vs. 于). A more concrete analysis, which also takes the act of writing into account, uses the brush or pen strokes as basic units of kanji. This is the one chosen here.

The stroke form of kanji is predetermined by the tool used in former times, the brush. There are no long strokes from right to left or from the bottom up, which would go against the direction of the hairs of the brush when one writes with the right hand (not only doest it seem to be more difficult to write kanji with the left hand; left-handed persons have even trouble using graphical character recognition for PDAs or tablet computers). An analysis of strokes distinguishes 25 basic forms of strokes. This analysis considers stroke direction, bending of the strokes, endings (blunt or with a short bend) and so on.

Strokes can be grouped together not only to build a full kanji but also to build smaller units. Kanji dictionaries include radicals, but there are other stroke groups too, which frequently occur in kanji, although, even in Japanese, there does not seem to be any technical term for them. For exemple, a Japanese graphic program for building non-standard kanji (gaiji, 外字) just calls them "parts" (pâtsu, パーツ). We chose the term grapheme elements.

The analysis of the strokes and of the grapheme elements depend on each other. If one can recognize grapheme elements within a kanji, one can use existing data to get the description of its component strokes. Furthermore, one can look for stroke combinations to find grapheme elements.

The analysis of the grapheme elements uses mostly existing kanji or given radicals. This might be extended.

To display the collected information about a kanji and its components, path data is needed. This data is put together with a vector graphics program (namely, Adobe Illustrator). The stroke order is identical with the input of the paths. The database for grapheme elements can be used to look up the components of a kanji, and the basic elements can be added together with copy and paste. The main work is then to put everything at the right place. For later review and to have more flexible data, numbers for the stroke order are put beside the strokes.

The data for a given kanji can then be exported as SVG. Figure 1 shows one of thousands of kanji made available through this project. The source code for the SVG data in Figure 2 shows clearly the hierarchical organization and the order of the strokes through the use of g and path elements.


Figure 1: A kanji in SVG

<svg width="109" height="109" viewBox="0 0 109 109" style="fill:none;stroke:black;stroke-width:4;stroke-linecap:round;stroke-linejoin:round;">
  <g id="01_pos_left">
    <g id="01_rad_general">
      <g id="01_strgr_さんずい/水V">
        <path id="01_1" d="M21.38,19.75c3.31,1.47,8.54,6.05,9.37,8.34"/>
        <path id="02_1" d="M16.75,45.5c3.79,1.15,9.8,4.72,10.75,6.5"/>
        <path id="03_6" d="M14.75,88.21c1.5,1.31,3.31,1.36,4.25-0.25c2.75-4.7,5.5-10.45,8-16.45"/>
  <g id="04_pos_right">
    <g id="04_strgr_くさかんむり/艸V">
      <path id="04_2" d="M38.07,23.44c0.99,0.4,2.15,0.5,3.14,0.4c10.79-1.09,34.04-4.59,44.49-4.63c1.66-0.01,2.65,0.19,3.48,0.39"/>
      <path id="05_3" d="M49.35,13.24c1.34,1.1,1.77,1.43,1.92,2.11c1.34,6.33,1.99,11.1,2.38,13.4"/>
      <path id="06_3" d="M71.38,10c0.32,0.9,0.67,1.55,0.38,2.68c-1.01,4.07-1.26,9.57-2.37,13.91"/>
    <g id="07_strgr_口">
      <path id="07_3" d="M41.75,33.46c0.42,0.31,0.85,0.57,1.04,0.96C44.25,37.5,46,46,47,51.53"/>
      <path id="08_11" d="M44.38,35.55c8.62-1.61,31.12-4.38,36.06-4.79c1.8-0.15,2.89,0.88,2.63,1.75c-1.07,3.6-2.32,7.94-3.91,12.92"/>
      <path id="09_2" d="M47.2,48.83c5.8-0.46,25.8-2.25,33.77-2.95"/>
    <g id="10_strgr_夫">
      <path id="10_2" d="M45.21,60.14c0.71,0.29,2.03,0.43,2.74,0.29c2.89-0.57,24.08-3.06,30.52-2.96c1.19,0.02,1.91,0.14,2.5,0.28"/>
      <g id="11_strgr_大">
        <path id="11_2" d="M38.84,72.88c0.71,0.29,2.03,0.43,2.74,0.29c2.89-0.56,39.3-4.58,45.74-4.48c1.19,0.02,1.91,0.14,2.5,0.28"/>
        <path id="12_4" d="M61.85,34.3c0.48,0.95,0.66,1.84,0.66,3.68c0,42.02-6.27,50.77-27.77,58.52"/>
        <path id="13_5" d="M61,72c9.25,7.5,21.02,17.37,29,21c2.69,1.22,4.94,1.62,6,1.79"/>

Figure 2: SVG source code for the kanji

3.2 Possible Applications

The data allows new approaches for searching kanji, such as:

It also allows new ways to display kanji:


Figure 3: Building a character by its strokes in order


Figure 4: An example practice sheet with arrows for stroke direction and numbers for stroke order

The SVG representation can be manipulated easily through style sheets or scripts to provide these modes of display. Altering the data (to show variations of the stroke orders, for instance) is easy to achieve as all the information necessary is already present.

Applications to graphical character recognition or correction of drawing mistakes are two more examples.

4. Three Example Applications

4.1 Searching for Kanji by Strokes

The easiest way to find a kanji in a dictionary, be it on paper or in electronic form, is to know its pronunciation. In the case of an unknown word, when the pronunciation information is not available, one must usually go through the index in dictionary, which is organized by number of strokes and radicals. For beginners, this is extremely difficult and frustrating as it is hard to guess which part is the radical, and exactly how many strokes there are.

However, even a beginner (who has learnt to read and write at least a few dozens of kanji) is able to recognize basic strokes and even some components of a kanji, though not necessarily the radical. Using the graphemic data described above, one can search for kanji by inputing the strokes or components that can be recognized with confidence.

Note that for the search itself, the graphical data is not used; the characters are represented by a list of numbers, each number representing a type of stroke, in the correct order. So for instance the character shizuka, 静, corresponds to the list of strokes 2, 3, 2, 2, 4, 17, 2, 2, 4, 15, 11, 2, 2, 8.

By inputting a series of stroke numbers, the user can get all the kanji that match this description. Moreover, the request can also include kanji which appear as components of more complex characters. For instance, one can search for 静 by including one of its components, 青, in a request.

Note that the more strokes the user can distinguish (and add to her query), the better the precision will be; however, if a stroke is not correct, the right character may not appear in the result set. Achieving the best precision and recall, as well as the best relevance ordering of results is a classical task of information retrieval.

The current implementation of this application is still tentative and many improvements are needed. The first one is that we should take into account the actual groups of stroke (so far, only a flat, albeit ordered list is used), which would improve precision dramatically. There are also many ways to input the request, which leads to graphical user-interface design questions.

4.2 Automatic Animation of Kanji Strokes

In this application, the kanji are not shown statically but dynamically, by drawing every stroke progressively as if they were actually drawn by someone behind the screen. This is illlustrated by Figure 5 where every stroke is rendered in sequence, with a pause between strokes; the stroke currently drawn is highlighted for clarity. (Reload to restart the animation from the beginning.)


Figure 5: Animated strokes

The original, static, SVG document contains all the data necessary for such an application. The strokes are in the correct order, and the points defining the paths are set so that the stroke is drawn in the correct direction (e.g. the first stroke of Figure 5 is drawn correctly from the left to right). Each stroke corresponds to a path element, defined by successive Bézier curves.

Although SVG provides a solid framework for animation, there is no "out-of-the-box" solution for the progressive drawing of a path. Some hacks have been suggested, but it quickly became clear that the animation steps had to be computed beforehand. The method chosen consists of splitting the paths in shorter segments, each segment being a curve of its own. It is then possible to draw every segment in turn, therefore providing animated content. Of course, since there are thousands of kanji, with possibly dozens of complex paths in the dictionary, the cost of splitting the paths manually is prohibitive. Moreover, automatic splitting allows the user to control parameters of the animation, such as the number of steps for every stroke and the overall speed of the animation.

Remember that for a given character, we have an order list of strokes, with a one-to-one correspondance between strokes and paths. In order to create the animated stroke, the following steps are undertaken:

  1. the d attribute of the path element is parsed using regular expressions. After parsing, all coordinates are transformed into absolute coordinates; e.g. a c curve is transformed into a C curve. After this step, the path data consists only of an absolute moveto (M) command followed by a series of absolute curveto commands (C).
  2. Every curve is then split in two using the algorithm of de Casteljau. Each half can then be split further in the same way; the number of steps is controlled by a parameter n (therefore, a curve is split into 2 n segments). In order example, n is set to 3.
  3. The d attribute of the path is then set to an empty string, and an animate child element is added to provide the successive values of d over time. Two parameters control the speed of the animation: first, the duration d during which a segment is drawn (defaults to 0.15 second), and the pause p between two strokes (defaults to 0.1 second). These two values are used to compute the total time it takes to draw a stroke and when the next segment should be drawn.

We create for every "static" character an animated counterpart and store it in a compressed SVG file. The path coordinates are truncated to two decimal places in order to save some space. The result files have an average size of about 5 kilobytes, which is very reasonable. Figure 6 shows (part of) a path element in the animated file. In this example, the path being drawn is shown in red (as indicated by the set child); it then turns to black when the next step begins.

<path d="" id="01_1">
  <animate accumulate="sum" attributeName="d" attributeType="XML" begin="0" dur="0.45" fill="freeze" values="M21.38 19.75C21.79 19.93 22.23 20.16 22.69 20.43;M21.38 19.75C21.79 19.93 22.23 20.16 22.69 20.43C23.16 20.7 23.63 21.01 24.12 21.35;..."/>
  <set attributeName="stroke" attributeType="CSS" begin="0" dur="0.55" fill="remove" to="red"/>

Figure 6: An animated path (excerpt)

The current implementation of the splitting technique is, however, not interactive. The data is processed once on the server side. Therefore, it is not possible for the user to make any changes to settings chosen when the files where generated. In the future, the splitting will be done either on demand, on the server side, or on the client side, using the scripting abilities of SVG.

4.3 Highlighting of Kanji Components

A simple but powerful idea is to highlight the kanji components with color (or some other sort of visual mean of making a component stand out from the others). In Figure 7 , the four main components of the character are highlighted when the mouse is moved over them.


Figure 7: Highlighting different grapheme elements with colors

This highlighting part can be added automatically to the SVG data by appending a set element child to every group that will be highlightable (in this example, the stroke attribute is affected). But since most components correspond to an actual kanji, we can go further and make every component clickable, providing a link to the full kanji form. This is especially useful when the component has a different form from its standalone coutnerpart; for instance, in Figure 7 the leftmost component, highlighted in red, actually stands for 水. Once again, these links can be added automatically.

5. Prospects of the Project and Future Work: Animation Details and Variations in Stroke Weight

The use of Bézier curves for the description of strokes gives very good result when the paths are animated. As the control points of the curves were carefully chosen, the automatic splitting of the paths into segments leads to a rhythm of animation that feels quite natural. However, there are still some subtleties of writing that are not catched by this technique: some vertical strokes should be drawn slower, whereas the "hook" at the end of a vertical stroke should be drawn faster.

Another issue is the actual appearance of the strokes. Japanese characters are normally written with a tool like a brush, a pencil, a pen, a ball pen etc. According to the pressure applied to the tool while writing, the stroke weight varies. The correct change of stroke weight is important for a nice handwriting and aesthetic characters: see the two model characters in Figure 8 . Often, the direction of a stroke can only be recognized by the changing stroke weight, when for example at the starting point the stroke is thicker and tapers to a point at the end.

The path data itself doesn't contain information about stroke weight, but we can use the analysis of the different kinds of strokes from the kanji descriptions. This gives us information about a stroke's shape at its starting point, its course and its end. The problem is to relate this information to the path data.

We demonstrate two approaches. First, we tried the Art Brush function of Adobe Illustrator. To our knowledge, newer versions of SVG will support a similar function. A graphic element depicting a stroke is stretched over the length of a path. According to the kind of the strokes of a given kanji, we could assign different Art Brushes. This sounds very much like a suitable solution for our problem, because Art Brush can simulate actual brush strokes. In our case however, the control on the Art Brush isn’t sufficient for really correct and aesthetic characters (see the paths using Illustrator's Art Brush in Figure 8 ). Start and end points are stretched or compressed too much; the change of stroke weight in the course of a stroke, at bends for example, is very difficult. Art Brushes seem also to be very difficult to animate.

The other approach was it to divide the paths into smaller segments, like for the animation, and assign different stroke weights to these sectors. If the sectors are small enough, the stroke looks as if it had no transitions (see the example with divided vector paths in Figure 8 ). This approach gives full control of the stroke weight and enables us to work out the characteristics of every kind of stroke. For pedagogical purposes, this is an excellent result. As the strokes and the whole character look very well balanced, this approach also satisfies aesthetic demands. This version of the character is also very easy to animate. The result stroke reminds of a stroke by pencil or felt-tip pen, which are the most frequently used writing tools nowadays. The simulation of a real brush stroke, however, seems to be very difficult with this approach too.

If Art Brushes are introduced in future versions of the SVG recommendation, it would benefit from giving more control to the brushes than does Illustrator (for instance) today.


Figure 8: Changes in stroke weight

6. Conclusion

Although neither the analysis of kanji into strokes or grapheme elements, nor animated kanji showing stroke order and direction are totally new, the problem until now was that for every one of these different approaches new sets of data had to be created for every different purposes. The SVG data described here was shown to be adequate for different kinds of applications, three of which have been highlighted. Many more are possible. The data is also extensible, for exemple to the characters of JIS level III and IV, or other character sets such as Chinese character sets, in traditional or simplified form.


[Bußmann, 1990]
Hadumod Bußmann. Lexikon der Sprachwissenschaft. Kröner, Stuttgart, 1990.
[Emori, 2003]
Emori Kenji. Kai gyô sô—hitsujun jitai jiten. Sanseido, Tokyo, 2003.
[Fazzioli, 1987]
Gemalte Wörter. 214 chinesische Schriftzeichen—Vom Bild zum Begriff. Lübbe, Bergisch Gladbach, 1987.
[Frutiger, 1978.]
Adrian Frutiger. Der Mensch und seine Zeichen: Schriften, Symbole, Signete, Signale. Fourier, Wiesbaden, 1978.
[Hadamitzky, 1995]
Wolfgang Hadamitzky. Langenscheidts Handbuch der japanischen Schrift—Kanji und Kana 1. Handbuch. Langenscheidt, Berlin et al., 1995.
[Kôdansha, 1998]
Kôdansha Encyclopedia of Japan. Kôdansha, Tokyo, 1998.
[Schmidt, 1995]
Berthold Schmidt. Einfürhung in die Schrift und Aussprache des Japanischen. Buske, Hamburg, 1995.

XHTML rendition created by gcapaper Web Publisher v2.0, © 2001-3 Schema Software Inc.