Technical Reports

Unicode Standard Annex #14

Line Breaking Properties

Summary

This report presents the specification of line breaking properties for Unicode characters.

Status

This document has been reviewed by Unicode members and other interested parties, and has been approved by the Unicode Technical Committee as a Unicode Standard Annex. It is a stable document and may be used as reference material or cited as a normative reference from another document.

A Unicode Standard Annex (UAX) forms an integral part of the Unicode Standard, carrying the same version number, but is published as a separate document. Note that conformance to a version of the Unicode Standard includes conformance to its Unicode Standard Annexes.

A list of current Unicode Technical Reports is found on http://www.unicode.org/unicode/reports/. For more information about versions of the Unicode Standard, see http://www.unicode.org/unicode/standard/versions/.

The References provide related information that is useful in understanding this document. Please mail corrigenda and other comments to the author(s).

Contents

• 1 Overview and Scope
• 2 Definitions
• 3 Description
• 4 Conformance
• 5 Line Breaking Properties
  • 5.1 Description of Line Breaking Properties
  • 5.2 Additional Details on Dictionary Usage
• 6 Line Breaking Algorithm
• 7 Pair-table Based Implementation
  • 7.1 Minimal Table
  • 7.2 Extended Context
  • 7.3 Example Table
  • 7.4 Sample Code
  • 7.5 Combining Marks
  • 7.6 Customization
  • 7.7 Examples of Customization
• References
• Acknowledgments
• Modifications

1 Overview and Scope

The Unicode Standard [U3.0] presents a summary of basic line-breaking behavior, but does not give a complete specification. This technical report provides the needed information in a way that reflects best practices. Normative line-breaking properties are assigned to those characters whose line breaking behavior must be identical across all implementations. For all other classes of characters informative line-breaking properties are provided.

Following the formal definitions and summary of line breaking properties, there are three main sections:

1. Section 5 contains a textual description of the line breaking behavior of the characters of the Unicode Standard and their grouping by line breaking property. These descriptions do not take account of the order of precedence. 
2. Section 6 provides a set of rules listed in order of precedence that constitute a line breaking algorithm.
3. Section 7 provides the detailed description of an efficient pair table based implementation of the algorithm.

2 Definitions

All terms not defined here shall be as defined in the Unicode Standard. The notation defined in this technical report differs somewhat from the notation defined in the Unicode Standard. All notation used here without an explicit definition shall be as defined in the Unicode Standard.

Line fitting - the process of determining the how much text will fit on a line of text, given the available space between the margins and the actual display width of the text.

Line Break - the position in the text where one line ends and the next one starts.

Line Break Opportunity - a place where a line is allowed to end. Whether a given position in the text is a valid line break opportunity depends on the line breaking rules in force, as well as on context.

Line Breaking - the process of selecting that part of a text that can be displayed on a line. In other words, selecting one among several line breaking opportunities such that the resulting line is optimal (unless the user requested an explicit line break).

Line Breaking Property - A character property with mutually exclusive values, as set out in Table 1 and arranged in approximate order of precedence. Line breaking properties are used to determine the type of break.

Mandatory Break - a line must break following a character that has the mandatory break property. Also known as a forced break. This is indicated in the rules as B !, where B is the character with the mandatory break property. (In the notation of the Unicode Standard, Version 3.0 [U3.0], this would be: B ×, although the standard doesn't specify whether or not a break is forced or just an opportunity.)

Direct Break - a line breaking opportunity exists between two adjacent characters of the given line breaking properties. This indicated in the rules below as B ÷ A, where B is the character class of the character before and A is the character class of the character after the break. If they are separated by one or more space characters, a break opportunity also exists after the last space. In the pair table, the optional space characters are not shown.

Indirect Break - a line breaking opportunity exists between two characters of the given line breaking properties only if they are separated by one or more spaces. In this case, a break opportunity exists after the last space. No break opportunity exists if the characters are immediately adjacent. This is indicated in the pair table below as B % A, where B is the character class of the character before and A is the character class of the character after the break. Even though no space characters are not shown in the pair table, an indirect break can only occur if one or more spaces follow B. In the notation of the Unicode Standard, this would be represented as two rules: B × A and B SP+ ÷ A.

Prohibited Break - no line breaking opportunity exists between two characters of the given line breaking properties, even if they are separated by one or more space characters. This is indicated in the pair table below as B ^ A, where B is the character class of the character before and A is the character class of the character after the break and the optional space characters are not shown. In the notation of the Unicode Standard, this would be expressed as a rule of the form: B SP* × A.

Hyphenation - Hyphenation uses language specific rules to provide additional line breaking opportunities within a word. Hyphenation improves the layout of narrow columns, especially for languages with many longer words, such as German or Finnish. For the purpose of this document, it is assumed that hyphenation is equivalent to insertion of soft hyphen characters. All other aspects of hyphenation are outside the scope of this document.

Table 1  Line Breaking Properties (* = normative)

3 Description

Lines are broken as result of either of two conditions. The first condition is the presence of an explicit line breaking character. The second condition results from a formatting algorithm having selected among available line breaking opportunities the particular one that results in the optimal layout of the text.

The definition of optimal line break is outside the scope of this document. Different formatting algorithms may use different methods of determining an optimal break. For example, simple implementations just consider a line at a time, trying find a locally optimal line break. A common approach is to allow no compression and consider the longest line that fits. When compression is allowed, a locally optimal line break seeks to balance the relative merits of the resulting amounts compression and expansion for different line break candidates.

More complex algorithms may take into account the interaction of line breaking decisions for the whole paragraph. The well known text layout system,  [T_EX] implements a well known example of such a globally optimal strategy that may make complex tradeoffs to avoid unnecessary hyphenation and other legal, but inferior breaks. For a description of a globally optimizing line fitting algorithm, see [Knuth78]. 

For the purpose of this document, what is important is not so much what defines the optimal amount of text on the line, but how line breaking opportunities are determined. 

Three principal styles of context analysis determine line-breaking opportunities.

1. Western (spaces and hyphens are used to determine breaks)
2. East Asian (lines can break anywhere, unless prohibited)
3. South East Asian (require morphological analysis)

The first is commonly used for scripts employing the space character. Hyphenation is often used with space-based line breaking to provide additional line break opportunities - however, it requires knowledge of the language and potentially user interaction or overrides.

The second style of context is used with East Asian ideographic and syllabic scripts. 

NOTE: When Korean text is laid out justified, the second style is commonly used, even for interspersed Latin letters. But when ragged margins are used, the first style (relying on spaces) is commonly used instead.

For multilingual text, styles one and two can be unified into a single set of specifications, based on the information provided in this report. Some Unicode characters have explicit line breaking properties assigned to them. These can be utilized with these two styles of context analysis for line break opportunities. 

NOTE: Interpretation of line breaking properties in bidirectional text takes place before applying rule L1 of the Unicode Bidirectional Algorithm. However, it is strictly independent of directional properties of the characters or of any auxiliary information determined by the application of rules of that algorithm.

The third style is used for scripts such as Thai, which do not use spaces, but which restrict word-breaks to syllable boundaries, the determination of which requires knowledge of the language comparable to that required by a hyphenation algorithm. Such and algorithm is beyond the scope of this report.

4 Conformance

• The line breaking behavior of characters with normative line breaking properties is described in the Unicode Standard. (See The Unicode Standard, Version 3.0, Chapters 6 and 13). Unless otherwise stated, the information in this technical report is not intended to supersede the normative specifications found in the Unicode Standard, but to organize the description in a different context and provide additional informative detail.
• All line breaking properties are informative, except for the line breaking properties marked with a * in Table 1 Line Breaking Properties. The behavior for characters with normative line breaking properties must be the same for all conformant implementations.
• Conformant implementations must not tailor characters with normative line breaking properties to any of the informative properties, but may tailor characters with informative line breaking properties to one of the normative line breaking properties.
• Higher level protocols may further restrict, override, or extend the line breaking properties of certain characters in some contexts.

5 Line Breaking Properties

The main emphasis in this section is to provide additional description of the line breaking behavior and to summarize the membership of character classes for each line breaking property. 

The classification by properties defined here is used as input into two algorithms defined below that implement workable default line breaking methods. In a few instances, the descriptions in this section provide additional detail about handling a given character at the end of a line, which goes beyond the simple determination of line breaks.

Data File

The full classification of all Unicode characters by their line breaking properties is available in the current version of the file LineBreak.txt [Data] in the Unicode Character Database [UCD]. This is a tab-delimited, two column plain text file, with code position, line breaking class.  A comment at the end of each line indicates the character name. Ideographic, Hangul, Surrogate, and Private Use ranges are collapsed by giving a range in the first column.

As more scripts are added to the Unicode Standard, and more scripts become more widely implemented and used on computers, more line breaking classes may be added, or the assignment of linebreak class may be changed for some characters. Implementations should not make any assumptions to the contrary.  Any future updates will be reflected in the latest version of the data file.

5.1 Description of Line Breaking Properties

Line breaking properties are listed alphabetically. Each property is marked with an annotation in parenthesis for easy reference showing that

(A) - the property allows a break opportunity after in specified contexts

(XA) - the property prevents a break opportunity after in specified contexts

(B) - the property allows a break opportunity before in specified contexts

(XB) - the property prevents a break opportunity before in specified contexts

(P) - the property allows a break opportunity for a pair of same characters

(XP) - the property prevents a break opportunity for a pair of same characters

NOTE: The use of the letters B and A in these annotations marks the position of the break opportunity relative to the character. It is not to be confused with the use of the same letters in the other parts of this document, where they indicate position of the characters relative to the break opportunity.  

AI - Ambiguous (Alphabetic or Ideograph)

Characters with East Asian Width property A (ambiguous width), and which would otherwise be AL in this classification. They take on the AL line break class only when their resolved width is N (narrow) and take the ID line break class, when their resolved width is W (wide). For more information on East Asian Width, and how to resolve it, see Unicode Standard Annex #11, East Asian Width [EAW]. In the absence of information needed to resolve their East Asian Width, they are treated as class AL.

AL - Ordinary Alphabetic and Symbol Characters (XP)

Require other characters to provide break opportunities, otherwise no breaking between pairs of ordinary characters. However, this is tailorable. In some Far Eastern documents it may be desirable to allow breaking between pairs of ordinary characters.

NOTE:  use ZWSP as a manual override to provide break opportunities around alphabetic or symbol characters.

ALPHABETIC — all characters of General Categories Lu, Ll, Lt, Lm, Lo, except as they appear below.
SYMBOLS — all characters of General Categories Sm, Sk, So, except as they appear below.

BA - Break Opportunity After (A)

Like SP, but are not part in determining indirect breaks.

Breaking Spaces

These characters with General Category Zs

The preceding list of space characters all have a specific width, but behave otherwise as breaking spaces. See also the SP property.

Tabs

Except for the effect of the location of the tabstops, the tab character acts similarly to a space for the purpose of line breaking.

Breaking Hyphens

Breaking hyphens establish explicit break opportunities immediately after each occurrence.

There are three types of hyphens: Explicit hyphens, conditional hyphens, and dictionary-inserted hyphens (as a result of a hyphenation process). There is no character code for the third kind of hyphen; therefore if it is desired to make the distinction, the fact that a hyphen is dictionary-inserted must be represented out of band, or by using a privately assigned control code instead of SHY.

Hyphens are graphic characters with width. Since, unlike spaces, they print, they are included in the measured part of the preceding line

SHY is rendered invisibly and has no width, except at a line break. The rendering of the soft hyphen depends on the script. For the Latin script it is rendered as a hyphen, however, some languages require a change in spelling surrounding an optional hyphen, if it occurs at a line break. For example in Swedish the word “tuggummi” changes to “tugg-gummi” when hyphenated.

The action of a hyphenation algorithm is equivalent to the insertion of a SHY. However, when a word contains an explicit SHY it is customarily treated as overriding the action of the hyphenator for that word.

The Tibetan thseg is a visible mark, but it functions effectively like a space to separate words (or other units) in Tibetan. It provides a break opportunity after itself, like space.

The Ethiopian word space is a visible word delimiter and is kept on the line before.

The Ogham space mark is rendered visibly between words but should be elided at the end of a line.

Hyphenation point is primarily used to visibly indicate syllabification of words. Syllable breaks are potential line breaking opportunities in the middle of words. The hyphenation point It is mainly used in dictionaries and similar works. When an actual line break falls inside a word containing hyphenation point characters, the hyphenation point is rendered as a regular hyphen at the end of the line.

In some dictionaries, a vertical bar is used instead of a hyphenation point. In this usage, U+0323 COMBINING DOT BELOW is used to mark stressed syllables, so all breaks are marked by the vertical bar. For an actual break opportunity, the vertical bar is rendered as a hyphen.

BB - Break opportunities before characters (B)

In some dictionaries, stressed syllables are indicated with a spacing acute accent instead of the hyphenation point. In this case the accent would move to the next line, and the preceding line ended with a hyphen.

These characters are used in dictionaries to indicate stress and secondary stress when IPA is used. Both are prefixes to the stressed syllable in IPA. Therefore, the only sensible way to break them is to keep them with the syllable; that is to break before them.

NOTE: It is hard to find actual examples in most dictionaries, since the pronunciation fields usually occur right after the headword, and the columns are wide enough to prevent line breaks in most pronunciations.

The Mongolian Todo soft hyphen indicates an optional line break opportunity with hyphen, but unlike the soft-hyphen it stays with the following line.

B2 - Break Opportunity Before and After (B/A/XP)

The em dash character is used to set off parenthetical text, normally without spaces, however, this is language dependent, for example, in Swedish, spaces are used. Line breaks can occur before and after an em dash, but not between two em dashes. Pairs of em dashes are sometimes used instead of a single quotation dash. For that reason, the line should not be broken between em dashes.

BK - Mandatory Break (A) - (normative)

Explicit breaks act independently of the surrounding characters.

Form Feed separates pages. The text on the new page starts at the beginning of the line. No paragraph formatting is applied.

The text after the Line Separator starts at the beginning of the line. No paragraph formatting is applied.

This is similar to HTML <BR>

The text of the new paragraph starts at the beginning of the line. Paragraph formatting is applied.

"NEW LINE FUNCTION (NLF)"

New line functions provide additional explicit breaks. They are not individual characters, but are expressed as sequences of control characters NEL, LF, and CR. What particular sequence(s) form a NLF depends on the implementation and other circumstances as described in Unicode Technical Report 13, Unicode Newline Guidelines.

If a character sequence for a new line function contains more than one character, it is kept together. The default behavior is to break after LF or CR, but not between CR and LF. Two additional line break classes have been added for convenience in this operation.

CB - Contingent Break Opportunity (B/A) - (normative)

Contingent Break Opportunity Before and After

By default there is a break opportunity both before and after the object. Object-specific line break behavior is implemented in the associated object itself, and where available can override the default to prevent either or both of the break opportunities.

CL - Closing Punctuation (XB)

The closing character of any set of paired punctuation must be kept with the preceding character

plus any characters of General Category Pe in the Unicode Character Database.

CM - Attached Characters and Combining Marks (XB) - (normative)

Combining characters

Combining character sequences are treated as units for the purposes of line breaking. The line-breaking behavior of the sequence is that of the base character. If U+0020 SPACE is used as a base character, it is treated as AL instead of SP.

All characters with General Category Mc,  Me, and Mn.

Conjoining Jamos (non-initial)

A sequence of conjoining Jamos is used to make up a Hangul syllable. Breaks are only allowed around the entire Hangul syllable, and then the line break properties are the same for precomposed Hangul syllables as for conjoined sequence of Jamos.

NOTE: for the purpose of determining line break opportunities, non-initial conjoining Jamos behave like combining marks, while the initial combining Jamos have the same property as Hangul Syllables.  

Control and formatting characters

Most controls and formatting characters are ignored in line breaking and do not contribute to the line width. By giving them class CM, the line breaking behavior of the last preceding character that is not of class CM affects the line breaking behavior. 

NOTE: When control codes and format characters are rendered visibly during editing, more graceful layout might be achieved by assigning them the AL or ID class instead. 

All characters of General Category Cc and Cf, unless listed explicitly elsewhere.

CR - Carriage Return (A) - (normative)

Do not break if followed by a LF, mandatory break after otherwise

EX - Exclamation / Interrogation (XB)

These behave like closing characters, except in relation to postfix and ‘non-starter’ characters

GL - Non-breaking ("Glue") (XB/XA) - (normative)

The action of these characters is to glue together both left and right neighbor character such that they are kept on the same line. If they follow a space character, they still allow a break.

Since this character is not visible, it is the preferred choice for keeping characters together that would otherwise be split across the line at a direct break.

NO-BREAK SPACE is the preferred character to use where two words should be visually separated but kept on the same line, as in the case of a title and a name "Dr.<NBSP>Joseph Becker". NARROW NO-BREAK SPACE is used in Mongolian.

This is the preferred space to use in numbers. It has the same width as a digit and keeps the number together for the purpose of line breaking.

This is the preferred character to use where words must be hyphenated but may not be broken at the hyphen.

This looks exactly like a Tibetan tsheg, but can be used to prevent a break. It inhibits breaking on either side, like no-break space.

Some dictionaries use a character that looks like a vertical series of four dots to indicate places where there is a syllable, but no allowable break. This character has not been encoded in Unicode yet, but is an example of a character that should be given the GL property.

HY - Hyphen (XA)

Some additional context analysis is required to distinguish usage of this character as a hyphen from the use as minus sign (or indicator of numerical range). If used as hyphen, it acts like HYPHEN.

NOTE: In some practice, runs of HYPHEN-MINUS are used to stand in for longer dashes or horizontal rules. If it is desired to treat them like the characters or layout elements they stand for, and actual character code conversion is not performed, line breaking will need to support these special cases explicitly.

ID - Ideographic (B/A)

NOTE: The name ideographic for this line breaking class was chosen pars pro toto. The actual set of characters in this class includes characters other than Han ideographs.

Do not require other characters to provide break opportunities, can ordinarily break before and after and between pairs.

plus all of the FULLWIDTH LATIN letters and all of the 3000-33FF blocks not covered elsewhere

NOTE:  use ZWNBSP as a manual override to prevent break opportunities around characters of class ID.

IN - Inseparable characters (XP)

Leaders

These characters are intended to be used in consecutive sequence. They therefore prevent line breaks absolutely in a series of two character of this class.

Horizontal ellipsis can be used as a three-dot leader.

IS - Numeric Separator (Infix) (XB)

Characters that usually occur inside a numerical expression may not be separated from following numeric characters, unless space character intervenes. Since they are otherwise sentence ending punctuation, they prevent breaks before.

There is no break in “100.00” or “10,000”, nor in “12:59”

LF - Line Feed (A) - (normative)

There is a mandatory break after any LF character.

NS - Non-starters (XB)

Some characters cannot start a line, but unlike CL they may allow a break in some context when they are following one or more space characters.

All characters with General Category Lm (Letter, Modifier) and East Asian Width type W or H, and all characters with General Category Sk (Symbol, Modifier) and East Asian width type W plus the following characters:

Plus all Hiragana, Katakana, and Halfwidth Katakana “small” characters

NOTE: Optionally, the NS restriction may be relaxed and characters treated like ID, to achieve a more permissive style of line breaking.

NU - Numeric (XP)

Behave like ordinary characters in the context of ordinary characters, activate the prefix and postfix behavior of prefix and postfix characters

DECIMAL DIGITS (All characters of General Category Nd, except FULL WIDTH)

OP - Opening Punctuation (XA)

The opening character of any set of paired punctuation must be kept with the following character

Characters of General Category Ps in the Unicode Character Database.

PO - Postfix (Numeric) (XB)

Characters that usually follow a numerical expression may not be separated from preceding numeric characters or preceding closing characters, even if one or more space characters intervene.

For example, there is no break in “(12.00) %”

The list of post-fix characters is:

PR - Prefix (Numeric) (XA)

Characters that usually precede a numerical expression may not be separated from following numeric characters or following opening characters, EVEN if space character intervenes.

There is no break in “$ (100.00)”

All currency symbols (General Category Sc) except as listed explicitly in PO and the following:

QU - Ambiguous Quotation (XB/XA)

Some paired characters can be either opening or closing depending on usage. The default is to treat them as both opening and closing.

NOTE: If language information is available, it can be used to determine which character is used as opening and which as closing quote. (See the information in the Unicode Standard, Version 3.0, Chapter 6. [U3.0])

Characters of General Category Pf or Pi in the Unicode Character Database as well as:

SA - Complex-context Dependent Characters (South East Asian) (P)

Runs of these characters require morphological analysis to determine break opportunities. This is similar to e.g. a hyphenation algorithm. For the characters that have this property, no line breaks will be found otherwise, therefore complex context analysis is mandatory.

NOTE: These characters can be mapped into their equivalent line break classes as result of dictionary lookup, thus permitting a logical separation of this algorithm from the morphological analysis.

If dictionary lookup is not available they should be treated as XX.

All characters of General Category Lo or Lm in these ranges:

SG - Surrogates (XP) - (normative)

All characters with General Category Cs. There is no break between a high surrogate and a low surrogate.

NOTE: The line breaking class for a pair of surrogates depends on the character. A useful default is to treat characters in the range 0x00010000 to 0x0001FFFD as AL and characters in the range 0x00020000 to 0x0002FFFD as ID, until the implementation can be revised to take into account the actual line breaking properties for these characters, many of which have yet to be assigned.

SP - Space (A) - (normative)

The space characters are explicit break opportunities, but spaces at the end of a line are not measured for fit. If there is a sequence of space characters, and breaking after any of the space characters would result in the same visible line, the line breaking position after the last space character in the sequence is the locally most optimal one. In other words, since the last character measured for fit is BEFORE the space character, any number of space characters are kept together invisibly on the previous line and the first non-space character starts the next line.

NOTE: SPACE, but none of the other breaking spaces, is used in determining an indirect break.

SY - Symbols Allowing Break After (A)

URLs are common enough now in regular plain text, that they must be taken into account when assigning general-purpose line breaking properties. The SY line break property is intended to provide a break after, but not in front of digits so as to not break “1/2” or “06/07/99”.

Slash (SOLIDUS) is allowed as an additional, limited break opportunity to improve layout of web addresses

NOTE: Normally, symbols are treated as AL. If it is desired to allow other breaks, more symbols can be added to this line breaking class, or classes BA, BB, B2 by tailoring, for example “=”.  Mathematics requires additional specifications for line breaking, which are outside the scope of this document.

XX - Unknown (XP)

Unassigned code positions, private use characters and characters for which reliable line breaking information is not available are assigned this default line breaking property. The default behavior for this class is identical to class AL. Users can manually insert ZWSP or ZWNBSP around characters of class XX to force or prevent breaks as needed. 

In addition, implementations can override or tailor this default behavior, e.g. by assigning characters the property ID or another class, if that is more likely to give the correct default behavior for their users, or use other means to determine the correct behavior. For example one implementation might treat any private use character in ideographic context as ID, while another implementation might support a method for assigning specific properties to specific definitions of private use characters. The details of such use of private use characters are outside the scope of this standard.

All characters with General Category Co and all codepoints with General Category Cn.

ZW - Zero Width Space (A) - (normative)

This character does not have width. It is used to enable additional (invisible) break opportunities wherever SPACE cannot be used.

5.2 Additional Details on Dictionary Usage

Dictionaries follow strict standards that guide their use of characters to indicate features of the terms listed. Some of these conventions mark places that can also serve as line breaking opportunities and therefore interact with line breaking and are described here. Where appropriate, these characters have been inserted in the list of characters for the corresponding line breaking property above.

However, implementing the full conventions in dictionaries requires special support. Looking up the noun “syllable” in eight dictionaries yields eight different conventions, in one dictionary a natural hyphen in a word becomes a tilde dash if the word is split.

Dictionary of the English Language, Samuel Johnson, 1843 SY´LLABLE where ´ is a U+02B9 (and a large one at that) and follows the vowel of the main syllable (not the syllable itself).

Oxford English Dictionary (1st Edition) si·lâ'bl where · is a slightly above middle dot indicating the vowel of the stressed syllable (similar to Johnson's acute). The letter â is really U+0103. The ' is an apostrophe.

Oxford English Dictionary (2nd Edition) has gone to IPA 'sIleb(e)l where ' is U+02C8, I is U+026A, e is U+0259 (both times). The ' comes before the stressed syllable. The () indicate the schwa may be omitted.

Chambers English Dictionary (7th Edition) sil´e-bl where the stressed syllable is followed by ´  U+02B9, e is U+0259, - is a hyphen when splitting a word like abate´- ment the stress mark ´ goes after stressed syllable followed by the hyphen. No special convention if splitting at hyphen.

BBC English Dictionary sIlebl where I is U+026A U+0332, e is U+0259. The vowel of the stressed syllable is underlined.

Collins Cobuild English Language Dictionary sIlebe°l where I is U+026A U+0332, and means the same as the BBC. The e is U+0259 (both times). The ° is a U+2070 and indicates the schwa may be omitted.

Readers Digest Great Illustrated Dictionary. syl·la·ble (sílleb'l) The spelling of the word has hyphenation points (· is a U+2027) followed by phonetic spelling. The vowel of the stressed syllable is given an accent (rather than being followed by an accent). The letter e is a schwa in the actual example and ' is apostrophe.

Webster's 3rd New International Dictionary.  syl·la·ble /'silebel/ The spelling of the word has hyphenation points (· is a U+2027) and is followed by phonetic spelling. The stressed syllable is preceded by ' U+02C8. The e's are schwas as usual. Webster splits words at the end of a line with a normal hyphen. When a hyphenated word is split at the hyphen this is indicated by a double hyphen which looks like a light version of the German Fraktur hyphen (short equals sign with a slight slope up to the right).

6 Line Breaking Algorithm

The Unicode Standard, Version 3.0, describes a particular method for boundary detection in Chapter 5. It is based on a set of hierarchical rules and character classifications. That method would be well suited for implementation of some of the advanced heuristics.

A slightly simplified implementation of that can be devised that uses a two dimensional table to resolve break opportunities between pairs or characters. It is described in the following section.

The line breaking algorithm presented in this section can be expressed in a series of rules which take line breaking classes as input.

Line breaking rules

The line breaking rules are stated in terms of regular expressions over the line breaking classes defined above and three special symbols indicating the type of line break opportunity.

! Mandatory break at the indicated position

× No break allowed at the indicated position

÷ Break allowed at the indicated position

The rules are applied in order. That is, there is an implicit ”otherwise” at the front of each rule following the first. It is possible to construct alternate sets of such rules that are fully equivalent, i.e. they have the same effect.

The distinction between direct and indirect break is handled by explicitly considering the effect of SP in rule LB12. Because rules are applied in order, rule LB12 implies that a prohibited break in rules 13-19 is equivalent to an indirect break.

The examples for each rule use representative characters, where ’H’  stands for an ideographs, ’h’ for small kana, ’9’ for digits.

Resolve line break classes:

LB 1  Assign a line break category to each character of the input. Resolve AI, CB, SA, SG, XX into other line break classes depending on criteria outside this algorithm.

Start and end of text:

LB 2a  Never break at the start of text

× sot

LB 2b  Always break at the end of text

! eot

These two rules are designed to deal with degenerate cases. Their effect is to have at least one character on each line, and at least one line break for the whole text. Emergency line breaking behavior usually also allows line breaks anywhere on the line if a legal line break cannot be found. This has the effect of preventing text to run over the margins.

Mandatory breaks:

LB 3a  Always break after hard line breaks (but never between CR and LF).

BK !

CR !

LF !

CR × LF

LB 3b  Don’t break before hard line breaks.

× ( BK | CR | LF )

Explicit breaks and non-breaks:

LB 4  Don’t break before spaces or zero-width space.

× SP

× ZW

LB 5 Break after zero-width space.

ZW ÷

Combining Marks:

At any possible break opportunity between CM and a following character, CM behaves as if it had the type of its base character. Virama are treated as CM so they work correctly. Non-initial Jamo are treated as CM, so each syllable inherits the linebreak class of the initial Jamo and no breaks can occur in the middle of a syllable.

LB 6  Don’t break graphemes (before combining marks, around virama or on sequences of conjoining Jamos.

Treat X CM* as if it were X

(See the Unicode Standard [U3.0] for other rules regarding graphemes.)

As stated in section 7.9 of The Unicode Standard, Version 3.0 [U3.0], combining characters are shown in isolation by applying them to either U+0020 SPACE (SP) or U+00A0 NO BREAK SPACE. The visual appearance is the same, but the line breaking result is different. Correspondingly, if there is no base, or if the base character is SP, CM* or SP CM* behave like ID.

LB 7  In all of the following rules, if a space is the base character for a combining mark, the space is changed to type ID. In other words, break before SP CM* in the same cases as one would break before an ID.

Treat SP CM* as if it were ID

Opening and closing:

These have special behavior with respect to spaces, and so come before rule 12.

LB 8  Don’t break before ‘]’ or ‘!’ or ‘;’ or ‘/’,  even after spaces.

× CL

× EX

× IS

× SY

LB 9  Don’t break after ‘[’, even after spaces.

OP SP* ×

LB 10  Don’t break within ‘”[’, , even with intervening spaces.

QU SP* × OP

LB 11  Don’t break within ‘]h’, even with intervening spaces.

CL SP* × NS

LB 11a  Don’t break within ‘——’, even with intervening spaces.

B2 × B2

Spaces:

LB 12  Break after spaces

SP ÷

Special case rules:

LB 13  Don’t break before or after ZWNBSP

× GL

GL ×

LB 14  Don’t break before or after ‘”’

 × QU

QU ×

LB 15  Don’t break before hyphen-minus, other hyphens, fixed-width spaces, small kana and other non- starters,  or after acute accents:

× BA

× HY

× NS

BB ×

LB 15b  Break after hyphen-minus, and before acute accents:

HY ÷

÷ BB

LB 16  Don’t break between two ellipses, or between letters or numbers and ellipsis:

AL × IN

ID × IN

IN × IN

NU × IN

Examples: ’9...’, ‘a...’, ‘H...’

Numbers:

Don't break alphanumerics.

LB 17  Don’t break within ‘a9’, ‘3a’, or ‘H%’

ID × PO

AL × NU

NU × AL

Numbers are of the form PR ? ( OP | HY ) ? NU (NU | IS) * CL ?  PO ?

Examples:   $(12.35)    2,1234    (12)¢    12.54¢

This is approximated with the following rules. (Some cases already handled above, like ‘9,’, ‘[9’.)

LB 18  Don’t break between the following pairs of classes.

CL × PO

HY × NU

IS × NU

NU × NU

NU × PO

PR × AL

PR × HY

PR × ID

PR × NU

PR × OP

SY × NU

Example pairs: ‘$9’, ‘$[’, ‘$-‘, ‘-9’, ‘/9’, ‘99’, ‘,9’,  ‘9%’ ‘]%’

Finally, join alphabetic letters and break everything else.

LB 19  Don’t break between alphabetics (“at”)

AL × AL

LB 20  Break everywhere else

ALL ÷

÷ ALL

7 Pair-Table Based Implementation

A two dimensional table can be used to resolve break opportunities between pairs of characters. The rows of the table are labeled by the possible values of the line breaking property of the leading character in the pair; the columns are labeled by the line breaking property for the following character of the pair. Each intersection is labeled with the resulting line breaking opportunity.

The Japanese standard JIS X 4051-1995 [JIS] provides an example of such a table-based definition. However, it uses line breaking classes whose membership is not solely determined by line breaking property (as in this report), but in some cases by heuristic analysis or markup of the text.

The implementation provided here directly uses the line breaking classed defined above. 

7.1 Minimal Table

If two rows of the table have identical values and the corresponding columns also have identical values, the two line breaking classes can be coalesced. For example, the JIS standard uses 20 classes of which only 14 appear to be unique. A minimal table representation is unique, except for trivial reordering of rows and columns.

7.2 Extended Context

Rules LB 6, and LB 8 - LB11 require extended context for handling combining marks and spaces. This extended context must be built into the code that interprets the pair table.

By broadening the definition of pair from B A to B SP* A, where A and B are characters and SP* is an optional run of space characters, the same table can be used to distinguish between cases where SP can or cannot provide a line breaking opportunity (i.e. direct and indirect breaks). Rules equivalent to the ones given in section 6 can be formulated without explicit use of SP, by instead using % to express indirect breaks. These rules can then be simplified to involve only pairs of classes, e.g. only constructions of the form

B ÷ A

B % A

B ^ A

where either A or B may be empty. These simplified rules can then be automatically translated into a pair table, as in the example below. Line break analysis then proceeds by pair table lookup. 

7.3 Example Table

The following example table implements the line breaking behavior described in this Technical Report, within the limitation that only context of the form B SP* A is considered. BK and SP classes are handled explicitly in the outer loop as given in the code sample below. B CM* can be handled in the table, or explicitly in the driving loop, as explained in section 7.5. Using the table entries is equivalent to making the simplifying assumption that combining marks are only applied to AL. (Such an assumption does not hold when conjoining Jamos are used).

• ^ denotes a prohibited break: Never break here, even if one or more spaces intervene.
  • As a reminder, A ^ B is equivalent to B SP* × A.
• % denotes an indirect break opportunity: Don’t break here, unless one of more spaces intervene.
  • As a reminder, A % B is equivalent to B × A and B SP+ ÷ A.
• _ an empty cell denotes a direct break opportunity (÷ above). These are blank to make them easier to distinguish in the table.

7.4 Sample Code

The following two functions demonstrate how the pair table is used. For a complete implementation of the line breaking algorithm, if statements to handle the following line break classes need to be added: CR, LF, CB, SG, XX. They have been omitted here for brevity.

    // placeholder function for complex break analysis
    int findComplexBreak(int *pcls, int *pbrk, int cch)
    {
         if (!cch)
               return 0;
         int cls = pcls[0];
         for(int ich = 0; ich < cch; ich++) {
 
                  // .. do complex break analysis here
 
                  if (pcls[ich] != SA)
                         break;
                  }
         return ich;
    }

    enum break_action { 
       DBK = 0, // direct break     (blank in table)
       IBK,     // indirect break   (% in table)
       PBK };   // prohibited break (^ in table)

    // handle spaces separately, all others by table
    // pcls - pointer to array of line break classes (input)
    // pbrk - pointer to array of line break opportunities (output)
    // cch - number of elements in the arrays (“count of characters”) (input)
    // ich - current index into the arrays (variable)
    int findLineBrk1(int *pcls, int *pbrk, int cch)
    {
         if (!cch) 
              return O;
 
         int  cls = pcls[0];

         // loop over all pairs in the string
         for (int ich = 1; (ich < cch) && (cls != BK); ich++) {
 
              // handle spaces
              if (pcls[ich] == SP) {
                   pbrk[ich-1] = PBK;
                   continue;
              }
 
              // handle complex scripts
              if (pcls[ich] == SA) {
                    ich += findComplexBreak(&pcls[ich-1], &pbrk[ich-1], cch - (ich-1));
                    if (ich < cch)
                         cls = pcls[ich];
                    continue;
              }
 
              // lookup pair table information
              int brk = brkPairs[cls][pcls[ich]];
 
              if (brk == IBK) {
                    pbrk[ich-1] = ((pcls[ich - 1] == SP) ? IBK : PBK);
              } else {
                   pbrk[ich-1] = brk;
              }
              cls = pcls[ich];
        }
        // always break at the end
        pbrk[ich-1] = DBK;
 
        return ich;
   }

The function returns all the break opportunities in the array pointed to by pbrk, using the values in the table. The code assumes that the predefined value SS is used for break opportunities marked by an % entry in the table and the value XX for an entry marked by an ^ above.

7.5 Combining Marks

If one makes the simplifying assumption that combining marks are only applied to AL, or that applying a combining mark turns the combination into AL, then CM can be handled in the table as shown. (Such an assumption does not hold when conjoining Jamos are used). 

Therefore it is preferable to handle CM outside of the pair table in the driver code. Adding a simple statement in the loop

    // handle combining marks
    if (pcls[ich] == CM){
       if(pcls[ich-1] == SP){
          cls = ID;
          if (ich > 1)
	     pbrk[ich-2] = brkPairs[pcls[ich-2]][ID] == DBK ? DBK : PBK;
       }
       pbrk[ich-1] = PBK;
       continue;
    }

would have the effect of letting the CM take on the class of the preceding non-CM characters. It also takes care of rule LB7, treating a combining mark applied to a SP as if it was ID. This also requires a statement in the setup part before the loop to cover the case of a missing base character at the beginning of the line:

    // handle missing base character
    if (cls == CM)
            cls = ID;

7.6 Customization

A real world line breaking algorithm must be tailorable to some degree. There are three principal ways of tailoring a pair-table based algorithm:

1. Changing the line breaking class assignment for some characters. This is useful for cases where the line breaking properties of one class of characters are occasionally lumped together with the properties of another class to achieve a less restrictive line breaking behavior. 
   
2. Changing the table value assigned to a pair of character classes. This is particularly useful if the behavior can be expressed by a change at a limited number of pair intersections. This form of customization is equivalent to overriding some of the rules in section 6.
   
3. Changing the interpretation of the line breaking actions is equivalent to the preceding, but instead of changing the values for the pair intersection directly in the table, they can be labeled with special values that cause different actions for different customizations. This is most suitable when customizations need to be done at run time.

Beyond these three straightforward customization steps, it is always possible to augment the algorithm itself, which places no limits to possible changes.

7.7 Examples of Customization

Example 1. One method of implementing line breaks for complex scripts is to invoke context-based classification for all runs of characters with class SA. For example a dictionary-based algorithm could return different classes for Thai letters depending on their context: letters at the start of Thai words would become BB and other Thai letters would become AL. The sample code sketches a different approach where the dictionary-based algorithm directly reports break opportunities.

Example 2. To implement terminal style line breaks, it would be necessary to allow breaks inside a run of spaces. This cannot be done in the pair-table, but requires a change in the way the driver loop handles spaces.

Example 3. Depending on the nature of the document, Korean uses either implicit breaking around characters (type 2 as defined above in section 3 Description) or uses spaces (type 1). Space based layout is common in informal documents with ragged margins, such as magazines, while books, with both margins justified, use the other type, as it affords more line break opportunities and therefore leads to better justification. Reference [Suign98] shows how the necessary customizations can be elegantly handled by selectively altering the interpretation of the pair entries. Only the intersection of ID/ID, AL/ID and ID/AL are affected. For alphabetic style line breaking, breaks for these four cases require space, for ideographic style line breaking, these four cases do not require spaces. Therefore, he defines a pseudo-action, which is then resolved into either direct or indirect break action based on user selection of the preferred behavior for a given (piece of) text.

Example 4. Sometimes allowing alphabetic characters and digit strings to break anywhere is required in Far Eastern context. According to reference [Suign98] this can again be done in the same way, this time affecting the intersections of NU/NU, NU/AL, AL/AL, and AL/NU.

Example 5. Some users prefer to force Kana syllables to be kept together, i.e. the syllable kyu, spelled with the two kanas KI and "small yu” would be kept together even though KI and yu are normally atomic. This customization can be handled via the first method, by changing the classification of the Kana small characters from ID to NS as needed.

Reference [Cedar97] reports on a real world implementation of a pair-table based implementation of a line breaking algorithm substantially similar to the one presented here, and including the types of customizations presented in this section. This implementation was able to simultaneously meet the requirements of customers in many European and East Asian countries with a single implementation of the algorithm.

References

Acknowledgments

The initial assignments of properties are based on input by Michel Suignard. Mark Davis provided algorithmic verification and formulation of the rules. Ken Whistler, Rick McGowan and other members of the editorial committee provided valuable feedback. Tim Partridge enlarged the information on dictionary usage. Sun Gi Hong reviewed the information on Korean and provided copious printed samples. 

Modifications

Change from Revision 9:

• Change header for publication of Unicode 3.1. Fixed a few additional typos.

Change from Revision 8: 

• Fixed several typos, reformatted and sorted some lists by code points
• Reconciled the data file and the description for BB (00B4), XX (PUA), AI (2015,25C8,PUA), ID (FE6B), BA (00B4)
• Restored PUA to XX
• LB 7: Restored the rule, and fixed the note so it matches the rule and section 7.9 of [U3.0].
• LB11a: added a rule to reconcile the rules against pair-table entry B2 ^ B2
• LB19: added an entry to reconcile the rules against pair-table entry PR % ID
• Reworked section 7.5
• Removed two unused definitions (overfull and underfull)

Change from Revision 7: 

• New status section, changed format of references. Fixed several typos.
• Added headers to table 1
• Added a note on use of B and A
• Added mention of PUA to AI and removed mention of PUA from XX since the data file assigns AI to them.
• Clarified the membership and implication of class CM and ID.
• Updated class ID by the new ranges for 3.1.
• LB 6: Clarified the description of LB6 to clarify how it affects conjoining Jamo .
• LB 7: Fixed the note so it matches the rule.
• LB 17: Fixed the regular expression for numbers in the explanation for this rule.
• Reworded sections 7.6 and 7.7 to clarify the customization process.

Change from Revision 6: 

• Fixed several typos.
• New header.

Change from Revision 5: 

• Rewrite and reorganization of the text as part of the publication of the Unicode Standard, Version 3.0.

Copyright © 1998-2001 Unicode, Inc. All Rights Reserved. The Unicode Consortium makes no expressed or implied warranty of any kind, and assumes no liability for errors or omissions. No liability is assumed for incidental and consequential damages in connection with or arising out of the use of the information or programs contained or accompanying this technical report.

Unicode and the Unicode logo are trademarks of Unicode, Inc., and are registered in some jurisdictions.