C0 and C1 control codes

From Wikipedia, the free encyclopedia
(Redirected from ISO 6429)

The C0 and C1 control code or control character sets define control codes for use in text by computer systems that use ASCII and derivatives of ASCII. The codes represent additional information about the text, such as the position of a cursor, an instruction to start a new line, or a message that the text has been received.

C0 codes are the range 00HEX–1FHEX and the default C0 set was originally defined in ISO 646 (ASCII). C1 codes are the range 80HEX–9FHEX and the default C1 set was originally defined in ECMA-48 (harmonized later with ISO 6429). The ISO/IEC 2022 system of specifying control and graphic characters allows other C0 and C1 sets to be available for specialized applications, but they are rarely used.

C0 controls[edit]

ASCII defined 32 control characters, plus a necessary extra character for the DEL character, 7FHEX or 01111111BIN (needed to punch out all the holes on a paper tape and erase it).

This large number of codes was desirable at the time, as multi-byte controls would require implementation of a state machine in the terminal, which was very difficult with contemporary electronics and mechanical terminals.

Only a few codes have maintained their use: BEL, ESC, and the "Format Effector" (FEn) characters BS, TAB, LF, VT, FF, and CR. Others are unused or have acquired different meanings such as NUL being the C string terminator. Some data transfer protocols such as ANPA-1312, Kermit, and XMODEM do make extensive use of SOH, STX, ETX, EOT, ACK, NAK and SYN for purposes approximating their original definitions; and some file formats use the "Information Separators" (ISn) such as the Unix info format[1] and Python's splitlines string method.[2]

The names of some codes were changed in ISO 6429:1992 (or ECMA-48:1991) to be neutral with respect to writing direction. The abbreviations used were not changed, as the standard had already specified that those would remain unchanged when the standard is translated to other languages. In this table both new and old names are shown for the renamed controls (the old name is the one matching the abbreviation).

ASCII control codes, originally defined in ANSI X3.4.[3]
Decimal
Hexadecimal
Abbreviations Name Description
^@ 0 00 NUL Null \0 Does nothing. The code of blank paper tape, and also used for padding to slow transmission.
^A 1 01 TC1, SOH Start of Heading First character of the heading of a message.[4]
^B 2 02 TC2, STX Start of Text Terminates the header and starts the message text.
^C 3 03 TC3, ETX End of Text Ends the message text, starts a footer (up to the next TC character).[4][5]
^D 4 04 TC4, EOT End of Transmission Ends the transmission of one or more messages.[4][5] May place terminals on standby.[5]
^E 5 05 TC5, ENQ, WRU[a] Enquiry Trigger a response at the receiving end, to see if it is still present.
^F 6 06 TC6, ACK Acknowledge Indication of successful receipt of a message.
^G 7 07 BEL[b] Bell, Alert \a Call for attention from an operator.
^H 8 08 FE0, BS Backspace \b Move one position leftwards. Next character may overprint or replace the character that was there.
^I 9 09 FE1, HT Character Tabulation,
Horizontal Tabulation
\t Move right to the next tab stop.
^J 10 0A FE2, LF Line Feed \n Move down to the same position on the next line (some devices also moved to the left column).
^K 11 0B FE3, VT Line Tabulation,
Vertical Tabulation
\v Move down to the next vertical tab stop.
^L 12 0C FE4, FF Form Feed \f Move down to the top of the next page.
^M 13 0D FE5, CR Carriage Return \r Move to column zero while staying on the same line.
^N 14 0E SO, LS0[c] Shift Out Switch to an alternative character set.
^O 15 0F SI, LS1[c] Shift In Return to regular character set after SO.
^P 16 10 TC7, DC0,[d] DLE Data Link Escape Cause a limited number of contiguously following characters to be interpreted in some different way.[14][15]
^Q 17 11 DC1, XON Device Control One Turn on (DC1 and DC2) or off (DC3 and DC4) devices.

Teletype[6] used these for the paper tape reader and the paper tape punch. The first use became the de facto standard for software flow control.[16]

^R 18 12 DC2, TAPE Device Control Two
^S 19 13 DC3, XOFF Device Control Three
^T 20 14 DC4, TAPE Device Control Four
^U 21 15 TC8, NAK Negative Acknowledge Negative response to a sender, such as a detected error.
^V 22 16 TC9, SYN Synchronous Idle Sent in synchronous transmission systems when no other character is being transmitted.
^W 23 17 TC10, ETB End of Transmission Block End of a transmission block of data when data are divided into such blocks for transmission purposes.
^X 24 18 CAN Cancel Indicates that the data preceding it are in error or are to be disregarded.
^Y 25 19 EM End of medium Indicates on paper or magnetic tapes that the end of the usable portion of the tape had been reached.[3]
^Z 26 1A SUB Substitute Replaces a character that was found to be invalid or in error. Should be ignored.
^[ 27 1B ESC Escape \e
[e]
Alters the meaning of a limited number of following bytes.
Nowadays this is almost always used to introduce an ANSI escape sequence.
^\ 28 1C IS4, FS File Separator Can be used as delimiters to mark fields of data structures. US is the lowest level, while RS, GS, and FS are of increasing level to divide groups made up of items of the level beneath it. SP (space) could be considered an even lower level.
^] 29 1D IS3, GS Group Separator
^^ 30 1E IS2, RS Record Separator
^_ 31 1F IS1, US Unit Separator
While not technically part of the C0 control character range, the following two characters can be thought of as having some characteristics of control characters.
  32 20 SP Space Move right one character position.
^? 127 7F DEL Delete Should be ignored. Used to delete characters on punched tape by punching out all the holes.
  1. ^ Teletype labelled the key WRU for 'who are you?'[6]
  2. ^ The name BELL is assigned by Unicode to the unrelated emoji character 🔔 (U+1F514). While C0 and C1 control characters were not formally named by the Unicode standard itself at the time, this collided with existing use of BELL as the name of this control character in software following the previous versions of UTS#18 (the Unicode Regular Expressions standard),[7] e.g. in Perl.[8] Unicode now accepts ALERT and BEL (but not BELL) as formal aliases for the control character,[9] although the code chart still lists BELL as the ISO 6429 alias,[10] and the corresponding control picture code point is called SYMBOL FOR BELL. Perl subsequently switched to using BELL for the emoji in version 5.18.[11]
  3. ^ a b ISO/IEC 2022 (ECMA-35) refers to these as LS0 and LS1 in 8-bit environments, and as SI and SO in 7-bit environments.[12]
  4. ^ The first, 1963 edition of ASCII classified DLE as a device control, rather than a transmission control, and gave it the abbreviation DC0 ("device control reserved for data link escape").[13]
  5. ^ The '\e' escape sequence is not part of ISO C and many other language specifications. However, it is understood by several compilers, including GCC.

C1 controls[edit]

In 1973, ECMA-35 and ISO 2022[17] attempted to define a method so an 8-bit "extended ASCII" code could be converted to a corresponding 7-bit code, and vice versa.[18] In a 7-bit environment, the Shift Out (SO) would change the meaning of the 96 bytes 0x20 through 0x7F[a][20] (i.e. all but the C0 control codes), to be the characters that an 8-bit environment would print if it used the same code with the high bit set. This meant that the range 0x80 through 0x9F could not be printed in a 7-bit environment,[18] thus it was decided that no alternative character set could use them, and that these codes should be additional control codes, which become known as the C1 control codes. To allow a 7-bit environment to use these new controls, the sequences ESC @ through ESC _ were to be considered equivalent.[18] The later ISO 8859 standards abandoned support for 7-bit codes, but preserved this range of control characters.

The first C1 control code set to be registered for use with ISO 2022 was DIN 31626,[21] a specialised set for bibliographic use which was registered in 1979.[22]

The more common general-use ISO/IEC 6429 set was registered in 1983,[23] although the ECMA-48 specification upon which it was based had been first published in 1976[24] and JIS X 0211 (formerly JIS C 6323).[25] Symbolic names defined by RFC 1345 and early drafts of ISO 10646, but not in ISO/IEC 6429 (PAD, HOP and SGC) are also used.[8][26]

Except for SS2 and SS3 in EUC-JP text, and NEL in text transcoded from EBCDIC, the 8-bit forms of these codes were almost never used. CSI, DCS and OSC are used to control text terminals and terminal emulators, but almost always by using their 7-bit escape code representations. Nowadays if these codes are encountered it is far more likely they are intended to be printing characters from that position of Windows-1252 or Mac OS Roman.

ISO/IEC 6429 and RFC 1345 C1 control codes
ESC+
Decimal
Hex
Abbr Name Description[27]
@ 128 80 PAD[9] Padding Character[b] Proposed as a "padding" or "high byte" for single-byte characters to make them two bytes long for easier interoperability with multiple byte characters. Extended Unix Code (EUC) occasionally uses this.[28]
A 129 81 HOP[9] High Octet Preset[b] Proposed to set the high byte of a sequence of multiple byte characters so they only need one byte each, as a simple form of data compression.
B 130 82 BPH Break Permitted Here[c] Follows a graphic character where a line break is permitted. Roughly equivalent to a soft hyphen or zero-width space except it does not define what is printed at the line break.
C 131 83 NBH No Break Here[c] Follows the graphic character that is not to be broken. See also word joiner.
D 132 84 IND Index[d] Move down one line without moving horzontally, to eliminate ambiguity about the meaning of LF.
E 133 85 NEL Next Line Equivalent to CR+LF, to match the EBCDIC control character.
F 134 86 SSA Start of Selected Area Used by block-oriented terminals. In xterm ESC F moves to the lower-left corner of the screen, since certain software assumes this behaviour.[29]
G 135 87 ESA End of Selected Area
H 136 88 HTS
  • Character Tabulation Set
  • Horizontal Tabulation Set
Set a tab stop at the current position.
I 137 89 HTJ
  • Character Tabulation With Justification
  • Horizontal Tabulation With Justification
Right-justify the text since the last tab against the next tab stop.
J 138 8A VTS
  • Line Tabulation Set
  • Vertical Tabulation Set
Set a vertical tab stop.
K 139 8B PLD
  • Partial Line Forward
  • Partial Line Down
To produce subscripts and superscripts in ISO/IEC 6429.
Subscripts use PLD text PLU while superscripts use PLU text PLD.
L 140 8C PLU
  • Partial Line Backward
  • Partial Line Up
M 141 8D RI
  • Reverse Line Feed
  • Reverse Index
Move up one line.
N 142 8E SS2 Single-Shift 2 Next character is from the G2 or G3 sets, respectively.
O 143 8F SS3 Single-Shift 3
P 144 90 DCS Device Control String Followed by a string of printable characters (0x20 through 0x7E) and format effectors (0x08 through 0x0D), terminated by ST (0x9C). Xterm defined a number of these.[30]
Q 145 91 PU1 Private Use 1 Reserved for private function agreed on between the sender and the recipient of the data.
R 146 92 PU2 Private Use 2
S 147 93 STS Set Transmit State
T 148 94 CCH Cancel character Destructive backspace, to eliminate ambiguity about meaning of BS.
U 149 95 MW Message Waiting
V 150 96 SPA Start of Protected Area Used by block-oriented terminals.
W 151 97 EPA End of Protected Area
X 152 98 SOS Start of String[c] Followed by a control string terminated by ST (0x9C) which (unlike DCS, OSC, PM or APC) may contain any character except SOS or ST.
Y 153 99 SGC,[9] SGCI[31] Single Graphic Character Introducer[b] Intended to allow an arbitrary Unicode character to be printed; it would be followed by that character, most likely encoded in UTF-1.[31]
Z 154 9A SCI Single Character Introducer[c] To be followed by a single printable character (0x20 through 0x7E) or format effector (0x08 through 0x0D), and to print it as ASCII no matter what graphic or control sets were in use.
[ 155 9B CSI Control Sequence Introducer Used to introduce control sequences that take parameters. Used for ANSI escape sequences.
\ 156 9C ST String Terminator Terminates a string started by DCS, SOS, OSC, PM or APC.
] 157 9D OSC Operating System Command Followed by a string of printable characters (0x20 through 0x7E) and format effectors (0x08 through 0x0D), terminated by ST (0x9C), intended for use to allow in-band signaling of protocol information, but rarely used for that purpose.

Some terminal emulators, including xterm, use OSC sequences for setting the window title and changing the colour palette. They may also support terminating an OSC sequence with BEL instead of ST.[32] Kermit used APC to transmit commands.[33]

^ 158 9E PM Privacy Message
_ 159 9F APC Application Program Command
  1. ^ In early versions the range excluded SP and DEL[19]
  2. ^ a b c Not part of ISO/IEC 6429 (ECMA-48)[8][26]
  3. ^ a b c d Not part of the first edition of ISO/IEC 6429.[23]
  4. ^ Deprecated in 1988 and withdrawn in 1992 from ISO/IEC 6429 (1986 and 1991 respectively for ECMA-48).[citation needed]

Other control code sets[edit]

The ISO/IEC 2022 (ECMA-35) extension mechanism allowed escape sequences to change the C0 and C1 sets. The standard C0 control character set shown above is chosen with the sequence ESC ! @ and the above C1 set chosen with the sequence ESC " C.[23]

Several official and unofficial alternatives have been defined, but this is pretty much obsolete. Most were forced to retain a good deal of compatibility with the ASCII controls for interoperability. The standard makes ESC,[34][35] SP and DEL[a] "fixed" coded characters, which are available in their ASCII locations in all encodings that conform to the standard.[37] It also specifies that if a C0 set included transmission control (TCn) codes, they must be encoded at their ASCII locations[34] and could not be put in a C1 set,[38] and any new transmission controls must be in a C1 set.[34]

Other C0 control code sets[edit]

  • ANPA-1312, a text markup language used for news transmission, replaces several C0 control characters.
  • IPTC 7901, the newer international version of the above, has its own variations.
  • Videotex has a completely different set.
  • Teletext also defines a set similar to Videotex.
  • T.61/T.51,[39] and others[40] replaced EM and GS with SS2 and SS3 so these functions could be used in a 7-bit environment.
  • Some sets replaced FS with SS2,[41] (same as ANPA-1312).
  • The now-withdrawn JIS C 6225, designated JIS X 0207 in later sources.[42] replaced FS with CEX or "Control Extension"[43] which introduces control sequences for vertical text behaviour, superscripts and subscripts[44] and for transmitting custom character graphics.[42]

Replacement C1 character sets[edit]

  • A specialized C1 control code set is registered for bibliographic use (including string collation), such as by MARC-8.[22][45][46]
  • Various specialised C1 control code sets are registered for use by Videotex formats.[21]
  • EBCDIC defines up to 29 additional control codes besides those present in ASCII. When translating EBCDIC to Unicode (or to ISO 8859), these codes are mapped to C1 control characters in a manner specified by IBM's Character Data Representation Architecture (CDRA).[47][48] Although the New Line (NL) does translate to the ISO/IEC 6429 NEL (although it is often swapped with LF, following UNIX line ending convention),[47] the remainder of the control codes do not correspond. For example, the EBCDIC control SPS and the ECMA-48 control PLU are both used to begin a superscript or end a subscript, but are not mapped to one another. Extended-ASCII-mapped EBCDIC can therefore be regarded as having its own C1 set, although it is not registered with the ISO-IR registry for ISO/IEC 2022.[21]

Unicode[edit]

Unicode inherits its first 256 code points from ISO 8859-1, hence also the 65 code points described above, giving them the general category Cc (control). These are:

Unicode only specifies semantics for the C0 format controls HT, LF, VT, FF, and CR, (note BS is missing); the C0 information separators FS, GS, RS, US (and SP); and the C1 control NEL.[49] The rest of the codes are transparent to Unicode and their meanings are left to higher-level protocols, with ISO/IEC 6429 suggested as a default.[49]

Unicode includes many additional format effector characters besides these, such as marks, embeds, isolates and pops for explicit bidirectional formatting, and the zero-width joiner and non-joiner for controlling ligature use. However these are given the general category Cf (format) rather than Cc.

See also[edit]

Footnotes[edit]

  1. ^ ISO/IEC 4873 extends this requirement to the C1 SS2 and SS3,[36] although ISO/IEC 2022 itself does not.

References[edit]

  1. ^ Fox, Brian. "Adding a new node to Info". Info: The online, menu-driven GNU documentation system. GNU Project.
  2. ^ "Built-in Types § str.splitlines". The Python Standard Library. Python Software Foundation.
  3. ^ a b ISO/TC 97/SC 2 (1975). The set of control characters of the ISO 646 (PDF). ITSCJ/IPSJ. ISO-IR-1.{{citation}}: CS1 maint: numeric names: authors list (link)
  4. ^ a b c IPTC (1995). The IPTC Recommended Message Format (PDF) (5th ed.). IPTC TEC 7901.
  5. ^ a b c "end-of-transmission character (EOT)". Federal Standard 1037C. 1996. Archived from the original on 2016-03-09.
  6. ^ a b Robert McConnell; James Haynes; Richard Warren (December 2002). "Understanding ASCII Codes". NADCOMM.
  7. ^ Williamson, Karl. "Re: PRI #202: Extensions to NameAliases.txt for Unicode 6.1.0".
  8. ^ a b c Ken Whistler (July 20, 2011). "Formal Name Aliases for Control Characters, L2/11-281". Unicode Consortium.
  9. ^ a b c d "Name Aliases". Unicode Character Database. Unicode Consortium.
  10. ^ "C0 Controls and Basic Latin" (PDF). Unicode Consortium.
  11. ^ "charnames". Perl Programming Documentation.
  12. ^ ECMA (1994). "7.3: Invocation of character-set code elements". Character Code Structure and Extension Techniques (PDF) (ECMA Standard) (6th ed.). p. 14. ECMA-35.
  13. ^ American Standards Association (1963). American Standard Code for Information Interchange: 4. Legend. p. 6. ASA X3.4-1963.
  14. ^ "data link escape character (DLE)". Federal Standard 1037C. 1996. Archived from the original on 2016-08-01.
  15. ^ "Supplementary transmission control functions (an extension of the basic mode control procedures for data communication systems)". European Computer Manufacturers Association. 1972. ECMA-37.
  16. ^ "What is the point of Ctrl-S?". Unix and Linux Stack exchange. Retrieved 14 February 2019.
  17. ^ ECMA/TC 1 (1973). "Brief History". 7-bit Input/Output Coded Character Set (PDF) (4th ed.). ECMA. ECMA-6:1973.{{citation}}: CS1 maint: numeric names: authors list (link)
  18. ^ a b c ECMA/TC 1 (1971). "8.2: Correspondence between the 7-bit Code and an 8-bit Code". Extension of the 7-bit Coded Character Set (PDF) (1st ed.). ECMA. pp. 21–24. ECMA-35:1971.{{citation}}: CS1 maint: numeric names: authors list (link)
  19. ^ ECMA/TC 1 (1973). "4.2: Specific Control Characters". 7-bit Input/Output Coded Character Set (PDF) (4th ed.). ECMA. p. 16. ECMA-6:1973.{{citation}}: CS1 maint: numeric names: authors list (link)
  20. ^ ECMA/TC 1 (1985). "5.3.8: Sets of 96 graphic characters". Code Extension Techniques (PDF) (4th ed.). ECMA. pp. 17–18. ECMA-35:1985.{{citation}}: CS1 maint: numeric names: authors list (link)
  21. ^ a b c ISO/IEC International Register of Coded Character Sets To Be Used With Escape Sequences (PDF), ITSCJ/IPSJ, ISO-IR
  22. ^ a b DIN (1979-07-15). Additional Control Codes for Bibliographic Use according to German Standard DIN 31626 (PDF). ITSCJ/IPSJ. ISO-IR-40.
  23. ^ a b c ISO/TC97/SC2 (1983-10-01). C1 Control Set of ISO 6429:1983 (PDF). ITSCJ/IPSJ. ISO-IR-77.{{citation}}: CS1 maint: numeric names: authors list (link)
  24. ^ ECMA/TC 1 (1979). "Brief History". Additional Control Functions for Character-Imaging I/O Devices (PDF) (2nd ed.). ECMA. ECMA-48:1979.{{citation}}: CS1 maint: numeric names: authors list (link)
  25. ^ "JIS X 02xx 符号" (in Japanese).
  26. ^ a b Ken Whistler (2015-10-05). "Why Nothing Ever Goes Away". Unicode Mailing List.
  27. ^ ECMA (1991). Control Functions for Coded Character Sets. Standard ECMA-48.
  28. ^ Lunde, Ken (2008). CJKV Information Processing: Chinese, Japanese, Korean, and Vietnamese Computing. O'Reilly. p. 244. ISBN 9780596800925.
  29. ^ "VT100 Widget Resources (§ hpLowerleftBugCompat)". xterm - terminal emulator for X.
  30. ^ Moy, Edward; Gildea, Stephen; Dickey, Thomas. "Device-Control functions". XTerm Control Sequences.
  31. ^ a b Brender, Ronald F. (1989). "Ada 9x Project Report: Character Set Issues for Ada 9x". Carnegie Mellon University.
  32. ^ Moy, Edward; Gildea, Stephen; Dickey, Thomas. "Operating System Commands". XTerm Control Sequences.
  33. ^ Frank da Cruz; Christine Gianone (1997). Using C-Kermit. Digital Press. p. 278. ISBN 978-1-55558-164-0.
  34. ^ a b c ECMA (1994). "6.4.2: Primary sets of coded control functions". Character Code Structure and Extension Techniques (PDF) (ECMA Standard) (6th ed.). p. 11. ECMA-35.
  35. ^ ISO/TC97/SC2/WG-7; ECMA (1985-08-01). Minimum C0 set for ISO 4873 (PDF). ITSCJ/IPSJ. ISO-IR-104.{{citation}}: CS1 maint: numeric names: authors list (link)
  36. ^ ISO/TC97/SC2/WG-7; ECMA (1985-08-01). Minimum C1 Set for ISO 4873 (PDF). ITSCJ/IPSJ. ISO-IR-105.{{citation}}: CS1 maint: numeric names: authors list (link)
  37. ^ ECMA (1994). "6.2: Fixed coded characters". Character Code Structure and Extension Techniques (PDF) (ECMA Standard) (6th ed.). p. 7. ECMA-35.
  38. ^ ECMA (1994). "6.4.3: Supplementary sets of coded control functions". Character Code Structure and Extension Techniques (PDF) (ECMA Standard) (6th ed.). p. 11. ECMA-35.
  39. ^ ITU (1985). Teletex Primary Set of Control Functions (PDF). ITSCJ/IPSJ. ISO-IR-106.
  40. ^ Úřad pro normalizaci a měřeni (1987). The set of control characters of ISO 646, with EM replaced by SS2 (PDF). ITSCJ/IPSJ. ISO-IR-140.
  41. ^ ISO/TC 97/SC 2 (1977). The set of control characters of ISO 646, with IS4 replaced by Single Shift for G2 (SS2) (PDF). ITSCJ/IPSJ. ISO-IR-36.{{citation}}: CS1 maint: numeric names: authors list (link)
  42. ^ a b ISO/TC97/SC2/WG6. "Liaison statement to ISO/TC97/SC2/WG8 and ISO/TC97/SC18/WG8" (PDF). ISO/TC97/SC2/WG6 N317.rev. Archived from the original (PDF) on 2020-10-26.{{cite web}}: CS1 maint: numeric names: authors list (link)
  43. ^ ISO/TC 97/SC 2 (1982). The C0 set of Control Characters of Japanese Standard JIS C 6225-1979 (PDF). ITSCJ/IPSJ. ISO-IR-74.{{citation}}: CS1 maint: numeric names: authors list (link)
  44. ^ Printronix (2012). OKI® Programmer's Reference Manual (PDF). p. 26.
  45. ^ ISO/TC 46 (1983-06-01). Additional Control Codes for Bibliographic Use according to International Standard ISO 6630 (PDF). ITSCJ/IPSJ. ISO-IR-67.{{citation}}: CS1 maint: numeric names: authors list (link)
  46. ^ ISO/TC 46 (1986-02-01). Additional Control Codes for Bibliographic Use according to International Standard ISO 6630 (PDF). ITSCJ/IPSJ. ISO-IR-124.{{citation}}: CS1 maint: numeric names: authors list (link)
  47. ^ a b Umamaheswaran, V.S. (1999-11-08). "3.3 Step 2: Byte Conversion". UTF-EBCDIC. Unicode Consortium. Unicode Technical Report #16. The 64 control characters […], the ASCII DELETE character (U+007F)[…] are mapped respecting EBCDIC conventions, as defined in IBM Character Data Representation Architecture, CDRA, with one exception -- the pairing of EBCDIC Line Feed and New Line control characters are swapped from their CDRA default pairings to ISO/IEC 6429 Line Feed (U+000A) and Next Line (U+0085) control characters
  48. ^ Steele, Shawn (1996-04-24). cp037_IBMUSCanada to Unicode table. Microsoft/Unicode Consortium.
  49. ^ a b "23.1: Control Codes" (PDF). The Unicode Standard (12.0.0 ed.). Unicode Consortium. 2019. pp. 868–870. ISBN 978-1-936213-22-1.