1UNICODE(7) Linux Programmer's Manual UNICODE(7)
2
6 Unicode - the Universal Character Set
7
9 The international standard ISO 10646 defines the Universal Character
10 Set (UCS). UCS contains all characters of all other character set
11 standards. It also guarantees round-trip compatibility, i.e., conver‐
12 sion tables can be built such that no information is lost when a string
13 is converted from any other encoding to UCS and back.
14 UCS contains the characters required to represent practically all known
16 languages. This includes not only the Latin, Greek, Cyrillic, Hebrew,
17 Arabic, Armenian, and Georgian scripts, but also Chinese, Japanese and
18 Korean Han ideographs as well as scripts such as Hiragana, Katakana,
19 Hangul, Devanagari, Bengali, Gurmukhi, Gujarati, Oriya, Tamil, Telugu,
20 Kannada, Malayalam, Thai, Lao, Khmer, Bopomofo, Tibetan, Runic,
21 Ethiopic, Canadian Syllabics, Cherokee, Mongolian, Ogham, Myanmar, Sin‐
22 hala, Thaana, Yi, and others. For scripts not yet covered, research on
23 how to best encode them for computer usage is still going on and they
24 will be added eventually. This might eventually include not only Hiero‐
25 glyphs and various historic Indo-European languages, but even some
26 selected artistic scripts such as Tengwar, Cirth, and Klingon. UCS also
27 covers a large number of graphical, typographical, mathematical and
28 scientific symbols, including those provided by TeX, Postscript, APL,
29 MS-DOS, MS-Windows, Macintosh, OCR fonts, as well as many word process‐
30 ing and publishing systems, and more are being added.
31 The UCS standard (ISO 10646) describes a 31-bit character set architec‐
33 ture consisting of 128 24-bit groups, each divided into 256 16-bit
34 planes made up of 256 8-bit rows with 256 column positions, one for
35 each character. Part 1 of the standard (ISO 10646-1) defines the first
36 65534 code positions (0x0000 to 0xfffd), which form the Basic Multilin‐
37 gual Plane (BMP), that is plane 0 in group 0. Part 2 of the standard
38 (ISO 10646-2) adds characters to group 0 outside the BMP in several
39 supplementary planes in the range 0x10000 to 0x10ffff. There are no
40 plans to add characters beyond 0x10ffff to the standard, therefore of
41 the entire code space, only a small fraction of group 0 will ever be
42 actually used in the foreseeable future. The BMP contains all charac‐
43 ters found in the commonly used other character sets. The supplemental
44 planes added by ISO 10646-2 cover only more exotic characters for spe‐
45 cial scientific, dictionary printing, publishing industry, higher-level
46 protocol and enthusiast needs.
47 The representation of each UCS character as a 2-byte word is referred
49 to as the UCS-2 form (only for BMP characters), whereas UCS-4 is the
50 representation of each character by a 4-byte word. In addition, there
51 exist two encoding forms UTF-8 for backwards compatibility with ASCII
52 processing software and UTF-16 for the backwards compatible handling of
53 non-BMP characters up to 0x10ffff by UCS-2 software.
54 The UCS characters 0x0000 to 0x007f are identical to those of the clas‐
56 sic US-ASCII character set and the characters in the range 0x0000 to
57 0x00ff are identical to those in ISO 8859-1 Latin-1.
58
60 Some code points in UCS have been assigned to combining characters.
61 These are similar to the non-spacing accent keys on a typewriter. A
62 combining character just adds an accent to the previous character. The
63 most important accented characters have codes of their own in UCS, how‐
64 ever, the combining character mechanism allows us to add accents and
65 other diacritical marks to any character. The combining characters
66 always follow the character which they modify. For example, the German
67 character Umlaut-A ("Latin capital letter A with diaeresis") can either
68 be represented by the precomposed UCS code 0x00c4, or alternatively as
69 the combination of a normal "Latin capital letter A" followed by a
70 "combining diaeresis": 0x0041 0x0308.
71 Combining characters are essential for instance for encoding the Thai
73 script or for mathematical typesetting and users of the International
74 Phonetic Alphabet.
75
77 As not all systems are expected to support advanced mechanisms like
78 combining characters, ISO 10646-1 specifies the following three imple‐
79 mentation levels of UCS:
80 Level 1 Combining characters and Hangul Jamo (a variant encoding of
82 the Korean script, where a Hangul syllable glyph is coded as a
83 triplet or pair of vovel/consonant codes) are not supported.
84 Level 2 In addition to level 1, combining characters are now allowed
86 for some languages where they are essential (e.g., Thai, Lao,
87 Hebrew, Arabic, Devanagari, Malayalam, etc.).
88 Level 3 All UCS characters are supported.
90 The Unicode 3.0 Standard published by the Unicode Consortium contains
92 exactly the UCS Basic Multilingual Plane at implementation level 3, as
93 described in ISO 10646-1:2000. Unicode 3.1 added the supplemental
94 planes of ISO 10646-2. The Unicode standard and technical reports pub‐
95 lished by the Unicode Consortium provide much additional information on
96 the semantics and recommended usages of various characters. They pro‐
97 vide guidelines and algorithms for editing, sorting, comparing, normal‐
98 izing, converting and displaying Unicode strings.
99
101 Under GNU/Linux, the C type wchar_t is a signed 32-bit integer type.
102 Its values are always interpreted by the C library as UCS code values
103 (in all locales), a convention that is signaled by the GNU C library to
104 applications by defining the constant __STDC_ISO_10646__ as specified
105 in the ISO C 99 standard.
106 UCS/Unicode can be used just like ASCII in input/output streams, termi‐
108 nal communication, plaintext files, filenames, and environment vari‐
109 ables in the ASCII compatible UTF-8 multi-byte encoding. To signal the
110 use of UTF-8 as the character encoding to all applications, a suitable
111 locale has to be selected via environment variables (e.g.,
112 "LANG=en_GB.UTF-8").
113 The nl_langinfo(CODESET) function returns the name of the selected
115 encoding. Library functions such as wctomb(3) and mbsrtowcs(3) can be
116 used to transform the internal wchar_t characters and strings into the
117 system character encoding and back and wcwidth(3) tells, how many posi‐
118 tions (0–2) the cursor is advanced by the output of a character.
119 Under Linux, in general only the BMP at implementation level 1 should
121 be used at the moment. Up to two combining characters per base charac‐
122 ter for certain scripts (in particular Thai) are also supported by some
123 UTF-8 terminal emulators and ISO 10646 fonts (level 2), but in general
124 precomposed characters should be preferred where available (Unicode
125 calls this Normalization Form C).
126
128 In the BMP, the range 0xe000 to 0xf8ff will never be assigned to any
129 characters by the standard and is reserved for private usage. For the
130 Linux community, this private area has been subdivided further into the
131 range 0xe000 to 0xefff which can be used individually by any end-user
132 and the Linux zone in the range 0xf000 to 0xf8ff where extensions are
133 coordinated among all Linux users. The registry of the characters
134 assigned to the Linux zone is currently maintained by H. Peter Anvin
135 <Peter.Anvin@linux.org>.
136
138 * Information technology — Universal Multiple-Octet Coded Character Set
139 (UCS) — Part 1: Architecture and Basic Multilingual Plane. Interna‐
140 tional Standard ISO/IEC 10646-1, International Organization for Stan‐
141 dardization, Geneva, 2000.
142 This is the official specification of UCS. Available as a PDF file
144 on CD-ROM from http://www.iso.ch/.
145 * The Unicode Standard, Version 3.0. The Unicode Consortium, Addison-
147 Wesley, Reading, MA, 2000, ISBN 0-201-61633-5.
148 * S. Harbison, G. Steele. C: A Reference Manual. Fourth edition, Pren‐
150 tice Hall, Englewood Cliffs, 1995, ISBN 0-13-326224-3.
151 A good reference book about the C programming language. The fourth
153 edition covers the 1994 Amendment 1 to the ISO C 90 standard, which
154 adds a large number of new C library functions for handling wide and
155 multi-byte character encodings, but it does not yet cover ISO C 99,
156 which improved wide and multi-byte character support even further.
157 * Unicode Technical Reports.
159 http://www.unicode.org/unicode/reports/
160 * Markus Kuhn: UTF-8 and Unicode FAQ for Unix/Linux.
162 http://www.cl.cam.ac.uk/~mgk25/unicode.html
163 Provides subscription information for the linux-utf8 mailing list,
165 which is the best place to look for advice on using Unicode under
166 Linux.
167 * Bruno Haible: Unicode HOWTO.
169 ftp://ftp.ilog.fr/pub/Users/haible/utf8/Unicode-HOWTO.html
170
172 When this man page was last revised, the GNU C Library support for
173 UTF-8 locales was mature and XFree86 support was in an advanced state,
174 but work on making applications (most notably editors) suitable for use
175 in UTF-8 locales was still fully in progress. Current general UCS sup‐
176 port under Linux usually provides for CJK double-width characters and
177 sometimes even simple overstriking combining characters, but usually
178 does not include support for scripts with right-to-left writing direc‐
179 tion or ligature substitution requirements such as Hebrew, Arabic, or
180 the Indic scripts. These scripts are currently only supported in cer‐
181 tain GUI applications (HTML viewers, word processors) with sophisti‐
182 cated text rendering engines.
183
185 Markus Kuhn <mgk25@cl.cam.ac.uk>
186
188 setlocale(3), charsets(7), utf-8(7)
189GNU 2001-05-11 UNICODE(7)