Requirements for Japanese Text Layout

2.1.1 Characters Used for Japanese Composition

Japanese letters used for composing Japanese text mainly consist of ideographic (cl-19), hiragana (cl-15) and katakana (cl-16) characters (see [Fig.1]).

[Fig.1]: Kanji, hiragana and katakana.

(note 1)

In addition to ideographic (cl-19), hiragana (cl-15) and katakana (cl-16) characters, various punctuation marks (see [Fig.2]) as well as Western characters (cl-27), such as European numerals, Latin letters and/or Greek letters, may be used in Japanese text. In this document these characters are classified into character classes, for which explanations are given describing their behavior in type-setting. [Fig.2]: Examples of punctuation marks.

(note 2)

The details of character classes used in this document will be explained in 3.9 About Character Classes, as well as in Appendix B Spacing between Characters. Also, in "Spacing between Characters" all non-Kanji characters included in ISO/IEC 10646 (UCS) Annex A collection 285 (Basic Japanese character set) and collection 286 (Extended non-Kanji character set) are explicitly classified by character class.

2.1.2 Kanji, hiragana and katakana

Ideographic (cl-19), hiragana (cl-15) and katakana (cl-16) characters are the same size, and have square character frames of equal dimensions. Aligned with the vertical and horizontal center of the character frame, there is a smaller box called the letter face, which contains the actual symbol. Character size is measured by the size of the character frame (see [Fig.3]). "Character advance" is a term used to describe the advance width of the character frame of a character. By definition, it is equal to the "width" of a character in horizontal writing mode, whereas it is the height of a character in vertical writing mode (see [Fig.3]).

[Fig.3]: The size of kanji and hiragana, and the character frames.

(note 1)

In vertical writing mode, the letter face of small kana (cl-11) characters (ぁぃぅァィゥ etc.) is placed at the vertical center and to the right of the horizontal center of the character frame; in horizontal writing mode, it is placed at the horizontal center and below the vertical center (see [Fig.4]). Also there are punctuation marks with letter faces that are not placed at the vertical and horizontal center of the character frame. [Fig.4]: Small kana and the position of their letter face in the character frame.

2.1.3 Principles of Arrangement of Kanji and Kana Characters

In principle, when composing a line with ideographic (cl-19), hiragana (cl-15) and katakana (cl-16) characters no extra space appears between their character frame. This is called solid setting (see [Fig.5]).

[Fig.5]: Example of solid setting in horizontal writing mode.

2.2.1 Specification of Page Formats

The page format of a Japanese document is specified by:

• Firstly, preparing a template of the page format, which determines the basic appearance of pages of the document;

• Then, specifying the details of actual page elements based on the templates.

2.2.2 Basic Templates of Page Formats

Generally, books use only one template for page format, whereas magazines often use several templates.

Although in books, as will be mentioned later, there tends to be one template for the page format, the basic pattern is typically adapted. For example, the table of contents may contain small modifications. Furthermore, there are many examples of indexes with a different page format than the basic page format, and vertically set books often have indexes in horizontal writing mode and sometimes multiple columns. This still holds true where the goal is to make the size of the hanmen for indexes close to the size of hanmen in the basic page format.

Magazines gather articles of different kinds. Often the page format will differ depending on the content of the article. For example, one part may have 9 point character size and 3 columns, and another part 8 point character size and 4 columns.

2.2.3 Elements of Page Formats

[Fig.11]: Example of a page format in vertical writing mode.

The following are the basic elements of a page format. [Fig.11] illustrates an example of a page format in vertical writing mode).

1. Trim size and binding side (vertically set Japanese documents are bound on the right-hand side, and horizontally set documents are bound on the left-hand side. See [Fig.12].)

2. Principal text direction (vertical writing mode or horizontal writing mode).

3. Appearance of the kihon-hanmen and its position relative to the trim size.

4. Appearance of running heads and page numbers, and their positions relative to the trim size and kihon-hanmen.

[Fig.12]: Binding-side (bound on the right-hand side and bound on the left-hand side).

2.2.4 Elements of Kihon-hanmen

The kihon-hanmen is the hanmen style designed as the basis of a book. The following are the basic elements of the kihon-hanmen (see [Fig.13]).

[Fig.13]: Elements of kihon-hanmen. (Example in vertical writing mode.)

1. Character size and typeface name

2. Text direction (vertical writing mode or horizontal writing mode)

3. Number of columns and column gap when using multi-column format

4. Length of a line

5. Number of lines per page (number of lines per column when using multi-column format)

6. Line gap

2.2.5 Kihon-hanmen and Examples of Real Page Format

Below are several examples of how the basic page format is created, and how then various elements are placed on a real text page. (This and other aspects of how the various elements of the kihon-hanmen are arranged on each page are explained in 2.5 Page wise Arrangement of Kihon-hanmen Elements.)

1. Realm and position of headings

   To set a heading, first establish a rectangular space based on a number of lines in the kihon-hanmen. For example, a '3 line space' means (3 * the size of the character frame used for the kihon-hanmen + 2 * the line gap in the kihon-hanmen). (Details of this processing are defined in JIS X 4051, sec. 8.3.3.d). The heading text is usually set in the centre of the rectangular space in the block direction, and indented from the line head. The size of the indent is usually specified as a number of characters in the kihon-hanmen. For example, a '4 character indent' means (4 * the size of the character frames used for establishing the kihon-hanmen). (See the example at [Fig.14].)

   

   [Fig.14]: Layout example of a heading based on the line positions established by the kihon-hanmen.

   (note 1)

   Details of the different types of heading, creation of headings, methods for placing headings, etc. will be explained in 4.1 Headings in a future version of this document.

2. Size of illustrations

   In horizontal writing mode with two columns, for example, the width of illustrations should, if at all possible, be either the width of one kihon-hanmen column or the width of the kihon-hanmen (see [Fig.15]). The illustrations are usually set at the head or the foot of the page (see [Fig.15]).

   

   [Fig.15]: Example of illustrations in two columns, horizontal writing mode.

   Also, in vertical writing mode, for example with three columns, the height of illustrations should, if at all possible, be either the height of one kihon-hanmen column or the height of the kihon-hanmen (see [Fig.16]). The illustrations are usually set at the right side or left side of the kihon-hanmen (see [Fig.16]).

   

   [Fig.16]: Example of illustrations in three columns, vertical writing mode.

   (note 1)

   Details of illustration positioning will be explained in a subsequent version of this document.

3. Hanmen size for the table of contents

   The hanmen size for the table of contents of books is based on the size of the kihon-hanmen. There are many examples of tables of contents in vertical writing mode where the left-to-right size is identical to that of the kihon-hanmen, but the top-to-bottom size is a little bit smaller (see [Fig.17]).

   

   [Fig.17]: Example of the design of the table of contents (TOC) in vertical writing mode.

   (note 1)

   There are cases when a different hanmen than the kihon-hanmen is used for positioning of running heads and page numbers. This will be discussed in 2.6.2 Principles of Arrangement of Running Heads and Page Numbers (see [Fig.51]).

2.3.1 Directional Factors in Japanese Composition

Japanese composition has two text directions. One is vertical direction (vertical writing mode), the other is horizontal direction (horizontal writing mode).

(note 1)	Ideographic (cl-19), hiragana (cl-15) and katakana (cl-16) characters for Japanese composition have basically been designed to have a square character frame from the letterpress printing era on. Thus the same collection of printing types can be used in either vertical writing mode or horizontal writing mode, simply by changing the direction of text, (see [Fig.18]). There were some attempts to develop printing types designed exclusively for horizontal writing mode, but they were not widely accepted. [Fig.18]: Vertical writing mode and horizontal writing mode. (The arrows show the reading direction.)
(note 2)	There is little market data comparing the number of pages with vertical writing mode and horizontal writing mode, but it is said that both are almost the same.
(note 3)	For official (e.g. governmental) documentation, horizontal writing mode is recommended. Educational material (with the exception of certain topics) is mostly in horizontal writing mode. Readers of "mobile novels" are increasing, and it is expected that in the future horizontal writing mode will increase in this area as well. However, most of the large newspapers are written completely in vertical writing mode, and most of the large journals for ordinary readers are almost completely set in vertical writing mode. In addition, novels, which are the most widely read kind of book publication, are almost completely in vertical writing mode (some readers say that they cannot read a novel if it is not in vertical writing mode). Hence it can be expected that the importance of vertical writing mode for Japanese will not change for the time being.
(note 4)	There is usually only one direction for all text throughout a book, but there are cases where horizontal writing mode is used in certain parts of vertically composed books (see [Fig.19]). Tables, captions for illustrations, running heads, and page numbers are usually composed horizontally in a page with a vertical writing mode. [Fig.19]: Example of horizontal writing mode in parts of vertically set books.

2.3.2 Major Differences between Vertical Writing Mode and Horizontal Writing Mode

The following are major differences between vertical writing mode and horizontal writing mode.

1. Arrangement of characters, lines, columns and pages; direction of page progression.

   (note 1)

   The positioning of characters, lines and paragraphs in vertical and horizontal writing mode is defined in JIS X 4051, sec. 7.4.4.

   1. Vertical writing mode. See [Fig.20] for an example of vertical writing mode with two columns per page.

      

      [Fig.20]: Direction of arrangement of characters in vertical writing mode.

      1. Characters are arranged from top to bottom, lines are arranged from right to left.

      2. Columns are arranged from top to bottom. A book starts with the left (recto) side and progresses from right to left (see [Fig.21]).

         

         [Fig.21]: Progression of pages for a vertically set books.

   2. Horizontal composition. See [Fig.22] for an example of horizontal text layout with two-columns per page.

      

      [Fig.22]: Direction of arrangement of characters in horizontal writing mode.

      1. Characters are arranged from left to right, and lines are arranged from top to bottom.

      2. Columns are arranged from left to right. A book starts with the right (recto) side and progresses from left to right (see [Fig.23]).

         

         [Fig.23]: Progression of pages for a horizontally set book.

2. Orientation of Latin alphanumeric characters in a line.

   1. There are three ways to arrange Latin alphanumerics in vertical writing mode:

      1. One by one with the same normal orientation as that of Japanese characters. This is usually applied to one-letter alphanumerics or capitalized abbreviations (see [Fig.24]).

         

         [Fig.24]: Arrangement of alphanumerics in vertical writing mode - normal orientation.

         (note 1)

         The alphanumeric characters used for this arrangement have different typographic features than those with proportional width used for Western text. They are of fixed-width and full-width design, and have been used this way since the letterpress printing era.

      2. Rotated 90 degrees clockwise. This is usually applied to English words or sentences (see [Fig.25]).

         

         [Fig.25]: Arrangement of alphanumerics in vertical writing mode - rotated 90 degrees clockwise.

         (note 1)

         In [Fig.25], there are spaces before and after the character frame for the Western word "editor". These spaces are necessary for composition of mixed Japanese and Western text, and details will be provided in a later section.

      3. Set horizontally without changing orientation (called tate-chu-yoko, which means horizontal-in-vertical composition) (see [Fig.26]). This is usually applied to two-digit numbers (see JIS X 4051, sec. 4.8 for the definition).

         

         [Fig.26]: Arrangement of alphanumerics in vertical writing mode - tate-chu-yoko.

   2. In horizontal writing mode there is only one way of arranging alphanumerics, i.e. normal orientation.

3. Arrangement of tables and/or illustrations rotated 90 degrees clockwise or counter-clockwise for reasons of size. (This processing is defined in JIS X 4051, sec. 7.3.).

   1. In vertical writing mode, align the top of tables/illustrations to the right of the page (see [Fig.27]).

      

      [Fig.27]: Example of arrangement of a table rotated 90 degrees clockwise in vertical writing mode.

   2. In horizontal writing mode, align the top of tables/illustrations to the left of the page (see [Fig.28]).

      

      [Fig.28]: Example of arrangement of a table rotated 90 degrees counterclockwise in horizontal writing mode.

      (note 1)

      The orientation is chosen to minimize interference with the overall reading flow of the book.

4. Arrangement of an incomplete number of lines on a multi-column format page due to new recto, page break or other reasons. (The processing of new recto and page break is defined in JIS X 4051, sec. 8.1.1.).

   1. In vertical writing mode, just finish the line where it ends ("nariyuki"). The number of lines in each column is not uniform (see [Fig.29]).

      

      [Fig.29]: How to process incomplete number of lines on a multi-column format page (vertically set book).

   2. In horizontal writing mode, re-arrange columns so that each column has the same number of lines. In case the number of lines is not divisible by the number of columns, add the smallest number to make it divisible and re-arrange columns using the quotient as the number of lines so that only the last column shall have the incomplete number of lines (see [Fig.30]).

      

      [Fig.30]: How to process incomplete number of lines on a multi-column format page (horizontally set book).

      (note 1)

      Neither horizontal nor vertical balance of column arrangement would break the stability of vertical page layout very much, while horizontal balance of column arrangement is determinant for horizontal page layout. In vertical text it doesn't matter too much whether columns are balanced or not. For horizontally set text it is best to balance columns wherever possible.

2.4.1 Procedure for Defining the Kihon-hanmen

In Japanese composition, first the size of the kihon-hanmen is defined, using the square character frames of characters in solid setting. Taking this as a base, the position of the kihon-hanmen with regards to the trim size is then specified. The following are procedures for determining the size and position of the kihon-hanmen (see [Fig.31]).

1. Specifying the dimensions of the kihon-hanmen.

   1. For a document with a single column per page, specify the character size, the line length (the number of characters per line), the number of lines per page, and the line gap.

   2. For a document with multiple columns per page, specify the character size, the line length (the number of characters per line), the number of lines per column, the line-gap, and the number of columns and the column gap.

      

      [Fig.31]: Procedures to determine the size and position of the kihon-hanmen, step 1.

2. Determining the position of the kihon-hanmen relative to the trim size.

   There are various alternative methods for specifying the position of the kihon-hanmen relative to the trim size (see [Fig.32]):

   1. Position vertically by centering the kihon-hanmen. Position horizontally by centering the kihon-hanmen.

   2. Position vertically by specifying the space at the head (for horizontal writing mode) or the space at the foot (for vertical writing mode). Position horizontally by centering the kihon-hanmen.

   3. Position vertically by centering the kihon-hanmen. Position horizontally by specifying the space of the gutter.

   4. Position vertically by specifying the space at the head (for horizontal writing mode) or the space at the foot (for vertical writing mode). Position horizontally by specifying the space of the gutter.

   

   [Fig.32]: Procedures to determine the size and position of the kihon-hanmen, step 2.

   (note 1)	In most cases the kihon-hanmen is set at the horizontal and vertical center of the trim size, which should be the default positioning, but depending on the dimensions of the kihon-hanmen there may be cases where the default needs to be changed; for example, by moving the kihon-hanmen up, down, to the left or to the right of the default position.
   (note 2)	It is technically possible to determine the dimensions of the kihon-hanmen by specifying the trim size and margins of all sides, but this method is not common in the tradition of Japanese composition. If this is the only way an implementation allows, the margins of each side need to be determined beforehand in relation to the dimensions of the kihon-hanmen and its position in the trim size.

2.4.2 Considerations in Designing the Kihon-hanmen

The following are considerations to take into account when designing the kihon-hanmen. (This topic is not about processing, but rather an explanation of design preferences. The definition of kihon-hanmen is given in JIS X 4051, sec. 7.4.1.)

1. Trim size and margins. It would be best if the shape of the kihon-hanmen could be made similar to that of the trim size.

2. Character size. Generally 9 point (about 3.2mm) type is common. Except for specialized publications such as dictionaries, the minimum size of type is 8 point (about 2.8mm).

   (note 1)

   In Western text layout, 10 point (about 3.5mm) or 12 point (about 4.2mm) type is common. This is mainly because of a difference in design principles between Japanese and Latin characters.

3. Line length should be multiples of the character size (see [Fig.33]).

   

   [Fig.33]: Line length should be multiples of the character size.

   (note 1)

   There are basically two reasons why line length should be multiples of the character size. For Japanese composition, all line lengths except that of the last line of the paragraph should, in principle, be the same. In principle, for printing, Japanese characters like ideographic (cl-19), hiragana (cl-15) and katakana (cl-16) characters are uniformly designed in the same square character frame and they are set solid (no extra space between adjacent character frames).

   (note 2)

   The best line length (number of characters per line) is around 52 characters, maximum, in vertical writing mode, and 40 characters, maximum, in horizontal writing mode. If the trim size would take lines beyond the recommended length, consider using a multi-column format and making the line length shorter.

4. Use the same amount of line gap throughout the book, except for special cases. The size of the kihon-hanmen in the block direction is specified using the number of lines and the size of the line-gap.

   (note 1)	In Japanese composition, there are cases where ruby or emphasis (emphasis dots, bousen, underlines, etc.) are inserted between lines. In such cases the line gap is not changed but is kept constant (see [Fig.34]). It is also possible to insert reference marks to notes between lines within the main text. This case is handled in the same manner. If these elements are likely to occur in text, the line gap established during the kihon-hanmen design needs to be of an adequate size to accommodate them. Further explanations about the placement of ruby will be given in 3 Line Composition. [Fig.34]: Inserting ruby or other items between lines.
   (note 2)	Warichu (inline cutting note) juts into the line gap on either side of a line. The basic line gap isn't changed where warichu occurs (the line gap between warichu itself and the adjacent lines looks narrower than for the rest of the line), so when warichu is likely to occur in text, the line gap for the kihon-hanmen may be set slightly larger than normal to accommodate it. The same is true for tate-chu-yoko or subscript and superscript (ornament characters). Further explanation of the placement of warichu and other items is provided in 3 Line Composition. [Fig.35]: Example of inter-line processing with warichu between lines.
   (note 3)	It is common that the line gap for the kihon-hanmen is set to a value between a half em space and the one em space of the character frame used for the kihon-hanmen. A half em space can be chosen in cases where the line length is short, but a one em space or close to it is more appropriate when the line length is longer than 35 characters.
   (note 4)	Unless ruby or other design elements are placed in the space between lines (e.g. for books such as classics, with many annotations), there is no need to make the line-gap larger than full-width, since this would decrease legibility.
   (note 5)	It is said that the standard line-gap in Western text layout is a one third em space, which is smaller than that in Japanese composition. This difference again comes from the different approach to the design of Latin and Japanese characters.
   (note 6)	There is another method of specifying the kihon-hanmen that uses line feeds rather than line gaps. Line feed is the distance between two adjacent lines measured from their reference points (see [Fig.36]). The reference point differs from implementation to implementation, however, in vertical writing mode the horizontal center of the character frame is usually used, and with horizontal writing mode, the vertical center of the character frame is used. When the character size is the same for every character, the following calculation is used: line feed = character size/2 + line gap + character size/2 i.e. character size + line gap line gap = line feed - character size [Fig.36]: Specifying kihon-hanmen with line feed. The size of the kihon-hanmen in this case can be calculated by following method: Vertical writing mode with one column Width of kihon-hanmen = character size * number of lines per page + line gap * (number of lines per page -1) e.g. 9 point * 18 lines + 8 point (18 lines -1) = 298 point Height of kihon-hanmen = character size * number of characters per line e.g. 9 point * 52 characters = 468 points Vertical writing mode with multi columns Width of kihon-hanmen = character size * number of lines per column + line gap * (number of lines per column -1) e.g. 9 point * 21 lines + 6 point * (21 lines -1) = 309 points Height of kihon-hanmen = (character size * number of characters per line) * number of columns + column gap * (number of columns - 1) e.g. 9 point * 25 characters + 18 point * (2-1) = 468 points Horizontal writing mode with one column Width of kihon-hanmen = character size * number of characters per line e.g. 9 point * 35 characters = 315 points Height of kihon-hanmen = character size * number of lines per page + line gap * (number of lines per page -1) e.g. 9 point * 28 lines + 8 point * (28 lines - 1) = 468 points Horizontal writing mode with multi columns Width of kihon-hanmen = (character size * number of characters per line) * number of columns + column gap * (number of columns - 1) e.g. (8 point * 19 characters) * 2 + 16 point * (2-1) = 320 points Height of kihon-hanmen = character size * number of lines per column + line gap * (number of lines per column - 1) e.g. 8 point * 40 lines + 4 point * (40 lines -1) = 476 points

2.5.1 Examples of Items Jutting Out of the Kihon-hanmen

The various elements of a page should remain inside the boundaries of the kihon-hanmen. However, there are exceptions such as the following:

1. Ruby or emphasis marks (bousen, emphasis dots, etc.) at the before edge of the hanmen, are placed outside the hanmen (see [Fig.37]). The same applies in cases where ruby, underline, etc. appear beyond the after edge of the hanmen. Like the handling of exceptions mentioned below, the purpose here is to preserve the line positions established for the kihon-hanmen. This technique can also be used for reference marks associated with lines of text.

   

   [Fig.37]: Example of ruby annotation placed outside of the kihon-hanmen.

2. When there are inline elements whose dimensions extend beyond the before edge and the after edge of a line of characters as determined by the kihon-hanmen, and when those elements appear in the first or last line of the hanmen, the parts that jut out beyond the regular line of characters also jut out of the hanmen area. For example, this is the case when the width of a sequence of characters which are set to tate-chu-yoko is wider than the characters set for the kihon-hanmen. In addition, warichu (inline cutting note) or subscript and superscript (ornament characters) are handled in the same way. (The processing rules for this item and the previous item are defined in JIS X 4051, sec. 12.1.1.)

3. Line adjustment by hanging punctuation is only necessary for full stops (cl-06) and commas (cl-07) when they would otherwise need to be wrapped to the line head. The character is placed so that it touches the hanmen at the line end (see [Fig.38]). (Hanging punctuation is not defined in JIS X 4051, but there is an explanation in sec. 8.1, c.)

   

   [Fig.38]: Example of IDEOGRAPHIC COMMA and IDEOGRAPHIC FULL STOP placed below the kihon-hanmen.

   (note 1)	Line adjustment by hanging punctuation is a way of reducing the processing cost of line adjustment by reducing the need to change inter-character space.
   (note 2)	A lot of books apply hanging punctuation.

4. Illustrations and tables are normally placed inside the area defined by the kihon-hanmen. However, there may also be cases in which a particular illustration or table juts outside the kihon-hanmen.

   1. Cases in which it is necessary to make the illustration or table larger than the kihon-hanmen, because reducing its size would make it unreadable.

   2. For the sake of visual effect, the illustration may bleed into the complete paper area. This is not often used in books, but is often used in magazines (see [Fig.39]).

      

      [Fig.39]: Example of bleeds.

5. Magazines may place the captions of illustrations outside the column area or in the column gap. (Some people regard this as bad style.)

2.5.2 Line Positioning based on the Kihon-hanmen Design

In principle, pagewise positioning of lines relies on the line positions established for the kihon-hanmen. This holds for lines with ruby or emphasis dots as shown in [Fig.34]. Even when lines contain characters that are smaller than the character size used for the kihon-hanmen (as shown in [Fig.40]), the line positions used for the kihon-hanmen continue to be used as the basic guide lines. This is so that following lines with normal-sized characters still naturally fall into the line positions established for the kihon-hanmen.

[Fig.40]: Positioning of lines with a smaller size of text.

The following are exceptions when handling line position:

1. When inserting more than one illustration or table item in horizontal writing mode, assuming that there is no text to the left or right of the items, the items may either slip off the lines established for the kihon-hanmen (see [Fig.41]), or stick to the lines (see [Fig.42]). The former approach is used, whenever possible, to achieve inter-character spacing before and after illustrations or tables . (This method is often used in books.) (This processing method is defined in JIS X 4051, sec. 10.3.2., d.)

   

   [Fig.41]: Positioning of lines with multiple illustrations - 1.
   

   [Fig.42]: Positioning of lines with multiple illustrations - 2.

2. The size of characters in endnotes inserted between paragraphs or those in footnotes at the bottom of the page (in horizontal writing mode) is smaller than the character size established for the kihon-hanmen. As a result, the character size and line gap are also smaller, and so the line positions are no longer identical to those established for the kihon-hanmen. As an example, [Fig.43] shows the position of an endnote between paragraphs in vertical writing mode. (The processing of endnotes is defined in JIS X 4051, sec. 9.3, and the processing of footnotes in sec. 9.4.)

   

   [Fig.43]: Positioning of an endnote in vertical writing mode.

3. As mentioned above, the position of a heading may not be identical to the lines established for the kihon-hanmen. Nevertheless, in the block direction, headings base their alignment on the line positions established for the kihon-hanmen (see [Fig.14]).

2.5.3 Character Positioning based on Kihon-hanmen Design

In principle, the characters in each line follow the solid setting positions of characters established for the kihon-hanmen. However, as already shown in some of the previous figures, there are examples where this is not the case. Such cases are rather common, and here we will show some prototypical examples (details will be given in 3 Line Composition).

1. When 9pt is the character size used to establish the kihon-hanmen, characters smaller than 9pt may be inserted in part of a line (see [Fig.40]). In such cases, the parts set at 9pt and any parts set at a smaller, say, 8pt size both use solid setting, with character frames at the respective sizes for each part.

2. In cases where proportional Latin letters are rotated 90 degrees clockwise (see [Fig.25]), the proportional letters are placed according to their proportional widths. Hence, they do not fit to the character positions established for the kihon-hanmen (see [Fig.44]). Japanese letters following the Latin letters consequently slip away from the default positions as well.

   

   [Fig.44]: Positioning of a mixture of Western and Japanese letters in a line.

3. There are several methods for positioning opening brackets (cl-01) at the beginning of a line (details are explained in 3.1.5 Positioning of Opening Brackets at Line Head). Because an opening bracket is not a full-width character, in cases where the indentation of the first line of a paragraph is a one em space, or if the tentsuki position is used for the bracket (that is, there is no space at the line head), the character following the bracket will be in a position which does not fit to the character positions established for the kihon-hanmen (see [Fig.45]). However, the adaptations made during the alignment of line ends will ensure that the character at the end of a line is at a position that fits with the kihon-hanmen.

   

   [Fig.45]: Example of positioning of characters off the kihon-hanmen position due to opening brackets at the line head.

4. 3 Line Composition explains that full stops (cl-06), commas (cl-07), opening brackets (cl-01) and closing brackets (cl-02) are half-width. If these punctuation marks and brackets are adjacent to ideographic (cl-19), katakana (cl-16) or hiragana (cl-15) characters, in principle there should be a half em space before or after the punctuation mark or brackets, so that these occupy in effect a full-width size. However, if they are adjacent to other punctuation marks or brackets, the half em space is not used. This is done to improve the visual appearance. In such cases, the character positions are different than the positions established when defining the kihon-hanmen (see [Fig.46]).

   

   [Fig.46]: Example of lines with consecutive punctuation marks.

5. 3 Line Composition explains the principle that closing brackets (cl-02), full stops (cl-06) and commas (cl-07) should not be placed at the line head. If by simple sequential placement these characters would appear at the line head or at the line end, some kind of adjustment becomes necessary. A similar adjustment is required for characters that should not be placed at the end of a line, such as opening brackets (cl-01). As a result of such adjustment, it may happen that other characters are placed at positions which are different from those established for the kihon-hanmen.

   

   [Fig.47]: Example of line adjustment to avoid those characters which shall not start and end a line.

2.6.1 Positioning of Running Heads and Page Numbers

Typical positions of running heads and page numbers for vertically set books with double running heads (see 2.6.3 Ways of Arranging Running Heads and Page Numbers) are as shown in [Fig.48].

[Fig.48]: Typical positioning of running heads and page numbers for vertically set books with double running heads.

Typical positions of running heads and page numbers for horizontally set books with double running heads (see 2.6.3 Ways of Arranging Running Heads and Page Numbers) are as shown in [Fig.49].

[Fig.49]: Typical positioning of running heads and page numbers for horizontally set books with double running heads.

In principle, positions of running heads and page numbers should be specified relative to the kihon-hanmen, not with absolute coordinates in the trim size. (Positioning of running heads is defined in JIS X 4051, sec. 7.6.4. Positioning of page numbers is defined in JIS X 4051, sec. 7.5.4.)

[Fig.50]: Positioning of a running head (vertical writing mode).

The following recommendations should be taken into account when positioning running heads and page numbers with reference to the kihon-hanmen.

1. When positioning horizontal running heads and page numbers with reference to the kihon-hanmen in vertically set books, the amount of vertical space between the edge of the kihon-hanmen and the running head is a one em space as established for the kihon-hanmen. If the kihon-hanmen of the book is horizontally set, take more vertical space than the character size in the kihon-hanmen.

2. Regardless of the direction of text in the kihon-hanmen of a book, horizontal running heads and page numbers on the left page should be aligned either at the left edge of the kihon-hanmen or one em space to the right of the left edge. On the right page, the tail of the running heads or page numbers should be aligned either at the right edge of the kihon-hanmen or one full-width space to left of the right edge.

3. Regardless of the direction of text in a book, when arranging running heads and page numbers together on the same horizontal line, the space between the running head and the page number should be double or one and a half times the character size of the running head. On the left page, the page number should be set at the left side and the running head should be set at the right side. On right-hand pages, the page number should be set at the right side and the running head should be set at the left side. The exact positions of the page numbers are given by the instructions above (see b).

4. When positioning running heads and page numbers vertically to the fore-edge of the kihon-hanmen in a vertically set book (see spread (e) in [Fig.48], for example), the minimum horizontal distance from the kihon-hanmen should be the same as that of the line gap of the kihon-hanmen. The top of the running head should be positioned approximately four kihon-hanmen characters below the head, and the bottom of the page numbers should be positioned approximately five kihon-hanmen characters above the foot.

   (note 1)

   In general, ideographic numerals (一二三四五六七八九〇) are used for vertically set page numbers, and European numerals for horizontal pagination. When using independent pagination for the front matter, small Roman numerals are used for horizontal pagination.

2.6.2 Principles of Arrangement of Running Heads and Page Numbers

Positioning of all running heads and page numbers in the same book should be consistent.

(note 1)

Even on a page with a text area smaller in size than that of the kihon-hanmen, such as for a table of contents or index, positioning of the running head and page number relative to the trim size will remain the same. Therefore, the positioning of running heads and page numbers relative to those areas smaller than the kihon-hanmen is different. [Fig.51] below demonstrates the respective positions of the hanmen for a table of contents and running heads or page numbers. As shown in [Fig.17], this hanmen is smaller than the kihon-hanmen. [Fig.52] demonstrates the related positions of running heads and page numbers and the hanmen of indexes. These hanmen are not only 4 points smaller at the left and right, but also 5 points smaller at the top and bottom. [Fig.51]: Positioning of running heads and page numbers on TOC pages for which the hanmen is smaller in size than the kihon-hanmen. [Fig.52]: Positioning of running heads and page numbers on index pages for which hanmen is smaller in size than the kihon-hanmen.

Because the start of a page will be on the recto side, the right-hand page of a spread in a vertically set book is always an even page and the left-hand page is always an odd page (see [Fig.53]). Likewise, the left-hand page of a spread in a horizontally set book is always an even page and the right-hand page is always an odd page (see [Fig.54]).

[Fig.53]: Page numbers on a spread in a vertically set book.

[Fig.54]: Page numbers on a spread in a horizontally set book.

2.6.3 Ways of Arranging Running Heads and Page Numbers

There are two ways to arrange running heads. One is the single running head method and the other is the double running head method. (Arrangement of running heads is defined in JIS X 4051, sec. 7.6.2. Page Numbers are defined in sec. 7.5.2.).

• Double running head method: Place running heads on both even pages and odd pages (see [Fig.55]).

• Single running head method: Place running heads only on odd pages (see [Fig.56]).

[Fig.55]: Double running head method.

[Fig.56]: Single running head method.

1. In the double running head method, a higher-level title, such as that of the chapter or book, is used for the running heads on the even pages, and a lower-level title, such as that for a section, on the odd pages. Where there are no differing levels of titles, such as on the page containing the table of contents, the same running head is used on both even and odd pages.

   (note 1)

   Which information is used for the running heads depends on the content of the book. Given that the main purpose of running heads is to signpost to readers what is written on each page, or the content of the current page, it does not make much sense to use the book title for the running head. The most common approach for a book with three levels of headings, such as chapter, section and subsection, is to use the highest level heading (i.e. chapter) and the second level heading (i.e. section).

2. In the single running head method, one of the headings between the top and third levels is used.

3. In principle, the contents of running heads will be the same as those of headings with the following differences:

   1. Numbers and words in Latin alphanumeric characters in vertically set headings in vertically set books should be changed to horizontal notation for horizontally set running heads (see 2.3.2 Major Differences between Vertical Writing Mode and Horizontal Writing Mode).

   2. If headings are too long, they should be made shorter by paraphrasing them in fewer characters. Running heads with too many characters will not look good.

   3. For certain publications, such as a collection of monographs, the names of authors may be added in parentheses at the end of the running head.

4. In principle, the text direction of running heads and page numbers should be the same as that of the kihon-hanmen. For vertically set books, however, it is more common to set running heads and page numbers horizontally.

5. In principle, for the single running head method running heads are printed on all odd pages, and for the double running head method on all even and odd pages. However, for the sake of appearance, running heads may be omitted as follows:

   1. Pages on which running heads should be hidden:

      1. Naka-tobira and han-tobira.

      2. Pages where a running head overlaps with other elements such as illustrations.

      3. Blank pages.

   2. Pages on which running heads may be hidden:

      1. Pages where a figure or a table is positioned adjacent to the running head.

      2. Pages with a heading right after a new recto or new page.

6. In principle, page numbers are printed on all pages. However, for the sake of appearance, they may be omitted as follows:

   1. Pages on which page numbers should be hidden:

      1. Pages on which a illustration or a table is positioned adjacent to the page number.

      2. Blank pages.

   2. Pages on which page numbers may be hidden:

      1. Divisional title and simplified divisional title pages.

      2. Pages in horizontally set books with a page number placed in the margin at the top of the page, and with a heading at the beginning of a new recto or new page. (In this case, it is also possible to move the page numbers to the center of the margin at the foot of the page.)

   (note 1)

   Pages are not counted in cases such as the following: If a different type or color of paper is used for the main title page, if a frontispiece is inserted in the opening page of a book; or if an illustration of the enclosure or a divisional title is present in the main text.

7. There are two types of page numbering. "Continuous pagination" means that page numbers continue throughout the whole book. "Independent pagination" means that page numbers start from "1" separately at beginning of the front matter and back matter. There is also, for example in manuals, the method of starting each chapter from page number "1". (In such cases, it is common that the name of the chapter is added as a prefix before the page number.)

   (note 1)

   If the front matter and the main text have different page numbers, each starts with page number "1". In this case, it is common to use Roman numerals for the pages of the front matter, in order to distinguish them from the main text.

   (note 2)

   For vertically set books with indexes in horizontal writing mode, the following methods are available. Reverse pagination. The index reads from the end of the book, and page numbers are added starting with "1" from the end of the book and flow in the same order as the index. Continuous pagination. The index reads from the end of the book, but page numbers start with "1" and flow in the same order as the book. (The index pages flow in the reverse order to the page numbers.) Both reverse pagination and continuous pagination. In this case, the page numbers for continuous pagination are in the same position as the page numbers of the main text, and page numbers in reverse pagination are in a different position (for example, if serial pagination is in the foot of the page, reverse pagination is in the head). Often other methods are applied to distinguish the different paginations. For example, Arabic numbers are used for both continuous pagination and reverse pagination, but for reverse pagination, brackets are added around the numbers.

3.1.1 Differences in Vertical and Horizontal Composition in Use of Punctuation Marks

There are some punctuation marks that are used uniquely in either vertical writing mode or horizontal writing mode. In this document, characters and symbols are treated as members of a character class, classified by their behavior for composition. Each class name is followed by class id, such as opening brackets (cl-01). Details are explained in 3.9 About Character Classes. The following are some typical examples:

1. Full stops (cl-06) and commas (cl-07)

   1. In vertical writing mode, IDEOGRAPHIC FULL STOP "。" and IDEOGRAPHIC COMMA "、" are used for full stops (cl-06) and commas (cl-07).

   2. In horizontal writing mode, there are three conventions in choice of symbols for full stops (cl-06) and commas (cl-07):

      1. Using COMMA "，" and FULL STOP "．" (see [Fig.57]).

         

         [Fig.57]: Example text using COMMA and FULL STOP.

      2. Using COMMA "，" and IDEOGRAPHIC FULL STOP "。" (see [Fig.58]).

         

         [Fig.58]: Example text using COMMA and IDEOGRAPHIC FULL STOP.

      3. Using IDEOGRAPHIC COMMA "、" and IDEOGRAPHIC FULL STOP "。" (see [Fig.59]).

         

         [Fig.59]: Example text using IDEOGRAPHIC COMMA and IDEOGRAPHIC FULL STOP.

      (note 1)

      In horizontal writing mode, there are many cases of composition that mixes Japanese and Western text. The convention shown in (i) is a way to apply the same comma and full stop to both Western and Japanese texts for consistency, and is common in books on science and technology. The convention shown in (ii) was invented because in (i) FULL STOP "." appears too small for Japanese texts and using IDEOGRAPHIC FULL STOP "。" for period looks better. This convention has been adopted for Japanese official publications. (In the past, COMMA "," and FULL STOP "." were used for some official publications.)

2. LEFT CORNER BRACKET "「", RIGHT CORNER BRACKET "」", LEFT DOUBLE QUOTATION MARK "“" and RIGHT DOUBLE QUOTATION MARK "”"

   1. In vertical writing mode, LEFT CORNER BRACKET "「" and RIGHT CORNER BRACKET "」" are used for quotations (see [Fig.60]).

      

      [Fig.60]: Examples of quoted text using LEFT CORNER BRACKET and RIGHT CORNER BRACKET.

   2. In horizontal writing mode, pairs of LEFT DOUBLE QUOTATION MARK "“" and RIGHT DOUBLE QUOTATION MARK "”" or pairs of LEFT SINGLE QUOTATION MARK "‘" and RIGHT SINGLE QUOTATION MARK "’" may be used in place of LEFT CORNER BRACKET "「" and RIGHT CORNER BRACKET "」" (see [Fig.61]).

      

      [Fig.61]: Examples of quoted text using LEFT DOUBLE QUOTATION MARK and RIGHT DOUBLE QUOTATION MARK.

      (note 1)	This is because LEFT CORNER BRACKET "「" and (especially) RIGHT CORNER BRACKET "」" may not look good in horizontal writing mode, but adoption of corner brackets for horizontal writing mode seems to be increasing.
      (note 2)	Though LOW DOUBLE PRIME QUOTATION MARK "〟" and REVERSED DOUBLE PRIME QUOTATION MARK "〝" are similar to double quotation marks in appearance (see [Fig.62]) they are exclusively used for vertical writing mode and are not to be used in horizontal writing mode. [Fig.62]: Examples of quoted text using LOW DOUBLE PRIME QUOTATION MARK and REVERSED DOUBLE PRIME QUOTATION MARK.
      (note 3)	LEFT DOUBLE QUOTATION MARK "“" and RIGHT DOUBLE QUOTATION MARK "”" are exclusively for horizontal writing mode and not to be used in vertical writing mode. Also, LEFT SINGLE QUOTATION MARK "‘" and RIGHT SINGLE QUOTATION MARK "’" are exclusively for horizontal writing mode and not to be used in vertical writing mode. However, in vertical writing mode, when Western characters (cl-27) are composed rotated 90 degrees clockwise, these quotation marks are sometimes used.

3. LEFT SQUARE BRACKET "［", RIGHT SQUARE BRACKET "］", LEFT TORTOISE SHELL BRACKET "〔" and RIGHT TORTOISE SHELL BRACKET "〕"

   LEFT TORTOISE SHELL BRACKET "〔" and RIGHT TORTOISE SHELL BRACKET "〕" are vertical variants of LEFT SQUARE BRACKET "［" and RIGHT SQUARE BRACKET "］", which are used in horizontal writing mode. Square brackets should be used in horizontal writing mode except for special cases.

[Fig.63]: KATAKANA-HIRAGANA PROLONGED SOUND MARK for vertical and horizontal writing modes.

3.1.2 Positioning of Punctuation Marks (Commas, Periods and Brackets)

The positioning of punctuation marks (commas, periods and brackets) in a line proceeds as follows.

The character advance of commas (cl-07), full stops (cl-06), opening brackets (cl-01), closing brackets (cl-02) and middle dots (cl-05) is half-width (half em). But when those punctuation marks are placed side-by-side with ideographic (cl-19), hiragana (cl-15), or katakana (cl-16) characters, in principle, a given amount of space will be inserted before or after the symbols, which makes them appear as if they were intrinsically full-width (one em) (see [Fig.64]). Space is inserted before and after middle dots (cl-05). This principle makes the symbols consistent with ideographic (cl-19), hiragana (cl-15) and katakana (cl-16) characters in character width, and at the same time the space for punctuation helps to make the organization of text clearer. The space added before or after punctuation marks is subject, in principle, to line adjustment and may eventually be removed, except for that added after full stops (cl-06). (Details of line adjustment are discussed in 3.8 Line Adjustment).

1. After commas (cl-07), a half em space is added, in principle.

2. After full stops (cl-06), in the middle of a line, a half em space is added. At the end of a line, a half em space is added, in principle.

3. Before opening brackets (cl-01), a half em space is added, in principle.

4. After closing brackets (cl-02), a half em space is added, in principle.

5. Before and after middle dots (cl-05), a quarter em space is added, in principle.

[Fig.64]: Character widths of commas, periods, and the space appended before and/or after the symbols.

[Fig.65]: Positioning of parentheses and brackets. (The left-hand side shows an example of setting them solid.)

3.1.3 Exceptional Positioning of Ideographic Comma and Katakana Middle Dot

The space usually added after IDEOGRAPHIC COMMA "、" and the space before and after KATAKANA MIDDLE DOT "・" are omitted, in principle, for cosmetic reasons in the following cases.

1. In vertical writing mode, ideographic numerals and IDEOGRAPHIC COMMA "、" used as a decimal separator are set solid (as in the right line in [Fig.66]).

   

   [Fig.66]: Example of exceptional positioning of the IDEOGRAPHIC COMMA.

   (note 1)

   In vertical writing mode, ideographic digits used with IDEOGRAPHIC COMMA "、" to represent an approximate number are expected to be set solid too (as in the right line in [Fig.67]). [Fig.67]: Example of the positioning of IDEOGRAPHIC COMMA with ideographic digits to represent an approximate number.

2. Ideographic digits and KATAKANA MIDDLE DOT "・" representing a decimal point are set solid (as in the right line in [Fig.68]). In vertical writing mode, when KATAKANA MIDDLE DOT "・" is used as a member of unit symbols (cl-25) in unit symbols, grouped numerals (cl-24), and Western characters (cl-27) in mathematical and chemical formulae, before and after KATAKANA MIDDLE DOT "・" is set solid.

   

   [Fig.68]: Example of the exceptional positioning of KATAKANA MIDDLE DOT.

3.1.4 Positioning of Consecutive Opening Brackets, Closing Brackets, Commas, Full Stops and Middle Dots

In cases where multiple punctuation marks, such as opening brackets (cl-01), closing brackets (cl-02), commas (cl-07), full stops (cl-06) and middle dots (cl-05), come one after the other, the following space adjustments are made for aesthetic reasons (see [Fig.69]). Note also that the half em and quarter em spaces added before or after punctuation marks, including the half em space after full stops (cl-06) appearing in the middle of a line, are subject, in principle, to line adjustment and may eventually be removed, except for those added after full stops (cl-06). (See 3.8 Line Adjustment for more about line adjustment.) For more information about the positioning of closing brackets (cl-02), full stops (cl-06), commas (cl-07) and middle dots (cl-05) at line end, see 3.1.9 Positioning of Closing Brackets, Full Stops, Commas and Middle Dots at Line End.

1. When closing brackets (cl-02) come immediately after commas (cl-07) or full stops (cl-06), remove the default half em space between them and, in principle, add a half em space after the closing brackets (see [Fig.69] (1)).

2. When commas (cl-07) come immediately after closing brackets (cl-02), remove the default half em space between them and, in principle, add a half em space after the comma. When full stops (cl-06) come immediately after closing brackets (cl-02), remove the default half em space between them and, in middle of a line, add a half em space after the full stop; at the end of a line, in principle, add a half em space after the full stop (see [Fig.69] (2)).

3. When opening brackets (cl-01) come immediately after commas (cl-07), in principle, add a half em space between them (see [Fig.69] (3)). When opening brackets (cl-01) come immediately after full stops (cl-06) in the middle of a line, add a half em space between them. Note that when full stops (cl-06) come in the bottom of lines, in principle, insert a half space after full stops (cl-06).

4. When opening brackets (cl-01) come immediately after closing brackets (cl-02), in principle, add a half em space between them (see [Fig.69] (4)).

5. When opening brackets (cl-01) come immediately after other opening brackets (cl-01), set them solid and, in principle, add a half em space before the first one (see [Fig.69] (5)).

6. When closing brackets (cl-02) come immediately after other closing brackets (cl-02), set them solid and, in principle, add a half em space after the last closing bracket (see [Fig.69] (6)).

7. When middle dots (cl-05) come immediately after closing brackets (cl-02), in principle, add a quarter em space before the following middle dot (see [Fig.69] (7)).

8. When opening brackets (cl-01) come immediately after middle dots (cl-05), in principle, add a quarter em space after the preceding middle dot (see [Fig.69] (7)).

[Fig.69]: Examples of line adjustment with multiple opening brackets, closing brackets, commas, full stops or middle dots.

The line adjustment rules shown above have been established because the default half em space before or after consecutive punctuation marks, or quarter em space before and after them, makes the line look sparse and doesn't make the line appear well-proportioned (see [Fig.70]).

[Fig.70]: Examples of bad line composition with unadjusted spaces between multiple opening brackets, closing brackets, commas, full stops or middle dots.

3.1.5 Positioning of Opening Brackets at Line Head

When starting a new line with opening brackets (cl-01) there are some patterns as shown in [Fig.71]. Note that the amount of line indent after the line feed (the first line indent of a new paragraph) is assumed to be a one em space across all the patterns.

1. The first line indent after the line feed is set full-width (one em) and the next line after the first line break starts with no space (so-called tentsuki) (see [Fig.71] (1)).

2. The first line indent after the line feed is set one and a half em and the next line indent after the first line break is set to a half em (see [Fig.71] (2)).

3. The first line indent after the line feed is set at a half em and the next line after the first line break is set tentsuki (see [Fig.71] (3)).

[Fig.71]: Examples of positioning of opening brackets at line head.

3.1.6 Positioning of Dividing Punctuation Marks (Question Mark and Exclamation Mark) and Hyphens

The dividing punctuation marks (cl-04) (QUESTION MARK "?" and EXCLAMATION MARK "!") should be full-width, and they are typeset as follows.

1. Basically, add no space before dividing punctuation marks (cl-04) at the end of a sentence and add a one em space after them (see [Fig.73]). However when a closing bracket (cl-02) follows right after the dividing punctuation mark, add no space after the dividing punctuation mark and add a half em space after the closing bracket (see [Fig.73]).

   

   [Fig.73]: Positioning of dividing punctuation marks (Examples in vertical writing mode).

   (note 1)	Many implementations use full-width ideographic space (cl-14) for the one em space appended after dividing punctuation marks (cl-04).
   (note 2)	No full stops (cl-06) should be appended after dividing punctuation marks (cl-04) at the end of a sentence.
   (note 3)	There are some cases where dividing punctuation marks (cl-04) are used in the middle of a sentence, not at the end. In those cases, either add no space or a quarter em space before and after the dividing punctuation mark see [Fig.74]). [Fig.74]: Examples of positioning of dividing punctuation marks in the middle of a sentence (in vertical writing mode).
   (note 4)	The details of composition rules for dividing punctuation marks (cl-04) and hyphens (cl-03) are described in Appendix B Spacing between Characters as a complete table, following the descriptions of character classes in 3.9 About Character Classes.

2. When dividing punctuation marks (cl-04) at the end of a sentence reach the end of a line, apply the following rules (see [Fig.75]).

   

   [Fig.75]: Examples of positioning of dividing punctuation marks at the end of a line (in vertical writing mode).

   1. If the line length is 13 character widths and a dividing punctuation mark (cl-04) occurs in the 12th character position, a one em space should be appended after it.

   2. If the line length is 13 character widths and a dividing punctuation mark (cl-04) occurs in the 13th character position, no space should be appended after it. In addition, do not carry over the one em space usually appended after the dividing punctuation marks to the line head of the next line; the line in this case should be set tentsuki.

The character width of hyphens (cl-03) varies according to the type of hyphen. HYPHEN "‐" should be quarter em width (i.e. one quarter of an em width), EN DASH "–" and KATAKANA-HIRAGANA DOUBLE HYPHEN "゠" should be half-width (a half em width), WAVE DASH "〜" should be full-width. Basically there should be no space before and after hyphens (cl-03). However, a half em space should be appended, in principle, when opening brackets (cl-01) follow right after a hyphen (cl-03) and a quarter em space when middle dots (cl-05) follow a hyphen (cl-03).

3.1.7 Characters Not Starting a Line

In principle, no line should begin with closing brackets (cl-02), hyphens (cl-03), dividing punctuation marks (cl-04), middle dots (cl-05), full stops (cl-06), commas (cl-07), iteration marks (cl-09), a prolonged sound mark (cl-10), small kana (cl-11) or warichu closing brackets (cl-29) (line-start prohibition rule). Otherwise the line would have an odd appearance.

3.1.8 Characters Not Ending a Line

No line should end with opening brackets (cl-01) or warichu opening brackets (cl-28) (line-end prohibition rules). Otherwise the line would have an odd appearance.

3.1.9 Positioning of Closing Brackets, Full Stops, Commas and Middle Dots at Line End

In principle, closing brackets (cl-02), commas (cl-07) or full stops (cl-06) at the line end have a half em space after them (see [Fig.76]). This half em space can be deleted for line adjustment (for more about line adjustment, see 3.8 Line Adjustment). However, the possibilities are only half em space or solid. Other spaces, such as a quarter em space should not be used. In principle, the middle dot (cl-05) character at the line end also has a quarter em space before and after, and is handled like a full-width character (see [Fig.76]). This quarter em space can also be deleted for line adjustment, namely middle dots (cl-05) can be set solid before and after (about line adjustment, see 3.8 Line Adjustment). However, in this case also, the only possibilities are quarter em space or solid setting. Other intermediate-sized spacing should not be used.

[Fig.76]: Example of handling closing brackets, full stops, commas and middle dots at the line end like full-width characters.

(note 1)	With regard to closing brackets (cl-02), full stops (cl-06), commas (cl-07) and middle dots (cl-05) at line end, the following processing is defined in JIS X 4051 (see [Fig.77]). full stops (cl-06) After full stops (cl-06), there must be a half em space, including at the line end. This half em space must not be a target for reduction during line adjustment. commas (cl-07) After commas (cl-07)，solid setting is applied. closing brackets (cl-02) After closing brackets (cl-02), solid setting is applied. middle dots (cl-05) In principal, before middle dots (cl-05) there is a quarter em space, and after middle dots (cl-05) solid setting is applied. [Fig.77]: Example of handling closing brackets, full stops, commas and middle dots at the line end in JIS X 4051.
(note 2)	In the letterpress printing era, the following methods were common (see [Fig.78]). For full stops (cl-06), commas (cl-07) and closing brackets (cl-02), if possible, a half em space was preserved. Using a half em space was the general approach. For middle dots (cl-05), if possible, quarter em space was preserved. Using a quarter em space was the general approach. If the line length was not sufficient or too great and line adjustment processing became necessary (see below), the first priority was to replace the half em space after full stops (cl-06), commas (cl-07) and closing brackets (cl-02) with solid setting. The reason was that this was at the line end, and no problems would arise, even though the half em space became set solid. The option of replacing the half em space after punctuation marks with a quarter em space, instead of removing the whole half em space was not used. That meant that there was the choice between either a half em space after punctuation marks or solid setting. The second priority was to replace quarter em spaces before and after middle dots (cl-05) with solid setting. [Fig.78]: Examples of closing brackets, commas and full stops at the end of a line with either a half em space after or set solid.
(note 3)	In some DTP systems etc., after full stops (cl-06), commas (cl-07) or closing brackets (cl-02) at the line end, the line end is always set solid (see [Fig.79]). [Fig.79]: Example of always applying solid setting after closing brackets, full stops, and commas at the line end.

3.1.10 Unbreakable Character Sequences

If the following characters and symbols appear in sequence there will be no line break between them. The reason is that these characters and symbols are to be handled as one unit.

1. Between a sequence of EM DASH "—" characters (to be more specific, for a double dash, see [Fig.80]). Note that some systems implement HORIZONTAL BAR "―" with very similar behavior to EM DASH "—".

   

   [Fig.80]: Sequence of EM DASH characters is unbreakable.

   (note 1)

   Other characters and punctuation marks before and after EM DASH "—" are set solid. However, in the following cases some space should be set between EM DASH "—" and other characters. As mentioned in the next note, HORIZONTAL ELLIPSIS "…", TWO DOT LEADER "‥", prefixed abbreviations (cl-12) and postfixed abbreviations (cl-13) behave similarly to EM DASH "—". Where closing brackets (cl-02) or commas (cl-07) are followed by EM DASH "—", a half em space is inserted between them, in principle. Also, in the middle of lines, where full stops (cl-06) are followed by EM DASH "—", a half em space is inserted between them. However, when full stops (cl-06) are placed at the end of a line a half em space is added after the full stops (cl-06). Where EM DASH "—" is followed by opening brackets (cl-01), a half em space is inserted between them, in principle. Where EM DASH "—" and middle dots (cl-05) are set side by side, a quarter em space is inserted between them, in principle.

   (note 2)

   A double dash is handled as one unit, hence a line break between them is forbidden. In letterpress printing, breaking of the double dash was forbidden very strongly by the fact that the double dash was created as a double full body (so it was actually not possible to break it). Nevertheless, if it was not possible to avoid a break, two EM DASH characters were used instead the double dash. That made it possible to have a line break between them.

2. Between sequences of HORIZONTAL ELLIPSIS "…" or TWO DOT LEADER "‥" (to be more specific, double HORIZONTAL ELLIPSIS "……" or double TWO DOT LEADER "‥‥").

   

   [Fig.81]: Unbreakable sequence of HORIZONTAL ELLIPSIS.

   (note 1)

   In the letterpress printing era, double HORIZONTAL ELLIPSIS etc. was created as a sequence of HORIZONTAL ELLIPSIS. Hence, compared to double dash, the line break between them was not so strongly forbidden.

3. Between European numerals (see [Fig.82], [Fig.83], and [Fig.84].). European numerals indicate ranks via the position of a numeral.

   (note 1)

   It is possible to have a line break between ideographic numerals. Also it is possible to have a line break after IDEOGRAPHIC COMMA "、" used as a decimal separator or an indicator for approximate number, and KATAKANA MIDDLE DOT "・" as a decimal point. However, the position between ideographic numerals and IDEOGRAPHIC COMMA "、" used as a decimal separator or an indicator for approximate number, and KATAKANA MIDDLE DOT "・" used as a decimal point is unbreakable. The reason is that the natural usage of ideographic numerals is to write them like "二百三十五", that is with inserted rank indicators (Explanation: "二" means "two", "百" means "hundred". "二百" means "two hundred". "三" means "three", "十" means "ten". "三十" means "thirty"."五" means "five". The complete sequence "二百三十五" means "Two hundred and thirty five".). Hence, it is not necessary to express rank via position. In contrast, line breaks are forbidden for European numerals, since it is necessary to express the rank via the position. Furthermore, if in vertical writing mode European numerals are placed in the appropriate direction side by side, they are used like the ideographic numerals, and it is possible to have a line break between them.

   (note 2)

   When writing European numerals, FULL STOP "." is used as a decimal point, COMMA "," or space is used as a rank indicator. Line breaks cannot occur before and after these characters (see [Fig.84]: the space before "4" expresses a rank).

4. Between prefixed abbreviations (cl-12) (YEN SIGN "¥"，DOLLAR SIGN "$"，CENT SIGN "¢" etc.) and the following arabic or ideographic numeral (see [Fig.82]). The reason is that such character sequences are to be handled as one unit.

   

   [Fig.82]: Unbreakable sequences between prefixed abbreviations and the following European numeral.

5. Between postfixed abbreviations (cl-13) (PERCENT SIGN "%", PER MILLE SIGN "‰" etc.) and the preceding European numeral or ideographic numeral (see [Fig.83]). The reason is that such character sequences are to be handled as one unit.

   

   [Fig.83]: Unbreakable sequences between postfixed abbreviations and the preceding European numeral.

   (note 1)

   Some people think that it is appropriate to have a line break between the PERCENT SIGN "%" and the preceding arabic or ideographic numeral. The reason seems to be that PERCENT SIGN "%" has a high level of independence. Furthermore it is possible to have a line break between "0" and "パ" in cases like "50パーセント" (meaning "50 percent", where "percent" is written in katakana).

6. Inter-letter space among Western characters (cl-27) in a word (or, sequence of letters, which it is not possible to hyphenate), or unit indicators (km, kg, mm etc.) in Latin letters (see [Fig.84]).

   

   [Fig.84]: It is not possible to break a line between letters in unit symbols using Latin letters.

   (note 1)	When using HYPHEN "‐" at the end of a line, it becomes possible to have a line break within Western characters (cl-27).
   (note 2)	In this document, description of units with proportional Western characters, such as km and kg, are treated as unit symbols (cl-25).
   (note 3)	In [Fig.84], there is a quarter em space between "4" and "k", because of the convention to insert a quarter em between unit symbols (cl-25) and following European numerals or Western characters (cl-27). It is permitted to break a line between "4" and "k". In this case, there is no quarter em space in either the head or the end of the line. Note that the space between "3" and "4" expresses a rank.

7. Inter-letter space among ruby letters, when composed as mono-ruby. Note that it is possible to break a line between base characters with mono-ruby (see [Fig.85]).

8. Inter-letter space among ruby letters or base characters, composed as group-ruby (see [Fig.85]).

   

   [Fig.85]: Example of unbreakable sequences of ruby.

   (note 1)

   With ruby used for compound words (jukugo-ruby), a group of ruby characters is attached to each base character. It is possible to have a line break between such groups of ruby letters and base characters (see [Fig.86]), however, a line break should not occur between ruby characters related to a given base character. [Fig.86]: Example of a line break for jukugo-ruby.

9. Between a subscript or superscript and an adjacent base character (preceding or following) (see [Fig.87]), or between base characters with ornament characters, or between ornament characters themselves. The reason is that these character sequences are to be handled as one object.

   

   [Fig.87]: Unbreakable sequences between a character and its related subscripts.

10. In order to create a correspondence between notes and the related main text, reference marks (aijirushi) are often added. Line breaks are not allowed before the reference mark or between letters of the reference mark itself (see [Fig.88]). The application of the no-line-break rule here is a matter of style.

    

    [Fig.88]: Unbreakable sequences before an aijirushi (reference marks, European numerals or ideographic numerals).

    (note 1)	In this document, characters in reference marks are treated as characters as reference marks (cl-20).
    (note 2)	Often there are full stops (cl-06) after reference marks. In these cases, line breaks are not allowed between the reference marks and the full stops. The reason is that full stops should not be set at the head of lines (see [Fig.88]).

11. After warichu opening brackets (cl-28), which open warichu, or before warichu closing brackets (cl-29), which close warichu.

12. A unit of furiwake. A unit of furiwake is handled as one object.

3.1.11 Character Sequences which Do Not Allow Space Insertion as Part of Line Adjustment Processing

For line adjustment processing, space must not be added between the following characters. (This is called the inseparable characters rule.) The reason is that these characters or symbols should appear as one unit (for more about line adjustment, see 3.8 Line Adjustment).

1. There must be no space between any characters described in 3.1.10 Unbreakable Character Sequences.

   (note 1)

   Some people think that it is only permissible to increase space between letters in Western words for cases where there is no possibility of regular line adjustment processing.

2. In addition to the cases mentioned above, the inseparable character rule has to be applied to the following cases.

   1. Before or after opening brackets (cl-01) or closing brackets (cl-02).

      (note 1)

      After opening brackets (cl-01) or before closing brackets (cl-02) the inseparable characters rule is always applied. In contrast, before opening brackets or after closing brackets the rule is not applied. Full stops (cl-06) and commas (cl-07) are handled the same as closing brackets.

   2. Before or after full stops (cl-06) or commas (cl-07).

   3. Before or after middle dots (cl-05).

   4. Before or after dividing punctuation marks (cl-04).

   5. Before or after hyphens (cl-03).

   6. Before or after one em, etc. spaces between Japanese characters.

   7. Among base characters with jukugo-ruby.

3.1.12 Examples of Line Adjustment

Methods of line adjustment processing are discussed in 3.8 Line Adjustment. However, since layout processing of punctuation marks is one reason for the need for line adjustment processing, we will here introduce two main examples of cases where line adjustment processing is necessary, and show adjustment examples (see [Fig.89]).

1. The principal approach in Japanese composition is that with the exception of the last line of a paragraph, the length of all lines is the same, so all lines are aligned. As explained before, the line length is set to be n-times the character size established for the kihon-hanmen. Hence, as long as only full-width characters are used, all lines have the same length (see (1) at [Fig.89]).

2. In [Fig.89] at (2), there is an IDEOGRAPHIC COMMA "、" followed by a LEFT CORNER BRACKET "「", and the complete space taken by the two characters is one and a half em. That means that the line overshoots or runs short of the edge of the kihon-hanmen by a half em. To restore a uniform line length, line adjustment is applied as shown at (3) in [Fig.89]. The half em space overshoot or shortage is recovered by reducing inter-character space to a quarter em before the LEFT CORNER BRACKET "「" and after the RIGHT CORNER BRACKET "」".

3. At (4) in [Fig.89], the 15th character is an opening bracket (cl-01). This should not appear at the line end. Ideally, a full width space reduction would be applied, and the character "前" on the second line would be moved onto the first line in the 15th position. In that way, the problem could be avoided. However, in this example a full-width space reduction is not possible, so line adjustment processing is applied as shown at (5) in [Fig.89]. The opening bracket (cl-01) is moved to the second line, and line adjustment by inter-character space expansion is applied. That means that space is inserted in the first line at places where it is allowed.

   

   [Fig.89]: Examples of line adjustment.

3.2.1 Composition of Japanese and Western Mixed Texts

There are a lot of examples of Japanese text in which Western and/or Greek letters are mixed among Japanese letters. Examples are as follows:

1. Using one Latin letter as a symbol for something, like "A" and "B".

2. Using a Western word in a Japanese context, like "editor".

3. Using acronyms of things and organization names, like "DTP" and "GDP".

4. Writing Western book titles and authors in lists of referred books with original spelling.

Latin letters are also used in itemized lists and numbering of headings, as well as symbols for units, symbols for chemical elements, and mathematical symbols. As can be judged from these examples, mixtures of Latin letters among Japanese letters are in daily use in Japanese composition.

3.2.2 Mixed Text Composition in Horizontal Writing Mode

In horizontal writing mode the basic approach is to use proportional Western fonts ([Fig.90]). For European numerals, both half-width fonts and proportional fonts are used. Note that Western word space (cl-26) is a one third em space, in principle, except at line head, line head of warichu, line end and line end of warichu. Western word space (cl-26) at line head, line head of warichu, line end and line end of warichu, is set solid.

[Fig.90]: Example of proportional Western fonts used in Japanese in horizontal writing mode.

[Fig.91]: Example of Western full-width fonts used in Japanese in horizontal writing mode. (In horizontal writing mode, Western full-width fonts are usually not recommended.)

(note 1)	As shown in [Fig.91], there are some examples of Western full-width fonts used in horizontal Japanese typesetting, this usage is not recommended, for cosmetic reasons.
(note 2)	Usually, in horizontal Japanese text, fonts are used with European numeric glyphs that are easy to balance and harmonize with Japanese fonts. Considering line adjustment, the use of fonts with half-width numeric glyphs is recommended. There are some cases of Japanese fonts with half-width glyphs for European digits.
(note 3)	There are two choices for including glyphs of Latin letters and numerals in Japanese and Western mixed text compositions. One way is to use the glyphs for Western characters built into the same Japanese font. The other is to combine an independent Western font for Western characters with a Japanese font for Japanese characters. (Example: [Fig.92] is composed using proportional glyphs for Latin letters and numerals included in Ryumin R-KL. [Fig.93] is composed with Ryumin R-KL for Japanese characters and Times New Roman for Latin letters and numerals.) [Fig.92]: Example of Japanese and Western mixed text with the same font Ryumin R-KL for both Japanese characters and proportional Western characters. [Fig.93]: Example of Japanese and Western mixed text with two distinct fonts - Ryumin R-KL for Japanese characters and Times New Roman for Western characters.

3.2.3 Mixed Text Composition in Vertical Writing Mode

As explained in 2.3.2 Major Differences between Vertical Writing Mode and Horizontal Writing Mode, there are three different styles for setting Latin letters and European numerals in vertical writing mode:

1. Setting Latin letters and/or European numerals one by one in inline direction with Japanese characters (see [Fig.94]). This style is typically adopted when composing a single Latin letter or European numeral between the adjacent Japanese letters. Full-width glyphs are specified for Latin letters and European numerals in this style.

   

   [Fig.94]: Example of Latin letters in normal orientation.

2. Setting Latin letters and/or European numerals rotated 90 degrees clockwise in vertical text mode ([Fig.95]). This style is usually adopted when Latin letters compose a word or sentence. Proportional fonts are specified for characters in this style, as in horizontal writing mode (or half-width fonts for European numerals).

   

   [Fig.95]: Example of Latin letters rotated 90 degrees clockwise.

3. Setting Latin letters and/or European numerals in tate-chu-yoko (horizontal-in-vertical setting, see [Fig.96]). Tate-chu-yoko layout is usually adopted when dealing with a two-digit number in European numerals, or a combination of two or three Latin letters, the length of which is equal to the default size of the line in paragraph direction or longer than that just to an acceptable extent. (A combination of two or three Latin letters may be rotated 90 degrees clockwise rather than set it in tate-chu-yoko layout.) Proportional glyphs (or half-width glyphs for European numerals) are used for characters in tate-chu-yoko layout.

   

   [Fig.96]: Example of European numerals in tate-chu-yoko (horizontal-in-vertical setting).

   (note 1)	Acronyms, such as "GNP", and abbreviations like "Web", are usually set one by one, character-wise in normal orientation (see [Fig.97]). However, there are some cases where acronyms and abbreviations are rotated 90 degrees clockwise (see [Fig.98]). [Fig.97]: Example of acronyms set one by one in normal orientation. [Fig.98]: Example of acronyms rotated 90 degrees clockwise.
   (note 2)	The ideographic numerals were traditionally used in vertical writing mode instead of European numerals. (Road numbers and car registration numbers were examples of a couple of exceptions). However, the more newspapers and other publications have been adopting European numerals in vertical writing mode, the more the use of tate-chu-yoko layout for European numerals has also been increasing.

3.2.4 Method for Setting Full-width Latin Letters and European Numerals

When full-width and fixed-width Western characters or European numerals are set in vertical writing mode as "quasi" Japanese characters, inter-character spaces between these characters and hiragana (cl-15), katakana (cl-16) or ideographic characters (cl-19) are set solid, similar to ordinary ideographic characters (cl-19) (see [Fig.99]). Also, in principle, when full-width and fixed-width Western characters or European numerals are set after full stops (cl-06), commas (cl-07) or closing brackets (cl-02), or before opening brackets (cl-01), insert a half em space after commas (cl-07) or closing brackets (cl-02), or before opening brackets (cl-01). In addition, in these cases, insert a half em space after full stops (cl-06). When full-width and fixed-width Western characters or European numerals are set before a full stop (cl-06), comma (cl-07) or closing bracket (cl-02), or after an opening bracket (cl-01), the inter-character space before the full stop, comma or closing bracket, or after the opening bracket is set solid.

[Fig.99]: Setting example of full-width Latin letters and European numerals.

(note 1)

In this document, full-width and fixed-space Western characters and European numerals are treated as members of the ideographic characters (cl-19) class. Accordingly, when KATAKANA MIDDLE DOT "・" appears before or after full-width and fixed-space Western characters and Western numerals, in principle, a quarter em space is inserted between KATAKANA MIDDLE DOT "・" and Western characters or Western numerals. However, when KATAKANA MIDDLE DOT "・" is used as a ranking symbol between Western numerals, in principle, it is set solid, like ideographic numerals. [Fig.100]: Example of setting KATAKANA MIDDLE DOT as a ranking symbol among full-width, fixed-space European numerals.

(note 2)

The details of ideographic characters (cl-19), including full-width and fixed-width Western characters and European numerals, are described as a complete table in Appendix B Spacing between Characters, following 3.9 About Character Classes.

3.2.5 Handling of Tate-chu-yoko (Horizontal-in-Vertical Settings)

To set strings as tate-chu-yoko (horizontal-in-vertical setting), first set from left to right using solid setting, then align the whole string to the center of the vertical line ([Fig.101]). When hiragana (cl-15), katakana (cl-16) or ideographic characters (cl-19) are set before/after tate-chu-yoko, the inter-character space is set solid. In principle, when tate-chu-yoko is set after a comma (cl-07) or closing bracket (cl-02), or before an opening bracket (cl-01), a half em space is inserted. In addition, when tate-chu-yoko is set after a full stop (cl-06) in the middle of a line, a half em space is inserted. When a full stop (cl-06) is set at the end of a line, a half em space is inserted after it, in principle. When tate-chu-yoko is set before full stops, commas or closing brackets, or after opening brackets, the inter-character space is set solid.

[Fig.101]: Example of setting tate-chu-yoko (horizontal-in-vertical text setting).

3.2.6 Handling of Western Text in Japanese Text using Proportional Western Fonts

Composition rules for Western characters, Western text and European numerals, set rotated 90 degrees clockwise in vertical writing mode, and horizontal writing mode, are as follows:

1. A sequence of Western characters in a Western word should not be broken across a line-break, except where hyphenation is allowed.

2. When line adjustment is done with line adjustment by inter-character space reduction, Western word space (cl-26) is used as first priority. Also, when line adjustment is done with line adjustment by inter-character space expansion, Western word spaces are used as first priority.

3. When line adjustment by inter-character space addition is used, inter-character spaces within Western words and European numerals are not used for expansion.

4. Inter-character space, between hiragana (cl-15), katakana (cl-16) or ideographic characters (cl-19) and Western characters or European numerals, is a quarter em space (see [Fig.102]). The issue as to whether the quarter em space can be used for line end adjustment or not is discussed in 3.8.2 Reduction and Addition of Inter-Character Space and 3.8.4 Procedures for Inter-Character Space Expansion.

   

   [Fig.102]: Example of a quarter em inter-character space between hiragana, katakana and ideographic characters, and Latin characters.

5. In the following cases, a quarter em space is not inserted (see [Fig.103]).

   1. At the start of a line, there is no space before Latin characters or European numerals. At the end of the line, there is no space after Latin characters or European numerals.

   2. In the case where Latin characters and European numerals follow a comma (cl-07) or closing bracket (cl-02), or are followed by opening brackets (cl-01), in principle, a half em space is inserted. In the case where Latin characters and European numerals follow a full stop (cl-06) in the middle of a line, a half em space is inserted. When the full stop (cl-06) is set at the end of a line, in principle, a half em space is inserted after the full stop (cl-06).

   3. In the case where Latin characters and European numerals are set before a full stop (cl-06), comma (cl-07) or closing bracket (cl-02), or after an opening bracket (cl-01), the inter-character space is set solid.

[Fig.103]: Example of no inter-character space before and after Latin characters and European numerals.

(note 1)

In this document, proportional Western characters and European numerals are treated as members of the Western characters (cl-27) class. Note that half- and fixed-width European numerals, when mixed with Japanese text, are treated as members of the grouped numerals (cl-24) class.

(note 2)

The reason a quarter em space is needed between Western characters or European numerals and hiragana (cl-15), katakana (cl-16) or ideographic characters (cl-19), is that the design concept of Latin fonts and Japanese fonts are different from each other, so it looks too tight without the spaces （[Fig.104]). [Fig.104]: Example of solid setting between, katakana and ideographic characters and Latin characters and European numerals. (This method is not recommended).

3.3.1 Usage of Ruby

Ruby is a small-sized, supplementary text attached to a character or a group of characters in the main text. A run of ruby text, usually attached to the right of the characters in vertical writing mode or immediately above them in horizontal writing mode, indicates the reading or the meaning of those characters (see [Fig.105]). The characters in the main text that are annotated by ruby are called "base characters". Kana characters are often used for ruby to indicate how to read kanji characters; this is known as ruby annotation or as "furigana".

[Fig.105]: Ruby and base characters.

In terms of annotation, ruby has several purposes as described below. Depending on the purpose and/or the function of the ruby, there are different ruby composition methods (the details of the composition methods will be provided in the later sections).

1. PURPOSE: Ruby annotation with kana (usually hiragana) to provide readings of kanji characters. There are two types of ruby for this purpose depending on the type of base character:

   1. Add one or more hiragana (cl-15) or katakana (cl-16) ruby character to indicate the reading (Japanese onyomi or kunyomi) for each base ideographic character (cl-19) (see [Fig.106]). This method, attaching one or several hiragana or katakana characters for each base ideographic character, is called mono-ruby.

      

      [Fig.106]: Example of ruby annotation per Kanji Character.

   2. In the Japanese writing system, compound words (jukugo) occasionally appear, usually constructed with a couple of kanji characters. There are two different methods of attaching ruby letters to base kanji characters for these compound words.

      1. Mono-ruby. Ruby letters are attached to each base kanji character, similarly to the previous section (see [Fig.107]).

         

         [Fig.107]: Example of mono-ruby method. Ruby letters are attached to each base kanji character in a compound word.

         (note 1)

         In [Fig.107], there is a quarter em space between the base characters "凝" and "視". So when this line happens to appear in the middle of a paragraph, there needs to be some line adjustment processing.

      2. Jukugo-ruby. Pronunciation is indicated for each base character, but the positioning takes into account the fact that together they make up a compound word (see [Fig.108]). The intention when using jukugo-ruby is to handle the kanji phrase as one object.

         

         [Fig.108]: Example of jukugo-ruby method. Ruby letters are attached to groups of kanji characters in compound words.

         (note 1)	There is no difference between [Fig.107] of mono-ruby and [Fig.108] of jukugo-ruby, when the number of ruby letters for each base kanji character is one or two. When the mono-ruby approach is adopted, the inter-character space between base kanji characters can be expanded for line adjustment (in [Fig.107], the inter-character space between "鬼" and "門", or, "方" and "角" can be expanded). When the jukugo-ruby approach is adopted, line adjustment cannot be applied to the base kanji characters.
         (note 2)	The following examples show the relationship between ruby letters and base kanji characters. Example of mono-ruby: "凝+(ぎよう)" "視+(し)" Example of jukugo-ruby 1: "凝+(ぎよう) 視+(し)" Example of jukugo-ruby 2: "(凝視)+(ぎよう/し)"
         (note 3)	Books commonly adopt kana-based jukugo-ruby for kanji compound words. However, due to technical difficulties for rendering jukugo-ruby in machine-assisted text layout, the adoption of kana-based mono-ruby is increasing. For example, newspapers do not use jukugo-ruby, and study aids generally use mono-ruby because it is considered more important to show the readings of each kanji character for students than to be concerned about the beauty of the layout.
         (note 4)	Multiple kanji compound words can form one compound phrase. In this case, there are two ways to attach ruby, i.e. attaching ruby to the compound phrase as a whole, or to each word which forms the compound (see [Fig.109]). Similarly, a Japanese personal name consists of a given name and a family name, which together form a compound of a full name, and it is an editorial decision whether to attach two runs of ruby, one each for given name and family name, or to attach the full ruby text to the compound which represents the reading of the full name. [Fig.109]: Examples of ruby attachment for a compound phrase.
         (note 5)	In most cases, the reading of a kanji compound word is just a concatenation of the readings of each individual kanji character, but some phrases have their own native readings (known as jukuji readings), which cannot be derived from the readings of each character. Attaching ruby to those phrases, which are usually sequences of two or three kanji characters, is essentially the same as attaching ruby in katakana to kanji and/or kana base characters (see [Fig.110]). [Fig.110]: Examples of ruby for jukuji readings.

2. PURPOSE: Ruby annotation that annotates a kanji or kana word with kana to provide its meaning, together with it's reading. In terms of ruby layout, attaching ruby text to a single character in kanji or kana is essentially the same as attaching the reading to a kanji character (e.g. attaching ruby text "バザール", 'Bazaar', in kana to a kanji character "市" is just like attaching the reading "いち" to that character). When attaching ruby text to a run of base text consisting of two or more characters in kanji and/or kana, the ruby text needs to be positioned as if it corresponds to the annotated text itself, no matter how the ruby characters are distributed across each base character. The most typical example of this is attaching ruby text to a kanji compound word to indicate a corresponding loan word in katakana (see [Fig.110]). The use of ruby text of this kind is on the increase in proportion to the growing need for translations and loan words. This type of ruby, namely ruby letters that are attached to two or more base characters as one object, is called group-ruby. Group-ruby and it's base characters are unbreakable, because of their behavior as one object (it is possible to break a line in the middle of the base characters where jukugo-ruby is in use).

   

   [Fig.111]: Examples of ruby for compound kanji words to indicate corresponding words in katakana.

   (note 1)

   Having said that, the layout of ruby text to a single kanji character is not entirely the same depending on the use of ruby, and may differ according to whether it is for indicating the reading or the meaning. In fact, the katatsuki, one of ruby layout per character which will be described in the later section, does allow the nakatsuki layout for those ruby which indicate meaning.

3. PURPOSE: Ruby annotation, usually with katakana characters, to indicate the reading or the meaning of a Western word used in base text (see [Fig.112]). There are opposite cases where a synonymous Western word in Latin characters is attached as a ruby annotation to a Japanese word in kana or kanji (see [Fig.112]). These cases are less used than a and b, however they are quite common in study guides, translated books and travel guides.

   

   [Fig.112]: Examples of Latin characters used either in base text or ruby text for Western words.

4. PURPOSE: Ruby annotation using kanji for a base text word in hiragana. This is called furikanji, and is very rarely found.

Hereafter, mainly the usage of (a) and (b) will be explained.

3.3.2 Choice of Base Characters to be Annotated by Ruby

There are several methods of choosing how to attach ruby annotations to which base characters.

1. 'General-ruby' is the method of attaching ruby annotations to all base characters in kanji.

2. 'Para-ruby' is the method of attaching ruby annotations to only those base characters in kanji for which readings are difficult.

   (note 1)

   The 'para-ruby' method is further divided into the method of attaching ruby annotations regardless of multiple occurrences of the same base kanji character, and the method of attaching ruby annotations to only the first instance of the same base kanji character. Furthermore, the latter method has several variations in counting the first instance for the entire book, each chapter, or each spread, etc.

Note that ruby should be attached to all kanji characters in a compound word, to reflect the unitary nature of the text. To attach ruby to only some of the kanji characters in a compound word is not recommended (see [Fig.113]).

[Fig.113]: Examples of ruby on kanji characters in a compound word. (Left side, recommended. Right side, not recommended.)

3.3.3 Choice of Size for Ruby Characters

The character size of ruby characters is, in principle, the half size of the base characters (see [Fig.114]).

[Fig.114]: Examples of ruby with half the size of the base characters.

The 'one-third-ruby' characters are used on rare occasions to attach three ruby characters to one full-width kanji character. One-third-ruby for vertical layout has the dimension of the half of the base character in width and the one third in height. Those for horizontal writing mode have the dimension of half of the base characters in height and one third in width (see [Fig.115]).

[Fig.115]: Examples of one third ruby.

When ruby is attached to twelve point or larger base characters (usually used for headings), the size of the ruby letter is generally smaller than half the size of the base characters, considering the proportion of the sizes of base characters and ruby. When all is said and done, these cases are very rare.

[Fig.116]: Examples of ruby at a size smaller than half the size of the base characters.

3.3.4 Choice of Sides for Ruby with Respect to Base Characters

In principle, ruby is attached to the right of base characters in vertical writing mode, and above in horizontal writing mode.

In some special cases, ruby can be seen to the left of base characters in vertical writing mode, and below in horizontal writing mode, but this is very rare.

There are cases where two kinds of ruby are attached, one to either side of the base characters, one for readings and the other for meanings (see [Fig.117]). This is also very rare.

[Fig.117]: An example of ruby attached to both sides of the base characters.

In the following sections, the ruby composition methods will be explained on the assumption that the size of ruby is half the size of the base characters, and they will be attached to the right in vertical writing mode and above in horizontal writing mode. First we look at the basic composition rules of mono-ruby, group-ruby and jukugo-ruby, then the rules of positioning of ruby with respect to those characters which come before and after the base characters, and finally the composition rules at the line head and at the line end.

3.3.5 Positioning of Mono-ruby with Respect to Base Characters

When mono-ruby characters are Japanese, they are set solid. If mono-ruby characters have their own character widths such as Western characters or European numerals, they are set according to their own widths and then the ruby text is placed so that its center matches that of its base character. There are more variations depending on the combination of the base character and ruby text and accordingly various composition rules have been invented, which will be explained with examples.

When attaching two kana ruby characters to a single base character, the lengths of the ruby text and the base text are the same and they are positioned as shown in [Fig.118].

[Fig.118]: An example of composition with two ruby characters.

When attaching a single ruby character to a single base character, there are two ways of positioning the ruby character.

1. In vertical writing mode, attach a ruby character so that its vertical center matches that of the base character (see [Fig.119]). In horizontal writing mode, attach a ruby character so that its horizontal center matches that of the base character (see [Fig.119]). This positioning of a ruby character is called 'nakatsuki' (center-alignment).

2. In vertical writing mode, attach a ruby character so that the top of its virtual body is aligned with the top of that of the base character (see [Fig.119]). This positioning of a ruby character is called 'katatsuki' (top-alignment). For horizontal writing mode, 'katatsuki' should not be adopted. If a ruby character is attached so that the left-edge of its virtual body is aligned with the left-edge of that of the base character, it would result in the loss of the center of balance, which doesn't look good (see [Fig.120]).

[Fig.119]: Examples of nakatsuki and katatsuki alignment.

[Fig.120]: Example of katatsuki alignment in horizontal layout (this is intentionally wrong and should not be applied).

When attaching more than three hiragana ruby characters to a single base character, the ruby characters are set solid. In this case, where the length of a ruby text is longer than that of its base character, positioning of the ruby text depends on which alignment has been adopted for a single ruby character. There is another issue: how to maintain the spatial balance of the ruby characters hanging over those characters which are not related base characters. The adjustment of inter-character spacing for those characters which come before and after the base character will be explained in a later section.

1. When nakatsuki alignment is adopted for a single ruby character, position a ruby text so that its vertical center is aligned with that of its base character in vertical writing mode (see [Fig.121]). In horizontal writing mode, position a ruby text so that its horizontal center is aligned with that of its base character (see [Fig.121]).

   

   [Fig.121]: Example 1 of positioning of ruby text with more than three characters.

2. When katatsuki alignment is adopted for a single ruby character, there are two methods, as follows.

   1. Position the ruby text so that its vertical center is aligned with that of its base character (see [Fig.121]).

   2. Depending on the type of script of the adjacent characters to the base character, and the number of ruby characters, a decision is made about whether ruby hangover is allowed on the character before its base character, or on the character after, or on both adjacent characters. At break-even situation, the hangover is usually on the character after its base character (see [Fig.122]).

[Fig.122]: Example 2 of positioning of ruby text with more than three characters (vertical writing mode).

For mono-ruby, base characters and adjacent ruby characters are handled as one object, and internal line-breaks are prohibited.

3.3.6 Positioning of Group-ruby with Respect to Base Characters

When the length of a sequence of base characters (number of characters * advance-width of each character) and that of the ruby text are the same, each text is set solid and the center of both texts are aligned with each other (see [Fig.123]).

[Fig.123]: Examples of group-ruby where the length is the same as that of the base text.

When the length of the ruby text is shorter than that of its base characters, set the base text solid and stretch the ruby text by adding a certain amount of inter-character space between each adjacent ruby character, so that both texts balance each other. To be more specific, where 2 units of inter-character space are used between ruby characters, add 1 unit of space between the start of the base text and the start of the ruby text, and between the end of the ruby text and the end of the base text. This will give a balanced appearance, and is the method specified in JIS X 4051 (see [Fig.124]). Another way is to first align the leading characters for both the base text and ruby text and the ends of both trailing characters, and then add the same amount of inter-character space between the rest of the ruby characters (see [Fig.125]).

[Fig.124]: Example 1 of distribution of group-ruby alongside base characters where the length of the ruby is shorter than that of the base characters.

[Fig.125]: Example 2 of distribution of group-ruby alongside base characters where the length of the ruby is shorter than that of the base characters.

(note 1)

In letterpress printing, there were not many choices for adjustment of inter-character space between ruby characters. Therefore, depending on the number of characters in the base text and its ruby text, the choice was whether to add a certain amount of space before the leading ruby character and after the trailing character, or not. In the former case it had been said that for 2 units of inter-character space between each adjacent ruby characters, adding 1 unit of the leading and trailing space would give a balanced appearance.

(note 2)

When the length of the ruby text is far shorter than that of the base text, the method specified in JIS X 4051 could result in space twice the size of a ruby character for the leading and the trailing space, which might give a misleading appearance. Therefore, a criterion for deciding whether or not to adopt the method of JIS X 4051 is to see if the amount of the leading and the trailing space exceeds the full-width size (or up to 1.5 times the size) of a ruby character (see [Fig.126]). [Fig.126]: Examples of distribution of group-ruby where the length is much shorter than that of the base text.

When the length of the ruby text is longer than that of the base characters, balance the base characters with the ruby text by setting the ruby text solid and adding a certain amount of inter-character space between each adjacent base character. To be more specific, for 2 units of inter-character space, add 1 unit of space between the start of the ruby text and the start of the base text, and between the end of the base text and the end of the ruby text, as specified in JIS X 4051 (see [Fig.127]). Another way is to first align the start of both the leading characters and the end of the trailing characters, and then add a certain amount of inter-character space between each adjacent base character (see [Fig.128]).

[Fig.127]: Example 1 of distribution of group-ruby where the length is longer than that of the base characters.

[Fig.128]: Example 2 of distribution of group-ruby where the length is longer than that of the base characters.

For group-ruby, base characters and adjacent ruby characters are handled as one object, and internal line-breaks are prohibited. Also, for an object constructed with base characters and adjacent ruby characters it is prohibited to insert additional spaces between each character for line adjustment.

3.3.7 Positioning of Jukugo-ruby with Respect to Base Characters

If the number of ruby characters are two or less for each kanji character which participates in a kanji compound word (or jukugo), then for each run of ruby text associated with each base character, compose ruby characters as described in 3.3.5 Positioning of Mono-ruby with Respect to Base Characters (see [Fig.129]).

[Fig.129]: Example 1 of distribution of jukugo-ruby.

If there is any kanji character in a given kanji compound word which needs more than three ruby characters, the jukugo-ruby layout cannot be used. In this case, attach the ruby text to the kanji compound word as a whole. The available methods include the layout as specified in JIS X 4051, which is similar to the group-ruby method described in 3.3.6 Positioning of Group-ruby with Respect to Base Characters (see [Fig.130]), and layout decided by the phonetic structure of the kanji compound word and the type of script of the adjacent characters (see [Fig.131]). The latter method can be used unless a run of ruby text for the base character hangs over another base character more than a full character width (or one and a half times the full-width) of a ruby character.

[Fig.130]: Example 2 distribution of jukugo-ruby.

[Fig.131]: Example 3 distribution of jukugo-ruby.

(note 1)

There are often cases where the jukugo-ruby consists of one ruby character followed by three ruby characters, and vice versa, for a kanji compound word of two base characters. If the mono-ruby layout were chosen for these cases, it would look like as shown in [Fig.132], which wouldn't be very beautiful. [Fig.132]: Example of distribution as mono-ruby for jukugo.

Jukugo-ruby can be split into two lines at the boundary of each unit of ruby text attached to one kanji character. When a kanji compound word consists of two characters, each unit will be processed using the mono-ruby method. When dividing a compound word that consists of three kanji, use the mono-ruby method for the first kanji character and use the jukugo-ruby method for the remaining two kanji characters, and vice versa. In order to maintain the correspondence of each kanji to its ruby annotation, the layout of the ruby may be different after the division (see [Fig.133]). Note that jukugo-ruby and its base characters cannot be the subject of inter-character space expansion for line adjustment.

[Fig.133]: Examples of distribution of jukugo-ruby split across two lines.

3.3.8 Adjustments of Ruby with Length Longer than that of the Base Characters

When the length of any ruby text is shorter than that of the base characters, the main text can be just set solid because there is no need for any adjustment of the inter-character spacing between base characters and their adjacent characters in the main text.

[Fig.134]: Set solid when the length of ruby text is shorter than that of base characters.

When the length of the ruby text is longer than that of the base characters, the method of composing the main text depends on how much the ruby text hangs over the kanji or kana characters, or punctuation marks, which are adjacent to the principal characters. The following are the general rules (see [Fig.135] and [Fig.136]). They were established especially in order to avoid misreading and to maintain the beauty of the layout.

1. Ruby text shall not hang over the ideographic characters (cl-19) adjacent to the base characters.

2. When the adjacent character is a hiragana (cl-15), katakana (cl-16), prolonged sound mark (cl-10) or small kana (cl-11), the ruby text may overhang the character up to the full-width size of the ruby characters.

3. The ruby letter may go over the base characters and overhang the half em spaces which are inserted after closing brackets (cl-02), full stops (cl-06) or commas (cl-07), set before the target ruby object, up to the full-width size of a ruby letter. Also, the ruby letter may go over the base characters and hang over the half em spaces which are inserted before opening brackets (cl-01), set after the target ruby object, up to the full -width size of a ruby letter. Note that when the spaces are reduced for line adjustment, the room for ruby letter overhang is also compressed. (For example, if the space is a quarter em in the base character size, the ruby letter can overhang by a half em in ruby letter size.)

4. When the adjacent character is an inseparable character (cl-08), the ruby text may overhang the character up to the full-width size of a ruby character.

5. When the adjacent character is one of the middle dots (cl-05), the ruby text may overhang the middle dots, in principle, up to the full-width size of a ruby character. But if there is any reduction of space before and after the middle dots as a result of the line adjustment, the amount of the extension shall be up to the amount of space after the middle dots plus 1/2 a ruby character size when the middle dots are set before the ruby object, or the space before the middle dots plus 1/2 a ruby character size when the middle dots are set after the ruby object.

6. When the adjacent character is one of the closing brackets (cl-02), the ruby text may go over the principal characters up to the full-width size of a ruby character. Note that the overhang must not go beyond the closing bracket itself.

7. When the adjacent character is a comma (cl-07) or full stop (cl-06), the ruby text may go over the base characters and overhang the comma or full stop up to the full-width size of a ruby character. Note that the overhang must not go beyond the comma or the period itself.

8. Also, when the adjacent character is one of the opening brackets (cl-01) before the ruby object, the ruby text may go over the base characters and hang over the opening brackets up to the full-width size of a ruby character. Note that the overhang must not go beyond the opening brackets..

[Fig.135]: Example 1 of distribution of ruby characters overhanging adjacent characters.

[Fig.136]: Example 2 of distribution of ruby characters overhanging adjacent characters.

(note 1)	There is an opinion that it is not good style to hang ruby characters over opening brackets (cl-01), especially LEFT CORNER BRACKET "「" and RIGHT CORNER BRACKET "」". When this opinion holds, there are two possibilities. One is not allow ruby characters to overhang the opening brackets (cl-01). The other is allow ruby characters to overhang the opening brackets up to 1/2 the size of a ruby character.
(note 2)	JIS X 4051 classifies katakana characters and kanji characters in same class. Accordingly, in JIS X 4051 it is prohibited for ruby characters to overhang adjacent katakana characters.
(note 3)	There is another variation that allows ruby text to overhang any ideographic characters (cl-19), hiragana (cl-15) or katakana (cl-16) up to the full-width size of a ruby character (see [Fig.137]). [Fig.137]: Example 3 of distribution of ruby characters overhanging adjacent characters.
(note 4)	There is a further variation that does not allow ruby text to overhang any ideographic (cl-19), hiragana (cl-15) or katakana (cl-16) characters (see [Fig.138]). [Fig.138]: Example 4 of distribution of ruby characters overhanging adjacent characters.

When the line head starts with ruby annotated text where the ruby text length is shorter than that of the base characters, compose the text so that the first base character is aligned with the line head. Similarly, when ruby annotated text ends at the line end and the ruby length is shorter than that of the basic characters, compose the text so that the last basic character is aligned with the line end.

When the line head starts with ruby annotated text where the ruby text length is longer than that of the base characters, compose the text so that the first ruby character which overhangs the base text is aligned with the line head, and vice versa (see [Fig.139]). Alternatively, there is a variation by which the text is composed so that the first base character is aligned with the line head, and vice versa (see [Fig.140]).

[Fig.139]: Example 1 of positioning of ruby characters at the line head and at the line end.

[Fig.140]: Example 2 of positioning of ruby characters at the line head and at the line end.

When aligning the first base character to the line head and the last base character to the line end, ruby text is not allowed to extend beyond the hanmen or the area of the column. If it does, the following adjustments should be considered in positioning base characters and ruby characters.

1. Mono-ruby at the line head: Make adjustments so that the top of the ruby text is aligned with that of the base characters (see [Fig.140]).

2. Mono-ruby at the line end: Make adjustments so that the bottom of the ruby text is aligned with that of the last base character (see [Fig.140]).

3. Group-ruby at the line head: Make adjustments so that the top of the ruby text is aligned with that of the first base character, and add the same amount of inter-character spacing between the base characters and between the end of the last base character and the end of the last ruby character after the last base character (the method specified in JIS X 4051) (see [Fig.141]).

   

   [Fig.141]: Example 3 of positioning of ruby characters at the line head and at the line end.

4. Group-ruby at the line end: Make adjustments so that the end of the ruby text is aligned with that of the last base character and add the same amount of inter-character space between the base characters and the space between the start of the base text and the start of the ruby text (the method specified in JIS X 4051) (see [Fig.141]).

5. Jukugo-ruby at the line head or at the line end: Make the same adjustments as described in (c) or (d) for the group-ruby.

6. Jukugo-ruby at the line head: Make adjustments so that the top of the ruby text is aligned with that of the first base character. A run of ruby characters for a base character may overhang the adjacent base characters of the same kanji compound word, up to the full-width size (or one and a half of it) of a ruby character. If the extension should go beyond the limit, just force the ruby text out of the base characters, or make a further adjustment by adding inter-character space between the base characters.

7. Jukugo-ruby at the line end: Make adjustments so that the end of the ruby text is aligned with that of the last base character. A run of ruby characters for a base character may overhang the adjacent base characters of the same kanji compound word, up to the full-width size (or one and a half of it) of the ruby characters. If the extension should go beyond the limit, just force the ruby text out of the base characters, or make a further adjustment by adding inter-character space between the base characters.

8. Jukugo-ruby split across two lines: jukugo-ruby can be split across two lines, with one part at the line end and the other at the line head. In the case of a compound word with two kanji characters, it is as the same as dealing with one kanji character with a mono-ruby text at the line end and the other kanji character with another mono-ruby text at the next line head. In the case of a phrase with three kanji characters, handle one kanji character with mono-ruby text and the remaining two kanji characters with jukugo-ruby, and vice versa. The layout of one kanji character with mono-ruby text will be composed by method (a) or (b) described above. The layout of two kanji characters with jukugo-ruby text will be composed by method (f) or (g) above.

3.3.9 Composition of Emphasis Dots

Emphasis dots (also known as bouten or side dots) are symbols placed alongside a run of kanji or kana characters to emphasize the text.

Composition of emphasis dots is as follows (see [Fig.142]).

1. The character size of emphasis dots is the half size of the base characters to be emphasized.

2. Emphasis dots are attached to the right of the base characters in vertical writing mode, or above them in horizontal writing mode. The center of emphasis dots is aligned with that of the base characters.

3. There are many symbols that could be specified for use as emphasis dots. SESAME DOT "﹅" in vertical writing mode and BULLET "•" in horizontal writing mode are those used for emphasis dots in general.

[Fig.142]: Composition of emphasis dots.

3.4.1 Where the Inline Cutting Note (Warichu) is used

Warichu (inline cutting note) is a type of inline notation, where two lines of small characters are inserted into the text. Warichu divides a line into two sub lines. The frequency of use of the inline cutting note is not so high. However, the inline cutting note is very important for study guides, travel guides, reference books, encyclopedias and manuals, because it is very effective for inserting notes at the point in the text where they are needed (see [Fig.143]). Inline cutting note is usually used in vertical writing mode. It is very infrequently used in horizontal writing mode.

[Fig.143]: Warichu (inline cutting note).

3.4.2 Character Size for Inline Cutting Notes and Line Gaps

Character size for an inline cutting note depends on the character size established for the kihon-hanmen. Usually, around six point size is used (see [Fig.143]).

The space between adjacent lines in an inline cutting note is zero, that is to say, there is no line gap between them (see [Fig.144]).

As shown in [Fig.144], an inline cutting note usually has two lines, and is surrounded by LEFT PARENTHESIS "(" and RIGHT PARENTHESIS ")" characters that are double the size of the characters in the inline cutting note itself. There is no space between the surrounding text and parentheses for the inline cutting note.

[Fig.144]: Example of construction of an inline cutting note.

Symbols, like opening brackets (cl-01), closing brackets (cl-02), commas (cl-07) and full stops (cl-06) are also used in inline cutting note text. In such cases, the handling of such symbols is the same as for the main text.

In vertical text, the horizontal width of the inline cutting note area is wider than the width of a kihon-hanmen line. The horizontal centers of the kihon-hanmen line and inline cutting note area are aligned. The line gap used to establish the kihon-hanmen should not be affected by the horizontal width of the inline cutting note area. In other words, the line gap for the kihon-hanmen needs to be designed wider than usual in preparation for the use of the inline cutting note. Warichu is used also in horizontal text, however it is not so common, and usually occurs only in study guides and encyclopedias.

The length of the two lines of the inline cutting note should be as near as possible the same. When the inline cutting note can be set in one kihon-hanmen line, the whole inline note text should be broken at a position where line breaking is permitted, and where the two resulting lines are as close as possible to the same length. The length of the second line should not be longer than the length of the first line. Note that the same line breaking rules are used as for basic text (see [Fig.145]).

[Fig.145]: Examples showing how the inline cutting note can be set in one line of base text.

3.4.3 Handling an Inline Cutting Note when it Straddles Two Kihon-hanmen Lines

When an inline cutting note will not fit on a single kihon-hanmen line, it will wrap onto the following line, and will be set as shown in [Fig.146] or [Fig.147].

[Fig.146]: Example of an inline cutting note straddling two base text lines.

[Fig.147]: Example of an inline cutting note straddling three base text lines.

3.5.1 Line Head Indent at the Beginning of Paragraphs

A paragraph, a section of a document which consists of one or more sentences to indicate a distinct idea, usually begins on a new line. For the related line head indent at the beginning of paragraphs (in JIS 4051, this is called the "paragraph line head indent") the following methods are available. The amount of space used for the indentation is, in principle, one em space using the character size in the paragraph.

1. Line head indent at the beginning of paragraphs is applied to all paragraphs. Nearly all books and magazines make use of this method (see [Fig.148]).

   

   [Fig.148]: Example of line head indent at the beginning of paragraphs.

   (note 1)

   In the case of written conversational text followed by, for example, "she said", if the conversational part is bounded by LEFT CORNER BRACKET "「" and RIGHT CORNER BRACKET "」" and the "she said" text appears after a line-break, the conversational text and the "she said" text are considered to be one continuous phrase. Hence, there is no line head indent at the beginning of the "she said" phrase, after the conversation part, i.e. so-called tentsuki (see [Fig.149]). When you have a (mathematical) formula on a separate line in horizontal writing mode, and a following line containing text such as "will be" there is also no line head indent at the beginning of the paragraph. On the other hand, there is also an approach in novels etc., where the first line indent is applied to paragraphs (see [Fig.150]).

   

   [Fig.149]: Layout example 1 of a line immediately following a written conversation.
   

   [Fig.150]: Layout example 2 of a line immediately following a written conversation.

2. Line head indent at the beginning of paragraphs is not applied for any paragraph at all, and the tentsuki position is used (see [Fig.151]). There are examples of this method being used in certain books and magazines for the sake of styling, but this is rather hard to read.

   

   [Fig.151]: Example of no line head indent at the beginning of paragraphs.

3. In principle, line head indent is applied at the beginning of a paragraph. However, a paragraph immediately following a tentsuki-set heading is also set tentsuki, so that the beginning of the heading and the paragraph are aligned (see [Fig.152]). In some books and magazines this method is applied to text in horizontal writing mode.

   

   [Fig.152]: Example of no line head indent at the beginning of paragraphs immediately following headings.

On the other hand, for example with itemization, there is also the method that indents the second and following lines of the paragraph (see [Fig.153]). This is the so-called "questions and answer" (Q&A) form. It has the effect that numbers (if used) stand out.

[Fig.153]: Layout example for itemization.

3.5.2 Line Head Indent and Line End Indent

The line head indent is the indentation of the line head by a fixed amount, starting from the line head side of the hanmen (in the case of one column) or of the column area (in the case of several columns). In contrast, the indentation of the line end position by a fixed amount, starting from the line head, is called line end indent.

There are examples of line head indent for quotations in separate lines (see [Fig.154]) or for headings in separate lines. Line end indent is used, for example, for headings or for quotations in separate lines.

[Fig.154]: Example of line head indent for a quotation in a separate line.

3.5.3 Single Line Alignment Processing

The Japanese "single line alignment method" is a process for setting alignment for a run of text that is shorter than a given line length. This method is frequently used for headings and poems. The following methods are available (see [Fig.155]).

1. Centering: The space between adjacent characters is, in principle, set solid. (If space is needed between Japanese text and western text, before opening brackets (cl-01) and after closing brackets (cl-02), that space is inserted based on the table in Appendix B Spacing between Characters.) Also, if there is an explicit instruction to insert spaces, such spaces are inserted. If there is not solid setting but a fixed space between characters, this is used; the amount of space at the line head and line end is made equal, and the center of the character sequence is unified with the center of the line.

2. Line head alignment: The space between adjacent characters is, in principle, set solid. (If space is needed between Japanese text and western text, before opening brackets (cl-01) and after closing brackets (cl-02), that space is inserted based on the table in Appendix B Spacing between Characters.) Also, if there is an explicit instruction to insert spaces, such spaces are inserted. If there is not solid setting but a fixed space between characters, this is used; the start of the character sequence is unified with the line head, and if the line is not full, the line end is kept empty.

3. Line end alignment: The space between adjacent characters is, in principle, set solid. (If space is needed between Japanese text and western text, before opening brackets (cl-01) and after closing brackets (cl-02), that space is inserted based on the table in Appendix B Spacing between Characters.) Also, if there is an explicit instruction to insert spaces, such spaces are inserted. If there is not solid setting but a fixed space between characters, this is used; the end of the character sequence is unified with the line end, and if the line is not full, the line head is kept empty.

4. Even spacing distribution: The space between adjacent characters is, in principle, set solid. (If space is needed between Japanese text and western text, before opening brackets (cl-01) and after closing brackets (cl-02), that space is inserted based on the table in Appendix B Spacing between Characters.) Also, if there is an explicit instruction to insert spaces, such spaces are inserted. In addition, using the space made available during line adjustment processing, equal character spacing is applied where possible (or character space reduction is applied where possible). The start of the character sequence is aligned to the position of the line head, and the end of the character sequence to the position of the line end.

[Fig.155]: Single line alignment processing.

(note 1)	Several justification methods are applied for positioning of headings or items of tables. For example, centering is often used for headings in horizontal writing mode, taking the left-right balance is taken into account. However, there are also examples of line head alignment.
(note 2)	Even spacing distribution is often used for printing Haiku in separate lines (see [Fig.156]). [Fig.156]: Example of Haiku positioning with even spacing distribution.

3.5.4 Widow Adjustment of Paragraphs

The intent of widow adjustment of paragraphs is to avoid that the last line of a paragraph contains less than a given number of characters. This is also called "widow" processing.

(note 1)

Widow adjustment of paragraphs is defined in JIS X 4051, sec. 4.20 "Widow Adjustment of Paragraphs".

(note 2)

Widow adjustment of paragraphs is not regarded as very important in Japanese composition. However, care is taken to avoid cases such as a single character in the last line of a paragraph (often this is tolerated), or - even more extreme - just one character on a page just before a page break or a new recto (see [Fig.157]). [Fig.157]: Example of just one character on a page just before a page break (to be avoided).

3.6.1 Usage of Tab Setting

Tab setting is useful for alignment of table data, itemized lists, etc. where a series of characters need to be set at specific alignment positions within a line (see [Fig.158]).

[Fig.158]: Example of tab setting.

For tab setting, it is necessary to identify tab positions, tab types (how to align the characters in the tab position), and the characters to be set. For this purpose, it is necessary to insert a tab sign before the tabbed character. The series of characters just after the tab sign are the target characters (see [Fig.159]). If there is more than one tab sign, it is necessary to set the same numbers of tab positions and tab types as the number of tab signs.

[Fig.159]: Tab signs and the target text of tab setting.

3.6.2 Types of Tab Settings

There are the following types of tab setting to align texts.

1. Start alignment tab setting: the start position of the text is aligned to the tab position (see [Fig.160]).

   

   [Fig.160]: Examples of start alignment tab settings.

2. End alignment tab setting: the end position of the text is aligned to the tab position (see [Fig.161]).

   

   [Fig.161]: Examples of end alignment tab settings.

3. Center alignment tab setting: the center of the text is aligned to the tab position (see [Fig.162])．

   

   [Fig.162]: Examples of center alignment tab settings.

4. Alignment with a specified character tab setting: the start position of a specified character or sign (for example, a period) in the text is aligned to the tab position (see [Fig.163]).

   

   [Fig.163]: Examples of specified character alignment tab settings.

3.6.3 The Method of Setting the Target Text

Set the text from the line head to the position before the tab sign in the first tab position, set the text from the first tab sign to the next tab sign in the second tab position, and so on. The behavior of opening brackets (cl-01) and closing brackets (cl-02), etc. is same as for the main text.

Following are some examples. The behavior of text before and after the tab positions are very difficult to anticipate, so it is necessary to design using trial and error.

1. If the target string is the first series of the line, the characters should be set in the first tab position from the start of the line, and so on, one after another (see [Fig.164]).

   

   [Fig.164]: Example of tab setting 1.

2. If the target string of text is too long to be set before the next tab position and overflows, the next string of text is aligned to the tag position after the end of the preceding string (see [Fig.165]).

   

   [Fig.165]: Example of tab setting 2.

3. If the beginning of the string overlaps with the end of the preceding string as the result of the tab setting indication, the following string is set just after the preceding string (see [Fig.166]).

   

   [Fig.166]: Example of tab setting 3.

4. If there is no tab position corresponding to the target string, the string should be set from the tab position of the next line, and so forth (see [Fig.167]).

   

   [Fig.167]: Example for tab setting 4.

3.7.1 Superscripts and Superscripts

Superscripts and subscripts are small letters associated with base characters, and typically used to indicate SI unit symbols, or used for mathematical or chemical formulae.

Superscripts and subscripts are usually set after the base character, with some exceptions for chemical formulae (which appear before the base character). They should be set solid.

For examples of superscripts and subscripts see [Fig.168]. In this document, superscripts and subscripts and their base characters are handled as ornamented character complex (cl-21) characters.

[Fig.168]: Examples of superscripts and subscripts.

JIS X 4051 specifies the character size and the block direction positioning of superscripts and subscripts alongside the base character to be implementation definable parameters. However it is recommended that the size of superscripts and subscripts are around 60% of the base character, depending on the size of the base character.

It is prohibited to break lines within an ornamented character complex (cl-21) sequence. Also, it is prohibited to use inter-character spacing within an ornamented character complex (cl-21) sequence for line adjustment.

3.7.2 Furiwake

Furiwake is a typesetting style for setting multiple phrases or sentences in in the middle of a line. Furiwake is also used to indicate options (see [Fig.169]). Study guides, manuals and reference books sometimes use furiwake. In many furiwake styles, multiple lines are indicated with opening brackets (cl-01) and closing brackets (cl-02), etc.

[Fig.169]: Example of furiwake.

Furiwake is usually done as follows (see [Fig.170]). In this document, the lines which combine to make the furiwake are called furiwake-gyou.

1. The character size of the furiwake-gyou is usually the same as that of the base paragraph. Sometimes, the furiwake-gyou is a little bit smaller than the size of the base paragraph. Sometimes, the font style of the furiwake-gyou is different typeface the style of the base paragraph.

2. In the same furiwake block, the top positions of all the furiwake-gyou lines are aligned.

3. The line length of the furiwake block is the line length of the longest furiwake-gyou. However, it is permitted to indicate the length of the furiwake block, and break the furiwake-gyou lines. In this style, the start positions of the broken lines should be explicitly indicated. When there are line break marks in the furiwake-gyou, the line is broken in the indicated places. In this style, the start positions of the wrapped lines are aligned to the first line. The space between wrapped lines should be set solid.

4. The line-feed space of each furiwake block should be explicitly indicated.

5. The center line of the furiwake block should be aligned with the center line of the main text.

6. When the furiwake block is enclosed by opening brackets (cl-01) and closing brackets (cl-02), etc. the width of brackets should be the same as the width of the furiwake block.

7. One furiwake block should be set per base text line. One furiwake block should not be extended across multiple base text lines.

[Fig.170]: Setting method of furiwake.

The size of the line-feed space of the paragraph which contains the furiwake block, should be explicitly indicated. The space should be decided by considering the content of the furiwake block, and may therefore differ from the size of the line-feed space of kihon-hanmen.

In general, the width of the furiwake block is larger than the width of an inline cutting note block. However, unlike in the case of the inline cutting block, the whole furiwake block should be set inside of the kihon-hanmen, or a column of the kihon-hanmen Setting a furiwake block that extends beyond the border of the kihon-hanmen is prohibited.

3.7.3 Jidori Processing

In cases such as lists of names of Japanese people, the length of some part of the text may be explicitly defined. In such cases, different numbers of characters are set, using adjustment of the inter-character spacing, so that they are all aligned to the same length. This is called jidori processing (see [Fig.171]).

[Fig.171]: Example 1 of jidori processing.

Sometimes, in horizontal writing mode, text in running heads (with the exception of chapter and section numbers) are set using jidori processing. For example, three to six characters are set in a 7 character space (based on the size of the characters in the running head (see [Fig.172]). Two characters are set in a 6 character space to avoid too much space. Seven characters are set solid in a seven character space, and eight or more characters are set solid in a space of eight or more characters. This rule can be applied to other numbers of characters, such as five, six and eight.

[Fig.172]: Example 2 of of jidori processing.

Jidori processing should be done as follows:

1. The length for the jidori processing should be defined as a whole number of full-width characters at the size defined for the surrounding text.

2. The jidori text should be adjusted using spacing between characters so that the sides of the text are aligned at the defined length. The following, however, should be set solid:

   1. Positions where line breaks are prohibited: inter-character spaces between European numerals; between two EM DASH "—" characters; between two TWO DOT LEADER "‥" characters; between two HORIZONTAL ELLIPSIS "…" characters; and so on. These sequences should be treated as a single block.

      (note 1)

      The handling of opening brackets (cl-01) and closing brackets (cl-02) in jidori processing is also very controversial. Usually, the space before opening brackets (cl-01) and space after closing brackets (cl-02) is set solid. In jidori processing, the space before opening brackets (cl-01) and the space after closing brackets (cl-02) may be used for line adjustment, but the space after opening brackets (cl-01) and the space before closing brackets (cl-02) should not used for adjustment, because it is prohibited to break lines in these positions. [Fig.173] shows one example with even spacing for all characters, a second example as explained here, and a third example that is set solid before and after the opening brackets (cl-01) and closing brackets (cl-02).

      

      [Fig.173]: Examples of jidori processing including opening brackets and closing brackets.

   2. Where Western word space (cl-26) or full-width ideographic space (cl-14) are inserted, the add same space as for other inter-character spaces to the value of the Western word space or the full-width ideographic space.

3. If there is only one character, it should be aligned to the left of the jidori block.

3.7.4 Processing of Math Symbols and Math Operators

Math symbols and math operators, such as EQUALS SIGN "=", APPROXIMATELY EQUAL TO OR THE IMAGE OF "≒", PLUS SIGN "+" and MINUS SIGN "−" are commonly used not only for scientific and technical documents but also for ordinary books. In the Japanese composition system, there are two different groups of math symbols, which are each treated differently. So in this document math symbols are classified into two different classes; math symbols (cl-17) and math operators (cl-18).

Composition rules for math symbols (cl-17) and math operators (cl-18) are as follows:

1. The width of math symbols (cl-17) and math operators (cl-18) is full-width, i.e. one em (see [Fig.174]).

2. The inter-character space between math symbols (cl-17) or math operators (cl-18) and before and after adjacent characters, such as grouped numerals (cl-24), Western characters (cl-27), and ornamented character complex (cl-21) in one line is set solid (see [Fig.174]). However, when the top and/or the bottom of the mathematical formula is grouped numerals (cl-24) or Western characters (cl-27), the space between ideographic characters (cl-19), hiragana (cl-15) and katakana (cl-16) is quarter em space.

   

   [Fig.174]: Example of math symbols and math operators set within an ordinary line.

   (note 1)

   The inter-character space between ideographic characters (cl-19), hiragana (cl-15) or katakana (cl-16) and math symbols (cl-17) or math operators (cl-18) is solid. However, it is better to insert a quarter em space between ideographic characters (cl-19),hiragana (cl-15) or katakana (cl-16) and adjacent math operators (cl-18) when the math operators (cl-18) are followed by grouped numerals (cl-24) or Western characters (cl-27) which represent some mathematical value (see [Fig.174]).

   

   [Fig.175]: Example of math symbols and math operators set within an ordinary line.

3. When math formulae or chemical formulae are set in one independent line, the inter-character space between math symbols (cl-17) and adjacent grouped numerals (cl-24), Western characters (cl-27) and ornamented character complex (cl-21) is quarter em. Also, when math formulae or chemical formulae are set in an individual line, the inter-character space between math operators (cl-18) and adjacent grouped numerals (cl-24), Western characters (cl-27) or ornamented character complex (cl-21) is set solid.

   

   [Fig.176]: Example of math symbols and math operators in one independent line.

   (note 1)

   In most case when a math formula or chemical formula is set in one independent line, the position of the formula is the center of the line in horizontal writing mode, and has is indented by some indicated number of characters from the head of line in vertical writing mode.

   (note 2)

   In the formula in one independent line, there is another method to set the inter-character space between grouped numerals (cl-24) or Western characters (cl-27) and math symbols (cl-17), solid or half em. When the inter-character space before and after the math symbols (cl-17) is set to a quarter em or a half em, there is another method to set the inter-character space, between math operators (cl-18) and grouped numerals (cl-24) or Western characters (cl-27), that is, a quarter em.

   

   [Fig.177]: Another example of setting math symbols and math operators in one independent formula line.

4. A line can be broken between math symbols (cl-17) or math operators (cl-18) and adjacent grouped numerals (cl-24), Western characters (cl-27) or ornamented character complex (cl-21).

   (note 1)	In an independent formula line, when there are more than one place where the line can be broken the first priority is before the math symbols (cl-17), and the next is before the math operators (cl-18).
   (note 2)	The inter-character space before and after KATAKANA MIDDLE DOT "・", before opening brackets (cl-01) and after closing brackets (cl-02) in an independent formula line is set solid, except for the case to set space between these characters and adjacent math symbols (cl-17) or math operators (cl-18).

3.8.1 Necessity for Line Adjustment

Line adjustment processing is applied where inter-character adjustments are needed to bring the line end into the correct alignment, e.g. because of line wrap or other reasons. Within a paragraph, lines are created by separating character sequences at places where line breaking is not prohibited. Except for the end of the last line of a paragraph, it is necessary to set the head and end of each line at predicable, aligned positions. For the last line of the paragraph, it is still necessary to set the head at the aligned position, however the line end may not aligned to the other alignment position. To achieve this, only inter-character spacing indicated in the table of Appendix B Spacing between Characters, or explicitly chosen spaces, are inserted, and other inter-character spaces are set solid.

There are various reasons for line adjustment processing. Examples of the most important ones will be given below.

1. Mixed use of characters and symbols (e.g. grouped numerals (cl-24) or Western characters (cl-27)) where not all characters are full-width (see [Fig.178]).

   

   [Fig.178]: Example of grouped numerals and Western characters.

2. Sequences of punctuation marks. For example, a sequence of a closing bracket (cl-02) and a full stop (cl-06) takes one and a half em spaces together (see [Fig.179]). However, if an opening bracket (cl-01) follows immediately after the full stop, these punctuation marks will need two em spaces together. Hence, no adjustment is needed to correctly align the line end (see [Fig.179]).

   

   [Fig.179]: Examples of sequences of punctuation marks.

3. Mixtures of characters with different sizes (see [Fig.180]).

   

   [Fig.180]: Example of characters within brackets which are made a level smaller than the normal character size.

   (note 1)

   In cases where additional information like page references, explanations of terminology, etc. appear within brackets, sometimes the character size is a level smaller than the character size established by the kihon-hanmen.

4. Cases where line head wrapping, line end wrapping or unbreakable character sequences should not be broken (see [Fig.89]).

3.8.2 Reduction and Addition of Inter-Character Space

Line adjustment processing targets places with a predefined space or solid setting. Methods for line adjustment are as follows.

1. Line adjustment by inter-character space reduction. This means that a half em space is reduced after commas (cl-07) or closing brackets (cl-02), or before opening brackets (cl-01), and Western word space (cl-26) is reduced within a defined limit.

2. Line adjustment by inter-character space expansion. Line adjustment by inter-character space expansion means expanding inter-character space for line adjustment, where inter-character space is allowed to be extended up to a defined limit, such as for Western word space (cl-26) or other places where it is not prohibited to extend inter-character space.

Normally line adjustment by inter-character space reduction is preferred. Only when there are no spaces that can be reduced is line adjustment by inter-character space expansion applied. The reason for the preference of line adjustment by inter-character space reduction comes from the thinking that characters in solid setting should not have more inter character space, if at all possible.

(note 1)

Line adjustment by hanging punctuation is a method of avoiding line head wrap of full stops (cl-06) and commas (cl-07). This method is not formally defined in JIS X 4051, however JIS X 4051 does provide explanatory material about it. Line adjustment by hanging punctuation is a method which is only applied to full stops (cl-06) and commas (cl-07). These touch the hanmen and are set in a place outside the defined line length (see [Fig.181]). This is also applied, for example, in books, in order to avoid the addition of inter character space and maintain solid setting. However, line adjustment by hanging punctuation is not an appropriate method for dealing with a mixture of Japanese and Latin script text, since the latter principally does not apply line adjustment by hanging punctuation. In addition, there is another argument against line adjustment by hanging punctuation. Originally it was a method used in letterpress printing, to make the task of line adjustment easier. Furthermore, as shown at the end of line 1 and 5 of [Fig.181], if possible the full stops (cl-06) or commas (cl-07) are placed at the line end (the 18th position). In DTP there are examples of hanging punctuation like in line 3, but this may be regarded as unnecessary processing. [Fig.181]: Examples of line adjustment by hanging punctuation.

3.8.3 Procedures for Inter-Character Space Reduction

For line adjustment by inter-character space reduction decisions must first be made about the preferred order in which reduction processing options are applied, and the maximum amount of space reduction needed. Inter-character space reduction is processed with following priorities.

1. Western word space (cl-26), which is usually one third em, is reduced by equal amounts, to leave a minimum of a quarter em space between words. The same space reduction is applied to all spaces on the target line at the same time.

2. The half em space after closing brackets (cl-02),commas (cl-07) and full stops (cl-06) at the end of a line, is deleted and set solid.

3. The quarter em spaces both before and after the middle dots (cl-05) are deleted and set solid.

4. The quarter em space before or after middle dots (cl-05), in the middle of a line, is reduced equally with proportional character size as far as solid setting.

5. The half em spaces before opening brackets (cl-01) or after closing brackets (cl-02) or commas (cl-07), in the middle of a line, are reduced equally with proportional character size, as far as solid setting.

   (note 1)	With the exception of the line end position, the half space after full stops (cl-06) should not adapted, since it plays an important role as a sentence separator.
   (note 2)	Commas (cl-07) fulfill different roles to opening brackets (cl-01) and closing brackets (cl-02). Hence, there are examples where it is preferred to reduce the half em space before or after brackets, rather than to reduce the half em space after commas.
   (note 3)	The reduction of the half em space before opening brackets (cl-01) or after closing brackets (cl-02) and commas (cl-07) up to solid setting is regarded as too much reduction. Hence, there are examples where the maximum amount of space reduction is up to quarter em space.

6. The quarter em space between Japanese text (hiragana (cl-15), katakana (cl-16) and ideographic characters (cl-19)) and Latin script text (grouped numerals (cl-24), Western characters (cl-27) and unit symbols (cl-25)), is reduced equally with proportional character size, as far as one eighth em space.

   (note 1)

   There are also examples where the quarter em space between Japanese text (hiragana (cl-15), katakana (cl-16) and ideographic characters (cl-19)) and Latin script text (Western characters (cl-27), grouped numerals (cl-24) or unit symbols (cl-25)) is regarded as fixed space, and space adaptation is not applied.

In JIS X 4051, the space after closing brackets (cl-02), commas (cl-07) and middle dots (cl-05) at the end of a line are set solid, and the space after commas (cl-07) at the end of a line is set to a half em. Accordingly, JIS X 4051 defines the priority of processing as follows:

1. Western word space (cl-26), which is usually one third em, is reduced by equal amounts, to leave a minimum of a quarter em space between words.

2. The quarter em space before and after middle dots (cl-05) is reduced equally with proportional character size as far as solid setting.

3. The half em spaces before opening brackets (cl-01) and after closing brackets (cl-02) or commas (cl-07), are reduced equally with proportional character size as far as solid setting.

4. The quarter em spaces between Japanese text (hiragana (cl-15), katakana (cl-16) and ideographic characters (cl-19)) and Latin script text (grouped numerals (cl-24), Western characters (cl-27) and unit symbols (cl-25)) are reduced equally with proportional character size, as far as a 1/8th em space.

3.8.4 Procedures for Inter-Character Space Expansion

As with line adjustment by inter-character space reduction, for line adjustment by inter-character space expansion at first the order of processing and the maximum amount of space to be added are defined. In JIS X 4051, the following processing order is defined.

1. Western word space (cl-26), which is usually one third em, is added equally with proportional character size up to a maximum of a half em size for each space.

2. The quarter em space between Japanese text (hiragana (cl-15), katakana (cl-16) and ideographic characters (cl-19)) and Latin script text (grouped numerals (cl-24), Western characters (cl-27) and unit symbols (cl-25)) is increased equally with proportional character size, up to half em space (or one third em space).

   (note 1)

   Like with inter-character space reduction, there are also examples there the quarter em space between Japanese text (hiragana (cl-15), katakana (cl-16) and ideographic characters (cl-19)) and Latin script text (Western characters (cl-27), grouped numerals (cl-24) and unit symbols (cl-25)) is regarded as a fixed space, and space adaptation is not applied.

3. For places which do not fall under (a) or (b) and which do not have bunrikinshi, space is added equally with proportional character size up to a quarter em space.

4. In addition to the adaptation in the manner of (a), (b) and (c), in cases where such processing is not possible, space is added equally with proportional character size, with the exception of places which require bunrikinshi.

   (note 1)	JIS X 4051 provides a definition in addition to (d). This says that it depends on each layout processing system whether inter-character space should be added equally. This includes the space between Western characters (cl-27).
   (note 2)	The detail of the places where space expansion is possible for line adjustment is described in Appendix B Spacing between Characters as a complete table, in accordance with the concept of character class in 3.9 About Character Classes.

3.9.1 Differences in Positioning of Characters and Symbols

The positioning of characters and symbols may vary depending on the following.

1. Is the character width full-width, half-width, or something else?

2. Is it allowed or forbidden to place the character or symbol at the line head? If it is allowed, how will it be placed?

3. Is it allowed or forbidden to place the character or symbol at the line end? If it is allowed, how will it be placed?

4. Are characters and symbols appearing in sequence in solid setting, or will there be a fixed space between them? For example, sequences of ideographic characters (cl-19) and hiragana (cl-15) are set solid, and for Western characters (cl-27) following hiragana (cl-15) there will be quarter em space.

5. Is it allowed to have a line break within a sequence of characters? For example, there must not be a line break in a sequence of grouped numerals (cl-24).

6. Is it allowed to use the space between characters in a sequence during line adjustment processing? For example, is inter-character space reduction or addition possible between the characters appearing in sequence? Another issue to be decided is the preferred order for adjustment processing, and the amount of the allowed adjustment.

3.9.2 Grouping of Characters and Symbols depending on their Positioning

During layout processing, the issues mentioned in the previous section are addressed by grouping characters and symbols according to their characteristics, and handling them as character classes.

JIS X 4051 also provides similar character classes but that are slightly different from this document. Furthermore JIS X 4051 states that it is implementation-defined how to handle characters that are not explicitly mentioned, e.g. whether they should belong to either class or not.

A few character classes of this document are modified from JIS X 4051. In Appendix A Character Classes, there is a whole mapping table to ISO/IEC 10646 Annex A collection 285 (BASIC JAPANESE) and collection 286 (JAPANESE NON IDEOGRAPHIC EXTENSION). All character classes of this document are as follows:

1. Opening brackets (cl-01)

   Example:

   ‘“（〔［｛〈《「『【

   etc.

2. Closing brackets (cl-02)

   Example:

   ’”）〕］｝〉》」』】

   etc.

   (note 1)

   In JIS X 4051, IDEOGRAPHIC COMMA "、" and COMMA "," are classified as closing brackets (cl-02), because they have similar positioning methods. However, in this document, the handling of IDEOGRAPHIC COMMA "、" and COMMA "," are described as an independent class, named commas (cl-07).

3. Hyphens (cl-03)

   Example:

   ‐〜

   etc.

4. Dividing punctuation marks (cl-04)

   Example:

   ？！

   etc.

5. Middle dots (cl-05)

   Example:

   ・：；

6. Full stops (cl-06)

   Example:

   。．

7. Commas (cl-07)

   Example:

   、，

8. Inseparable characters (cl-08)

   Example:

   —…‥

   etc.

9. Iteration marks (cl-09)

   Example:

   ヽヾゝゞ々

   etc.

   (note 1)	In JIS X 4051, iteration marks such as IDEOGRAPHIC ITERATION MARK "々" are classified as "no line break allowed before" characters. In this document, IDEOGRAPHIC ITERATION MARK "々" etc. are classified as iteration marks (cl-09).
   (note 2)	There is another method where it is permitted to break a line before IDEOGRAPHIC ITERATION MARK "々". In this case, IDEOGRAPHIC ITERATION MARK "々" is regarded as a member of ideographic characters (cl-19).

10. Prolonged sound marks (cl-10)

    Example:

    ー

    (note 1)	In JIS X 405, KATAKANA-HIRAGANA PROLONGED SOUND MARK "ー" is a member of "Japanese characters with no line break allowed before". In this document KATAKANA-HIRAGANA PROLONGED SOUND MARK "ー" is the only member of prolonged sound mark (cl-10).
    (note 2)	In JIS X 4051, it is permitted to exclude KATAKANA-HIRAGANA PROLONGED SOUND MARK "ー" from the "Japanese characters with no line break allowed before" character class.
    (note 3)	When it is permitted to break a line before KATAKANA-HIRAGANA PROLONGED SOUND MARK "ー", KATAKANA-HIRAGANA PROLONGED SOUND MARK "ー" is regarded as a member of the katakana (cl-16) class.

11. Small kana (cl-11)

    Example:

    ぁぃぅぇぉァィゥェォっゃゅょッャュョ

    etc.

    (note 1)	In JIS X 4051, small katakana letters, such as KATAKANA LETTER SMALL TU "ッ", are regarded as members of "Japanese characters with no line break allowed before" character class. In this document, small katakana letters are classified in small kana (cl-11). Accordingly, the "Japanese characters with no line break allowed before" character class in JIS X 4051 is divided to three separate classes; iteration marks (cl-09), prolonged sound mark (cl-10) and small kana (cl-11).
    (note 2)	In JIS X 4051, it is permitted to exclude small kana letters (ぁぃぅァィゥ etc.) from the "Japanese characters with no line break allowed before" character class as an implementation definable option.
    (note 3)	When it is permitted to break a line before small hiragana (ぁぃぅ etc.) are regarded as members of the hiragana (cl-15) class, and small katakana (ァィゥ etc.) are regarded as members of the katakana (cl-16) class.

12. Prefixed abbreviations (cl-12)

    Example:

    ￥＄￡＃

    etc.

13. Postfixed abbreviations (cl-13)

    Example:

    °′″℃￠％‰

    etc.

14. Full-width ideographic space (cl-14)

    Example:

    U+3000 IDEOGRAPHIC SPACE

15. Hiragana (cl-15)

    Example:

    あいうえおかがきぎ

    etc.

    (note 1)

    The reason kanji etc. (Japanese characters except 1. to 12.) and hiragana are regarded different classes in JIS X 4051, is the difference in the case of ruby, and specifically the handling in terms of characters jutting out of the base and overhanging adjacent characters.

16. Katakana (cl-16)

    Example:

    アイウエオカガキギ

    etc.

    (note 1)

    In JIS X 4051, katakana and kanji are included in the same class (Japanese characters except for 1. to 12.). However, in this document, when ruby characters jut out of the base characters and overhang adjacent hiragana or katakana, the handling is same. That is the reason that in this document, katakana (cl-16) is an independent character class.

17. Math symbols (cl-17)

    Example:

    ＝≠＜＞≦≧⊆⊇∪∩

    etc.

    (note 1)

    In JIS X 4051, math symbols (＋－÷×etc.) and math operators (＝≠＜＞≦≧⊆⊇∪∩etc.) are included in the "Japanese characters excluded from 1. to 12." class or Western character class. However, handling of math symbols and math operators adjacent to Western character or Arabic numerals is different from kanji. So, in this document, new math operators (cl-18) and math symbols (cl-17) classes are defined.

18. Math operators (cl-18)

    Example:

    ＋－÷×

    etc.

19. Ideographic characters (cl-19)

    Example:

    亜唖娃阿哀愛挨〃仝〆♂♀

    etc.

    (note 1)

    In JIS X 4051, corresponding character class for ideographic characters (cl-19) is "Japanese characters excluded from 1. to 12.".

20. Characters as reference marks (cl-20)

    Characters which are inside verification seal (those are characters inside a verification seal that appear in the line just after the item applicable for reference marks of notes)

21. Ornamented character complexes (cl-21)

    (note 1)

    The name of this class in JIS X 4051 is "Characters included in ornamented base characters complex". The meaning of "base characters complex" is characters in a complex including ruby, ornament characters and emphasis dots.

22. Simple-ruby character complexes (cl-22)

    (note 1)

    The name of this class in JIS X 4051 is "Characters included in base characters complex with ruby (excluding jukugo-ruby)".

23. Jukugo-ruby character complexes (cl-23)

    (note 1)

    The name of this class in JIS X 4051 is "Characters included in base characters complex with jukugo-ruby".

24. Grouped numerals (cl-24)

    Sequences of European numerals which are not full-width and are handled as Japanese text, the decimal point or the comma and space used as grade indicator in number.

25. Unit symbols (cl-25)

    Units described here include combinations of Latin script and Greek script characters used for international units (SI).

    (note 1)

    There are units created with combinations of Latin and Greek script characters with a full-width character frame (full-width units). Such units are not part of the characters for units described here. Furthermore, full-width characters for units are mainly used in vertical writing mode. Their usage in horizontal writing mode is regarded as bad style and should be avoided (see [Fig.182]).

    

    [Fig.182]: Example of a unit which encompasses a full-width unit character (upper part) and characters for Latin script text (lower part).

26. Western word space (cl-26)

27. Western characters (cl-27)

    (note 1)

    Western characters (cl-27) include punctuation marks, such as commas, used in Western context. Among these punctuation marks, several marks are used both in a Japanese context and Western context. However, these marks have different character shapes depending on whether they are used in a Japanese context or Western context. For example, LEFT PARENTHESIS "(" and RIGHT PARENTHESIS ")" has not only different width (Japanese, half em, Western, proportional) but are also different in line position (Japanese, center of the character frame in the inline direction, Western, base line and descender line dependent) and design (Japanese, slightly bent and constant line thickness, Western, strongly curved and dynamic line thickness). The usage of these two differently designed commas should be explicit. Usually, in a Japanese context Japanese design is used, and in Western context Western design is used. However, there are some ambiguous cases, such as "エディター（editor）は……". In this case, English spelling is indicated using parentheses in a Japanese line of text. In this particular case, Japanese design is better.

28. Warichu opening brackets (cl-28)

    Example:

    （〔［

    etc.

29. Warichu closing brackets (cl-29)

    Example:

    ）〕］

    etc.

    (note 1)

    warichu opening brackets (cl-28) and warichu closing brackets (cl-29) are used for surrounding inline cutting notes and the space before and after. They are in a separate class since they differ from normal brackets with regard to their processing.

30. Characters in tate-chu-yoko (cl-30)

3.9.3 Positioning Methods for each Character Class

For each character class it is possible to describe whether the characters may appear at the line head or line end or not, the positioning method for the line head or line end positions (if available), the amount of space between sequences of several characters, and the combination with character classes before or after the characters (in a 2 dimensional table). In JIS X 4051 this is shown in table 5 "Amount of space (between characters)".

Also, it can be defined for each combination of the character classes (in a two dimensional table) whether the characters of classes appearing in sequence allow for a line break between them, or whether it is possible during line adjustment processing to add inter character space between them. In JIS X 4051 these items are also shown in a two dimensional table. Table 6 shows whether a line break is possible, and table 7 shows if it is possible to add inter character space.

The width, in principle, of the space between each character or symbol in character classes used in this document is described in the table of Appendix B Spacing between Characters.

The combinations of adjacent characters and symbols in character classes used in this document, and where text is breakable or not, is described in the table of Appendix C Possibilities for Line-breaking between Characters.

The width of spaces between each character or symbol in character classes used in this document, and which can be reduced, is described in the table of Appendix D Opportunities for Inter-character Space Reduction during Line Adjustment. Also, expandable spaces are described in the table of Appendix E Opportunities for Inter-character Space Expansion during Line Adjustment.