[转]wiki:UTF-8

原文：http://en.wikipedia.org/wiki/UTF-8

 

From Wikipedia, the free encyclopedia

 

 

UTF-8 (U from Universal Character Set + Transformation Format—8-bit^[1]) is a character encoding capable of encoding all possible characters (called code points) in Unicode. The encoding is variable-length and uses 8-bit code units. It was designed for backward compatibility with ASCII and to avoid the complications of endianness and byte order marks in UTF-16 and UTF-32.

 

Graph indicating that UTF-8 (light blue) exceeded other main encodings of text on the Web in December 2007, and that by 2010 it was nearing 50% and up to 81% in 2014. Encodings were detected by examining the text, not from the encoding tag in the header,^[2] thus does not include ASCII tagged as UTF-8. As ASCII is valid UTF-8 it could be added to the UTF-8 to get over 65% usage.

UTF-8 has become the dominant character encoding for the World Wide Web, accounting for 81.4% of all Web pages in November 2014 (with most popular East Asian encoding at 1.4% and all of them combined under 5%).^[3][2][4] The Internet Mail Consortium (IMC) recommends that all e-mail programs be able to display and create mail using UTF-8.^[5] The W3C recommends UTF-8 as default encoding in their main standards (XML and HTML).

UTF-8 encodes each of the 1,112,064 valid code points in the Unicode code space (1,114,112 code points minus 2,048 surrogate code points) using one to four 8-bit bytes (a group of 8 bits is known as an octet in the Unicode Standard). Code points with lower numerical values (i.e. earlier code positions in the Unicode character set, which tend to occur more frequently) are encoded using fewer bytes. The first 128 characters of Unicode, which correspond one-to-one with ASCII, are encoded using a single octet with the same binary value as ASCII, making valid ASCII text valid UTF-8-encoded Unicode as well.

The official IANA code for the UTF-8 character encoding is UTF-8.^[6]

 

Contents

  [hide] 

• 1 History
• 2 Description
  • 2.1 Examples
  • 2.2 Codepage layout
  • 2.3 Overlong encodings
  • 2.4 Invalid byte sequences
  • 2.5 Invalid code points
  • 2.6 Sample code
• 3 Official name and variants
• 4 Derivatives
  • 4.1 CESU-8
  • 4.2 Modified UTF-8
• 5 Byte order mark
• 6 Advantages and disadvantages
  • 6.1 General
    • 6.1.1 Advantages
    • 6.1.2 Disadvantages
  • 6.2 Compared to single-byte encodings
    • 6.2.1 Advantages
    • 6.2.2 Disadvantages
  • 6.3 Compared to other multi-byte encodings
    • 6.3.1 Advantages
    • 6.3.2 Disadvantages
  • 6.4 Compared to UTF-16
    • 6.4.1 Advantages
    • 6.4.2 Disadvantages
• 7 See also
• 8 References
• 9 External links

 

History[edit]

By early 1992, the search was on for a good byte-stream encoding of multi-byte character sets. The draft ISO 10646 standard contained a non-required annex called UTF-1 that provided a byte-stream encoding of its 32-bit code points. This encoding was not satisfactory on performance grounds, but did introduce the notion that bytes in the range of 0–127 continue representing the ASCII characters in UTF, thereby providing backward compatibility with ASCII.

In July 1992, the X/Open committee XoJIG was looking for a better encoding. Dave Prosser of Unix System Laboratories submitted a proposal for one that had faster implementation characteristics and introduced the improvement that 7-bit ASCII characters would only represent themselves; all multibyte sequences would include only bytes where the high bit was set. This original proposal, the File System Safe UCS Transformation Format (FSS-UTF), was similar in concept to UTF-8, but lacked the crucial property of self-synchronization.^[7][8]

In August 1992, this proposal was circulated by an IBM X/Open representative to interested parties. Ken Thompson of the Plan 9 operating system group at Bell Labs made a small but crucial modification to the encoding, making it very slightly less bit-efficient than the previous proposal but allowing it to be self-synchronizing, meaning that it was no longer necessary to read from the beginning of the string to find code point boundaries. Thompson's design was outlined on September 2, 1992, on a placemat in a New Jersey diner with Rob Pike. In the following days, Pike and Thompson implemented it and updated Plan 9 to use it throughout, and then communicated their success back to X/Open.^[7]

UTF-8 was first officially presented at the USENIX conference in San Diego, from January 25 to 29, 1993.

Google reported that in 2008 UTF-8 (misleadingly labelled "Unicode") became the most common encoding for HTML files.^[9][10]

Description[edit]

The design of UTF-8 can be seen in this table of the scheme as originally proposed by Dave Prosser and subsequently modified by Ken Thompson (the x characters are replaced by the bits of the code point):

Bits of code point	First code point	Last code point	Bytes in sequence	Byte 1	Byte 2	Byte 3	Byte 4	Byte 5	Byte 6
7	U+0000	U+007F	1	0xxxxxxx
11	U+0080	U+07FF	2	110xxxxx	10xxxxxx
16	U+0800	U+FFFF	3	1110xxxx	10xxxxxx	10xxxxxx
21	U+10000	U+1FFFFF	4	11110xxx	10xxxxxx	10xxxxxx	10xxxxxx
The patterns below are not part of UTF-8, but were part of the first specification.
26	U+200000	U+3FFFFFF	5	111110xx	10xxxxxx	10xxxxxx	10xxxxxx	10xxxxxx
31	U+4000000	U+7FFFFFFF	6	1111110x	10xxxxxx	10xxxxxx	10xxxxxx	10xxxxxx	10xxxxxx

The original specification covered numbers up to 31 bits (the original limit of the Universal Character Set). In November 2003, UTF-8 was restricted by RFC 3629 to end at U+10FFFF, in order to match the constraints of the UTF-16character encoding. This removed all 5- and 6-byte sequences, and about half of the 4-byte sequences.

The salient features of this scheme are as follows:

• Backward compatibility: One-byte codes are used only for the ASCII values 0 through 127. In this case the UTF-8 code has the same value as the ASCII code. The high-order bit of these codes is always 0. This means that UTF-8 can be used for parsers expecting 8-bit extended ASCII even if they are not designed for UTF-8.
• Clear distinction between multi-byte and single-byte characters: Code points larger than 127 are represented by multi-byte sequences, composed of a leading byte and one or more continuation bytes. The leading byte has two or more high-order 1s followed by a 0, while continuation bytes all have '10' in the high-order position.
• Self synchronization: Single bytes, leading bytes, and continuation bytes do not share values. This makes the scheme self-synchronizing, allowing the start of a character to be found by backing up at most five bytes (three bytes in actual UTF‑8 per RFC 3629 restriction, see above).
  Bit patterns 0xxxxxxx and 11xxxxxx are synchronizing words used to mark the beginning of the next valid character.
• Clear indication of code sequence length: The number of high-order 1s in the leading byte of a multi-byte sequence indicates the number of bytes in the sequence, so that the length of the sequence can be determined without examining the continuation bytes.
• Code structure: The remaining bits of the encoding are used for the bits of the code point being encoded, padded with high-order 0s if necessary. The high-order bits go in the lead byte, lower-order bits in succeeding continuation bytes. The number of bytes in the encoding is the minimum required to hold all the significant bits of the code point.

The first 128 characters (US-ASCII) need one byte. The next 1,920 characters need two bytes to encode. This covers the remainder of almost all Latin alphabets, and also Greek, Cyrillic, Coptic, Armenian, Hebrew, Arabic, Syriac and Tānaalphabets, as well as Combining Diacritical Marks. Three bytes are needed for characters in the rest of the Basic Multilingual Plane (which contains virtually all characters in common use^[11]). Four bytes are needed for characters in theother planes of Unicode, which include less common CJK characters, various historic scripts, mathematical symbols, and emoji (pictographic symbols).

Examples[edit]

Consider the encoding of the Euro sign, €.

1. The Unicode code point for "€" is U+20AC.
2. According to the scheme table above, this will take three bytes to encode, since it is between U+0800 and U+FFFF.
3. Hexadecimal 20AC is binary 0010000010101100. The two leading zeros are added because, as the scheme table shows, a three-byte encoding needs exactly sixteen bits from the code point.
4. Because it is a three-byte encoding, the leading byte starts with three 1s, then a 0 (1110 ...)
5. The remaining bits of this byte are taken from the code point (11100010), leaving ... 000010101100.
6. Each of the continuation bytes starts with 10 and takes six bits of the code point (so 10000010, then 10101100).

The three bytes 11100010 10000010 10101100 can be more concisely written in hexadecimal, as E2 82 AC.

The following table summarises this conversion, as well as others with different lengths in UTF-8. The colors indicate how bits from the code point are distributed among the UTF-8 bytes. Additional bits added by the UTF-8 encoding process are shown in black.

Character		Binary code point	Binary UTF-8	Hexadecimal UTF-8
$	U+0024	0100100	00100100	24
¢	U+00A2	000 10100010	11000010 10100010	C2 A2
€	U+20AC	00100000 10101100	11100010 10000010 10101100	E2 82 AC
𤭢	U+24B62	00010 01001011 01100010	11110000 10100100 10101101 10100010	F0 A4 AD A2

Codepage layout[edit]

UTF-8

	_0	_1	_2	_3	_4	_5	_6	_7	_8	_9	_A	_B	_C	_D	_E	_F
0_	NUL 0000 0	SOH 0001 1	STX 0002 2	ETX 0003 3	EOT 0004 4	ENQ 0005 5	ACK 0006 6	BEL 0007 7	BS 0008 8	HT 0009 9	LF 000A 10	VT 000B 11	FF 000C 12	CR 000D 13	SO 000E 14	SI 000F 15
1_	DLE 0010 16	DC1 0011 17	DC2 0012 18	DC3 0013 19	DC4 0014 20	NAK 0015 21	SYN 0016 22	ETB 0017 23	CAN 0018 24	EM 0019 25	SUB 001A 26	ESC 001B 27	FS 001C 28	GS 001D 29	RS 001E 30	US 001F 31
2_	SP 0020 32	! 0021 33	" 0022 34	# 0023 35	$ 0024 36	% 0025 37	& 0026 38	' 0027 39	( 0028 40	) 0029 41	* 002A 42	+ 002B 43	, 002C 44	- 002D 45	. 002E 46	/ 002F 47
3_	0 0030 48	1 0031 49	2 0032 50	3 0033 51	4 0034 52	5 0035 53	6 0036 54	7 0037 55	8 0038 56	9 0039 57	: 003A 58	; 003B 59	< 003C 60	= 003D 61	> 003E 62	? 003F 63
4_	@ 0040 64	A 0041 65	B 0042 66	C 0043 67	D 0044 68	E 0045 69	F 0046 70	G 0047 71	H 0048 72	I 0049 73	J 004A 74	K 004B 75	L 004C 76	M 004D 77	N 004E 78	O 004F 79
5_	P 0050 80	Q 0051 81	R 0052 82	S 0053 83	T 0054 84	U 0055 85	V 0056 86	W 0057 87	X 0058 88	Y 0059 89	Z 005A 90	[ 005B 91	\ 005C 92	] 005D 93	^ 005E 94	_ 005F 95
6_	` 0060 96	a 0061 97	b 0062 98	c 0063 99	d 0064 100	e 0065 101	f 0066 102	g 0067 103	h 0068 104	i 0069 105	j 006A 106	k 006B 107	l 006C 108	m 006D 109	n 006E 110	o 006F 111
7_	p 0070 112	q 0071 113	r 0072 114	s 0073 115	t 0074 116	u 0075 117	v 0076 118	w 0077 119	x 0078 120	y 0079 121	z 007A 122	{ 007B 123	| 007C 124	} 007D 125	~ 007E 126	DEL 007F 127
8_	• +00 128	• +01 129	• +02 130	• +03 131	• +04 132	• +05 133	• +06 134	• +07 135	• +08 136	• +09 137	• +0A 138	• +0B 139	• +0C 140	• +0D 141	• +0E 142	• +0F 143
9_	• +10 144	• +11 145	• +12 146	• +13 147	• +14 148	• +15 149	• +16 150	• +17 151	• +18 152	• +19 153	• +1A 154	• +1B 155	• +1C 156	• +1D 157	• +1E 158	• +1F 159
A_	• +20 160	• +21 161	• +22 162	• +23 163	• +24 164	• +25 165	• +26 166	• +27 167	• +28 168	• +29 169	• +2A 170	• +2B 171	• +2C 172	• +2D 173	• +2E 174	• +2F 175
B_	• +30 176	• +31 177	• +32 178	• +33 179	• +34 180	• +35 181	• +36 182	• +37 183	• +38 184	• +39 185	• +3A 186	• +3B 187	• +3C 188	• +3D 189	• +3E 190	• +3F 191
2-byte C_	2-byte inval (0000) 192	2-byte inval (0040) 193	Latin-1 0080 194	Latin-1 00C0 195	Latin Ext-A 0100 196	Latin Ext-A 0140 197	Latin Ext-B 0180 198	Latin Ext-B 01C0 199	Latin Ext-B 0200 200	IPA 0240 201	IPA 0280 202	Spaci Modif 02C0 203	Combi Diacr 0300 204	Combi Diacr 0340 205	Greek 0380 206	Greek 03C0 207
2-byte D_	Cyril 0400 208	Cyril 0440 209	Cyril 0480 210	Cyril 04C0 211	Cyril 0500 212	Armen 0540 213	Hebrew 0580 214	Hebrew 05C0 215	Arabic 0600 216	Arabic 0640 217	Arabic 0680 218	Arabic 06C0 219	Syriac 0700 220	Arabic 0740 221	Thaana 0780 222	N'Ko 07C0 223
3-byte E_	Indic 0800* 224	Misc. 1000 225	Symbol 2000 226	Kana CJK 3000 227	CJK 4000 228	CJK 5000 229	CJK 6000 230	CJK 7000 231	CJK 8000 232	CJK 9000 233	Asian A000 234	Hangul B000 235	Hangul C000 236	Hangul Surr D000 237	Priv Use E000 238	Forms F000 239
4-byte F_	Ancient Sym,CJK 10000* 240	unall 40000 241	unall 80000 242	Tags Priv C0000 243	Priv Use 100000 244	4-byte inval 140000 245	4-byte inval 180000 246	4-byte inval 1C0000 247	5-byte inval 200000* 248	5-byte inval 1000000 249	5-byte inval 2000000 250	5-byte inval 3000000 251	6-byte inval 4000000* 252	6-byte inval 40000000 253	254	255

Legend: Yellow cells are control characters, blue cells are punctuation, purple cells are digits and green cells are ASCII letters.

Orange cells with a large dot are continuation bytes. The hexadecimal number shown after a "+" plus sign is the value of the 6 bits they add.

White cells are the start bytes for a sequence of multiple bytes, the length shown at the left edge of the row. The text shows the Unicode blocks encoded by sequences starting with this byte, and the hexadecimal code point shown in the cell is the lowest character value encoded using that start byte. When a start byte could form both overlong and valid encodings, the lowest non-overlong-encoded code point is shown, marked by an asterisk "*".

Red cells must never appear in a valid UTF-8 sequence. The first two (C0 and C1) could only be used for overlong encoding of basic ASCII characters (i.e., trying to encode a 7-bit ASCII value between 0 and 127 using 2 bytes instead of 1). The remaining red cells indicate start bytes of sequences that could only encode numbers larger than the 0x10FFFF limit of Unicode. The byte 244 (hex 0xF4) could also encode some values greater than 0x10FFFF; such a sequence would also be invalid if the subsequent bytes attempted to encode a value higher than 0x10FFFF.

Overlong encodings[edit]

In principle, it would be possible to inflate the number of bytes in an encoding by padding the code point with leading 0s. To encode the Euro sign € from the above example in four bytes instead of three, it could be padded with leading 0s until it was 21 bits long—000 000010 000010 101100, and encoded as 11110000 10000010 10000010 10101100 (or F0 82 82 AC in hexadecimal). This is called an overlong encoding.

The standard specifies that the correct encoding of a code point use only the minimum number of bytes required to hold the significant bits of the code point. Longer encodings are called overlong and are not valid UTF-8 representations of the code point. This rule maintains a one-to-one correspondence between code points and their valid encodings, so that there is a unique valid encoding for each code point, this makes string comparisons and searches well-defined.

Modified UTF-8 uses the 2-byte overlong encoding of U+0000 (the NUL character), 11000000 10000000 (hex C0 80), rather than 00000000 (hex 00). This allows the byte 00 to be used as a string terminator.

Invalid byte sequences[edit]

Not all sequences of bytes are valid UTF-8. A UTF-8 decoder should be prepared for:

• the red invalid bytes in the above table
• an unexpected continuation byte
• a start byte not followed by enough continuation bytes
• an Overlong Encoding as described above
• A 4-byte sequence (starting with 0xF4) that decodes to a value greater than U+10FFFF

Many earlier decoders would happily try to decode these. Carefully crafted invalid UTF-8 could make them either skip or create ASCII characters such as NUL, slash, or quotes. Invalid UTF-8 has been used to bypass security validations in high profile products including Microsoft's IIS web server^[12] and Apache's Tomcat servlet container.^[13]

RFC 3629 states "Implementations of the decoding algorithm MUST protect against decoding invalid sequences."^[14] The Unicode Standard requires decoders to "...treat any ill-formed code unit sequence as an error condition. This guarantees that it will neither interpret nor emit an ill-formed code unit sequence."

Many UTF-8 decoders throw exceptions on encountering errors.^[15] This can turn what would otherwise be harmless errors (producing a message such as "no such file") into a denial of service bug. Early versions of Python 3.0 would exit immediately if the command line or environment variables contained invalid UTF-8,^[16] making it impossible to handle such errors.

More recent converters translate the first byte of an invalid sequence to a replacement character and continue parsing with the next byte. These error bytes will always have the high bit set. This avoids denial-of-service bugs, and it is very common in text rendering such as browser display, since mangled text is probably more useful than nothing for helping the user figure out what the string was supposed to contain. Popular replacements include:

• The replacement character "�" (U+FFFD)
• The invalid Unicode code points U+DC80–U+DCFF where the low 8 bits are the byte's value.^[17] Sometimes it is called UTF-8B^[18] (where the B stands for Binary)
• The Unicode code points U+0080–U+00FF with the same value as the byte, thus interpreting the bytes according to ISO-8859-1^{[citation needed]}
• The Unicode code point for the character represented by the byte in CP1252,^{[citation needed]} which is similar to using ISO-8859-1, except that some characters in the range 0x80–0x9F are mapped into different Unicode code points. For example, 0x80 becomes the Euro sign, U+20AC.

These replacement algorithms are "lossy", as more than one sequence is translated to the same code point. This means that it would not be possible to reliably convert back to the original encoding, therefore losing information.

The large number of invalid byte sequences provides the advantage of making it easy to have a program accept both UTF-8 and legacy encodings such as ISO-8859-1. Thus, the software can check for UTF-8 correctness, and if that fails assume the input to be in the legacy encoding. It is technically true that this may detect an ISO-8859-1 string as UTF-8, but this is very unlikely if it contains any 8-bit bytes as they all have to be in unusual patterns of two or more in a row, such as "£".

Invalid code points[edit]

According to the UTF-8 definition (RFC 3629) the high and low surrogate halves used by UTF-16 (U+D800 through U+DFFF) are not legal Unicode values, and their UTF-8 encoding should be treated as an invalid byte sequence.

Whether an actual application should do this is debatable, as it makes it impossible to store invalid UTF-16 (that is, UTF-16 with unpaired surrogate halves) in a UTF-8 string. This is necessary to store unchecked UTF-16 such as Windows filenames as UTF-8. It is also incompatible with CESU encoding (described below).

Sample code[edit]

This code assumes ungetc can be called more than once. It translates any encoding errors into 0xDCxx, where xx is the value of the error byte.

void write_utf8(unsigned code_point)
{
  if (code_point < 0x80) {
    putchar(code_point);
  } else if (code_point <= 0x7FF) {
    putchar((code_point >> 6) + 0xC0);
    putchar((code_point & 0x3F) + 0x80);
  } else if (code_point <= 0xFFFF) {
    putchar((code_point >> 12) + 0xE0);
    putchar(((code_point >> 6) & 0x3F) + 0x80);
    putchar((code_point & 0x3F) + 0x80);
  } else if (code_point <= 0x10FFFF) {
    putchar((code_point >> 18) + 0xF0);
    putchar(((code_point >> 12) & 0x3F) + 0x80);
    putchar(((code_point >> 6) & 0x3F) + 0x80);
    putchar((code_point & 0x3F) + 0x80);
  } else {
    error("invalid code_point");
  }
}
 
unsigned read_code_point_from_utf8()
{
  int code_unit1, code_unit2, code_unit3, code_unit4;
 
  code_unit1 = getchar();
  if (code_unit1 < 0x80) {
    return code_unit1;
  } else if (code_unit1 < 0xC2) {
    /* continuation or overlong 2-byte sequence */
    goto ERROR1;
  } else if (code_unit1 < 0xE0) {
    /* 2-byte sequence */
    code_unit2 = getchar();
    if ((code_unit2 & 0xC0) != 0x80) goto ERROR2;
    return (code_unit1 << 6) + code_unit2 - 0x3080;
  } else if (code_unit1 < 0xF0) {
    /* 3-byte sequence */
    code_unit2 = getchar();
    if ((code_unit2 & 0xC0) != 0x80) goto ERROR2;
    if (code_unit1 == 0xE0 && code_unit2 < 0xA0) goto ERROR2; /* overlong */
    code_unit3 = getchar();
    if ((code_unit3 & 0xC0) != 0x80) goto ERROR3;
    return (code_unit1 << 12) + (code_unit2 << 6) + code_unit3 - 0xE2080;
  } else if (code_unit1 < 0xF5) {
    /* 4-byte sequence */
    code_unit2 = getchar();
    if ((code_unit2 & 0xC0) != 0x80) goto ERROR2;
    if (code_unit1 == 0xF0 && code_unit2 < 0x90) goto ERROR2; /* overlong */
    if (code_unit1 == 0xF4 && code_unit2 >= 0x90) goto ERROR2; /* > U+10FFFF */
    code_unit3 = getchar();
    if ((code_unit3 & 0xC0) != 0x80) goto ERROR3;
    code_unit4 = getchar();
    if ((code_unit4 & 0xC0) != 0x80) goto ERROR4;
    return (code_unit1 << 18) + (code_unit2 << 12) + (code_unit3 << 6) + code_unit4 - 0x3C82080;
  } else {
    /* > U+10FFFF */
    goto ERROR1;
  }
 
  ERROR4:
    ungetc(code_unit4, stdin);
  ERROR3:
    ungetc(code_unit3, stdin);
  ERROR2:
    ungetc(code_unit2, stdin);
  ERROR1:
    return code_unit1 + 0xDC00;
}

Official name and variants[edit]

The official name is "UTF-8". All letters are upper-case, and the name is hyphenated. This spelling is used in all the Unicode Consortium documents relating to the encoding.

Alternatively, the name "utf-8" may be used by all standards conforming to the Internet Assigned Numbers Authority (IANA) list (which include CSS, HTML, XML, and HTTP headers),^[19] as the declaration is case insensitive.^[20]

Other descriptions that omit the hyphen or replace it with a space, such as "utf8" or "UTF 8", are not accepted as correct by the governing standards.^[14] Despite this, most agents such as browsers can understand them, and so standards intended to describe existing practice (such as HTML5) may effectively require their recognition.

Unofficially, UTF-8-BOM or UTF-8-NOBOM are sometimes used to refer to text files which contain or lack a byte order mark (BOM). In Japan especially, UTF-8 encoding without BOM is sometimes called "UTF-8N".^[21][22]

Derivatives[edit]

The following implementations show slight differences from the UTF-8 specification. They are incompatible with the UTF-8 specification.

CESU-8[edit]

Main article: CESU-8

Many programs added UTF-8 conversions for UCS-2 data and did not alter this UTF-8 conversion when UCS-2 was replaced with the surrogate-pair using UTF-16. In such programs each half of a UTF-16 surrogate pair is encoded as its own 3-byte UTF-8 encoding, resulting in 6-byte sequences rather than 4 bytes for characters outside the Basic Multilingual Plane. Oracle and MySQL databases use this, as well as Java and Tcl as described below, and probably many Windows programs where the programmers were unaware of the complexities of UTF-16. Although this non-optimal encoding is generally not deliberate, a supposed benefit is that it preserves UTF-16 binary sorting order when CESU-8 is binary sorted.

Modified UTF-8[edit]

In Modified UTF-8,^[23] the null character (U+0000) is encoded as 0xC0,0x80; this is not valid UTF-8^[24] because it is not the shortest possible representation. Modified UTF-8 strings never contain any actual null bytes but can contain all Unicode code points including U+0000,^[25] which allows such strings (with a null byte appended) to be processed by traditional null-terminated string functions.

All known Modified UTF-8 implementations also treat the surrogate pairs as in CESU-8.

In normal usage, the Java programming language supports standard UTF-8 when reading and writing strings through InputStreamReader and OutputStreamWriter. However it uses Modified UTF-8 for object serialization,^[26] for the Java Native Interface,^[27] and for embedding constant strings in class files.^[28] The dex format defined by Dalvik also uses the same modified UTF-8 to represent string values.^[29] Tcl also uses the same modified UTF-8^[30] as Java for internal representation of Unicode data, but uses strict CESU-8 for external data.

Byte order mark[edit]

Many Windows programs (including Windows Notepad) add the bytes 0xEF, 0xBB, 0xBF at the start of any document saved as UTF-8. This is the UTF-8 encoding of the Unicode byte order mark (BOM), and is commonly referred to as a UTF-8 BOM, even though it is not relevant to byte order. A BOM can also appear if another encoding with a BOM is translated to UTF-8 without stripping it. Software that is not aware of multibyte encodings will display the BOM as three strange characters (e.g. "" in software interpreting the document as ISO 8859-1 or Windows-1252) at the start of the document.

The Unicode Standard neither requires nor recommends the use of the BOM for UTF-8.^[31] The presence of the UTF-8 BOM may cause interoperability problems with existing software that could otherwise handle UTF-8; for example:

• Programming language parsers not explicitly designed for UTF-8 can often handle UTF-8 in string constants and comments, but cannot parse the BOM at the start of the file.
• Programs that identify file types by leading characters may fail to identify the file if a BOM is present even if the user of the file could skip the BOM. An example is the Unix shebang syntax. Another example is Internet Explorer which will render pages in standards mode only when it starts with a document type declaration.
• Programs that insert information at the start of a file will break use of the BOM to identify UTF-8 (one example is offline browsers that add the originating URL to the start of the file).

If compatibility with existing programs is not important, the BOM could be used to identify UTF-8 encoding, but such use should not be necessary as UTF-8 can be identified with very high reliability since other encodings are extremely unlikely to contain valid UTF-8 byte sequences.

Advantages and disadvantages[edit]

This section needs additional citations for verification. Please help improve this article by adding citations to reliable sources. Unsourced material may be challenged and removed. (October 2009)

General[edit]

Advantages[edit]

• UTF-8 is the only encoding for XML entities that does not require a BOM or an indication of the encoding.^[32]
• UTF-8 and UTF-16 are the standard encodings for Unicode text in HTML documents, with UTF-8 as the preferred and most used encoding.
• UTF-8 strings can be fairly reliably recognized as such by a simple heuristic algorithm.^[33] The probability of a random string of bytes which is not pure ASCII being valid UTF-8 is 3.9% for a two-byte sequence,^[34] and decreases exponentially for longer sequences. ISO/IEC 8859-1 is even less likely to be mis-recognized as UTF-8: the only non-ASCII characters in it would have to be in sequences starting with either an accented letter or the multiplication symbol and ending with a symbol. This is an advantage that most other encodings do not have, and allows UTF-8 to be mixed with a legacy encoding without having to add data to identify which encoding is in use, avoiding errors (mojibake) typically encountered when trying to change a system to a new default encoding. Even word-based UTF-16 can be mistaken for byte encodings (like in the "bush hid the facts" bug).
• Sorting a set of UTF-8 encoded strings as strings of unsigned bytes yields the same order as sorting the corresponding Unicode strings lexicographically by codepoint.

Disadvantages[edit]

• A UTF-8 parser that is not compliant with current versions of the standard might accept a number of different pseudo-UTF-8 representations and convert them to the same Unicode output. This provides a way for information to leak past validation routines designed to process data in its eight-bit representation.^[14]

Compared to single-byte encodings[edit]

Advantages[edit]

• UTF-8 can encode any Unicode character, avoiding the need to figure out and set a "code page" or otherwise indicate what character set is in use, and allowing output in multiple scripts at the same time. For many scripts there have been more than one single-byte encoding in usage, so even knowing the script was insufficient information to display it correctly.
• The bytes 0xFE and 0xFF do not appear, so a valid UTF-8 stream never matches the UTF-16 byte order mark and thus cannot be confused with it. The absence of 0xFF (0377) also eliminates the need to escape this byte in Telnet (and FTP control connection).

Disadvantages[edit]

• UTF-8 encoded text is larger than specialized single-byte encodings except for plain ASCII characters. In the case of scripts which used 8-bit character sets with non-Latin characters encoded in the upper half (such as most Cyrillic andGreek alphabet code pages), characters in UTF-8 will be double the size. For some scripts, such as Thai and Hindi's Devanagari, characters will triple in size. This has caused objections in India and other countries.^{[citation needed]}
• It is possible in UTF-8 (or any other multi-byte encoding) to split or truncate a string in the middle of a character. This can result in an invalid string if the two halves are not concatenated later.
• If the code points are all the same size, measurements of a fixed number of them is easy. Due to ASCII-era documentation where "character" is used as a synonym for "byte" this is often considered important. However, by measuring string positions using bytes instead of "characters" most algorithms can be easily and efficiently adapted for UTF-8.^{[citation needed]}
• Some software, such as text editors, will refuse to correctly display or interpret UTF-8 unless the text starts with a Byte Order Mark, and will insert such a mark. This has the effect of making it impossible to use UTF-8 with any older software that can handle ASCII-like encodings but cannot handle the byte order mark. This is considered an incorrect implementation of the text editor, not the older software.

Compared to other multi-byte encodings[edit]

Advantages[edit]

• UTF-8 uses the codes 0–127 only for the ASCII characters. This means that UTF-8 is an ASCII extension and can be processed by software that supports 7-bit characters and assigns no meaning to non-ASCII bytes. By contrast, in Shift-JISa byte that can be a 7-bit ASCII character can also be used as part of a multi-byte character. The byte 0x5C, for example, might be part of a multibyte character, but in the context of a string some programming languages or application software would instead interpret it as a backslash ('\') and assume that it marks the beginning of an escape sequence, incorrectly influencing the interpretation of subsequent bytes.^[35]
• UTF-8 can encode any Unicode character. Files in different scripts can be displayed correctly without having to choose the correct code page or font. For instance Chinese and Arabic can be supported (in the same text) without special codes inserted or manual settings to switch the encoding.
• UTF-8 is self-synchronizing: character boundaries are easily identified by scanning for well-defined bit patterns in either direction. If bytes are lost due to error or corruption, one can always locate the beginning of the next valid character and resume processing. Many multi-byte encodings are much harder to resynchronize.
• Any byte oriented string searching algorithm can be used with UTF-8 data, since the sequence of bytes for a character cannot occur anywhere else. Some older variable-length encodings (such as Shift JIS) did not have this property and thus made string-matching algorithms rather complicated. In Shift JIS the end byte of a character and the first byte of the next character could look like another legal character, something that can't happen in UTF-8.
• Efficient to encode using simple bit operations. UTF-8 does not require slower mathematical operations such as multiplication or division (unlike the obsolete UTF-1 encoding).

Disadvantages[edit]

• UTF-8 will take more space than a multi-byte encoding designed for a specific script. East Asian legacy encodings generally used two bytes per character yet take three bytes per character in UTF-8.

Compared to UTF-16[edit]

Advantages[edit]

• Byte encodings and UTF-8 are represented by byte arrays in programs, and often nothing needs to be done to a function when converting from a byte encoding to UTF-8. UTF-16 is represented by 16-bit word arrays, and converting to UTF-16 while maintaining compatibility with existing programs (such as was done with Windows) requires every API and data structure that takes a string to be duplicated, one version accepting byte strings and another version accepting UTF-16. Different handling of invalid strings often makes the "duplicated" APIs not exactly map to each other, making it impossible to do some actions with one of them.
• Characters outside the basic multilingual plane are not a special case. UTF-16 is often mistaken to be the obsolete constant-length UCS-2 encoding, leading to code that works for most text but suddenly fails for non-BMP characters.
• Text encoded in UTF-8 will be smaller than the same text encoded in UTF-16 if there are more code points below U+0080 than in the range U+0800..U+FFFF. This is true of all modern European languages. As HTML markup characters, numbers (digits 0–9), spaces and line terminators are all code points below U+0080, this is often true even for Asian scripts.
• Most communication and storage was designed for a stream of bytes. A UTF-16 string must use a pair of bytes for each code unit:
  • The order of those two bytes becomes an issue and must be specified in the UTF-16 protocol, such as with a byte order mark.
  • If an odd number of bytes is missing from UTF-16, the whole rest of the string will be meaningless text. Any bytes missing from UTF-8 will still allow the text to be recovered accurately starting with the next character after the missing bytes. If any partial character is removed the corruption is always recognizable.
• A "UTF-8" byte stream (one that is allowed to contain invalid byte sequences) can store all possible arrangements of bytes, while also allowing the lossless translation of all possible "UTF-16" streams (streams that can contain any sequence of 16-bit words including invalid UTF-16 sequences), by encoding each unpaired UTF-16 surrogate half in the obvious way. This makes UTF-8 a portable solution if it is necessary to preserve invalid sequences. An example are both Unix "UTF-8" filenames and Windows "UTF-16" filenames (both systems do not prevent invalid file names).

Disadvantages[edit]

• Characters U+0800 through U+FFFF use three bytes in UTF-8, but only two in UTF-16. As a result, text in (for example) Chinese, Japanese or Hindi could take more space in UTF-8 if there are more of these characters than there are ASCII characters. This happens for pure text^[36] but rarely for HTML documents or documents in XML based formats such as .docx or .odt. For example, both the Japanese UTF-8 and the Hindi Unicode articles on Wikipedia take more space in UTF-16 than in UTF-8.^[37]

See also[edit]

• Alt code
• Character encodings in HTML
• Comparison of e-mail clients § Features
• Comparison of Unicode encodings
• GB 18030
• Iconv—a standardized API used to convert between different character encodings
• ISO/IEC 8859
• Specials (Unicode block)
• Unicode and e-mail
• Unicode and HTML
• Universal Character Set
• UTF-8 in URIs
• UTF-9 and UTF-18
• UTF-16/UCS-2

References[edit]

1. Jump up^ "Chapter 2. General Structure". The Unicode Standard(6.0 ed.). Mountain View, California, USA: The Unicode Consortium. ISBN 978-1-936213-01-6.
2. ^ Jump up to:^a b Davis, Mark (28 January 2010). "Unicode nearing 50% of the web". Official Google Blog. Google. Retrieved 5 December 2010.
3. Jump up^ "Usage Statistics of Character Encodings for Websites, November 2014". W3Techs. 5 November 2014. Retrieved 5 November 2014.
4. Jump up^ "UTF-8 Usage Statistics". BuiltWith. Retrieved March 28, 2011.
5. Jump up^ "Using International Characters in Internet Mail". Internet Mail Consortium. August 1, 1998. Retrieved November 8, 2007.
6. Jump up^ "CHARACTER SETS". Internet Assigned Numbers Authority. November 4, 2010. Retrieved 5 December 2010.
7. ^ Jump up to:^a b Pike, Rob (30 Apr 2003). "UTF-8 history". Retrieved September 7, 2012.
8. Jump up^ Pike, Rob (September 6, 2012). "UTF-8 turned 20 years old yesterday". Retrieved September 7, 2012.
9. Jump up^ Davis, Mark (5 May 2008). "Moving to Unicode 5.1". Retrieved 2013-03-01.
10. Jump up^ Goodger, David (6 May 2008). "Unicode misinformation". Retrieved 2013-03-01.
11. Jump up^ Allen, Julie D.; Anderson, Deborah; Becker, Joe; Cook, Richard, eds. (2012). "The Unicode Standard, Version 6.1". Mountain View, California: Unicode Consortium. "The Basic Multilingual Plane (BMP, or Plane 0) contains the common-use characters for all the modern scripts of the world as well as many historical and rare characters. By far the majority of all Unicode characters for almost all textual data can be found in the BMP."
12. Jump up^ Marin, Marvin (October 17, 2000). "Web Server Folder Traversal MS00-078".
13. Jump up^ "National Vulnerability Database - Summary for CVE-2008-2938".
14. ^ Jump up to:^a b c Yergeau, F. (2003). "UTF-8, a transformation format of ISO 10646". "RFC 3629". Internet Engineering Task Force.
15. Jump up^ decode() method of Java UTF8 object
16. Jump up^ "Non-decodable Bytes in System Character Interfaces".python.org. 2009-04-22. Retrieved 2014-08-13.
17. Jump up^ Kuhn, Markus (2000-07-23). "Substituting malformed UTF-8 sequences in a decoder". Retrieved 2014-09-25.
18. Jump up^ Sittler, B. (2006-04-02). "Binary vs. UTF-8, and why it need not matter". Retrieved 2014-09-25.
19. Jump up^ Dürst, Martin. "Setting the HTTP charset parameter".W3C. Retrieved February 8, 2013.
20. Jump up^ "Character Sets". Internet Assigned Numbers Authority. January 23, 2013. Retrieved February 8, 2013.
21. Jump up^ "BOM - suikawiki" (in Japanese). Retrieved 2013-04-26.
22. Jump up^ Mark Davis. "Forms of Unicode". IBM. Archived from the original on 6 May 2005. Retrieved 18 September 2013.
23. Jump up^ "Java SE 6 documentation for Interface java.io.DataInput, subsection on Modified UTF-8". Sun Microsystems. 2008. Retrieved May 22, 2009.
24. Jump up^ "Request for Comments 3629: "UTF-8, a transformation format of ISO 10646"". 2003. Retrieved May 22, 2009. "[...] the overlong UTF-8 sequence C0 80 [...]", "[...] the illegal two-octet sequence C0 80 [...]"
25. Jump up^ "The Java Virtual Machine Specification, 2nd Edition, section 4.4.7: "The CONSTANT_Utf8_info Structure"". Sun Microsystems. 1999. Retrieved May 24, 2009. "[...] Java virtual machine UTF-8 strings never have embedded nulls."
26. Jump up^ "Java Object Serialization Specification, chapter 6: Object Serialization Stream Protocol, section 2: Stream Elements". Sun Microsystems. 2005. Retrieved May 22, 2009. "[...] encoded in modified UTF-8."
27. Jump up^ "Java Native Interface Specification, chapter 3: JNI Types and Data Structures, section: Modified UTF-8 Strings". Sun Microsystems. 2003. Retrieved May 22, 2009. "The JNI uses modified UTF-8 strings to represent various string types."
28. Jump up^ "The Java Virtual Machine Specification, 2nd Edition, section 4.4.7: "The CONSTANT_Utf8_info Structure"". Sun Microsystems. 1999. Retrieved May 23, 2009. "[...] differences between this format and the "standard" UTF-8 format."
29. Jump up^ "dex — Dalvik Executable Format". Retrieved April 9, 2013. "[T]he dex format encodes its string data in a de facto standard modified UTF-8 form, hereafter referred to as MUTF-8."
30. Jump up^ "Tcler's Wiki: UTF-8 bit by bit (Revision 6)". April 25, 2009. Retrieved May 22, 2009. "In orthodox UTF-8, a NUL byte(\x00) is represented by a NUL byte. [...] But [...] we [...] want NUL bytes inside [...] strings [...]"
31. Jump up^ "The Unicode Standard - Chapter 2". p. 30.
32. Jump up^ "Extensible Markup Language (XML) 1.0 (Fifth Edition)". W3C. November 26, 2008. Retrieved February 8, 2013.
33. Jump up^ Dürst, Martin. "Multilingual Forms". W3C. Retrieved February 8, 2013.
34. Jump up^ There are 256 × 256 − 128 × 128 = 49152 2-byte sequences with at least one high bit set, only 1920 encode valid UTF-8 characters (the range U+0080 to U+07FF), 1920 ÷ 49152 = 3.9%
35. Jump up^ "#418058 - iconv: half-smart on ascii compatible code conversion (shift-jis) - Debian Bug report logs". Bugs.debian.org. 2007-04-06. Retrieved 2014-06-13.
36. Jump up^ Although the difference may not be great: the 2010-11-22 version of hi:यूनिकोड (Unicode in Hindi), when the pure text was pasted to Notepad, generated 19 KB when saved as UTF-16 and 22 KB when saved as UTF-8.
37. Jump up^ The 2010-10-27 version of ja:UTF-8 generated 169 KB when converted with Notepad to UTF-16, and only 101 KB when converted back to UTF-8. The 2010-11-22 version ofhi:यूनिकोड (Unicode in Hindi) required 119 KB in UTF-16 and 76 KB in UTF-8.

External links[edit]

Look up UTF-8 in Wiktionary, the free dictionary.

There are several current definitions of UTF-8 in various standards documents:

• RFC 3629 / STD 63 (2003), which establishes UTF-8 as a standard Internet protocol element
• The Unicode Standard, Version 6.0, §3.9 D92, §3.10 D95 (2011)
• ISO/IEC 10646:2012 §9.1

They supersede the definitions given in the following obsolete works:

• ISO/IEC 10646-1:1993 Amendment 2 / Annex R (1996)
• The Unicode Standard, Version 5.0, §3.9 D92, §3.10 D95 (2007)
• The Unicode Standard, Version 4.0, §3.9–§3.10 (2003)
• The Unicode Standard, Version 2.0, Appendix A (1996)
• RFC 2044 (1996)
• RFC 2279 (1998)
• The Unicode Standard, Version 3.0, §2.3 (2000) plus Corrigendum #1 : UTF-8 Shortest Form (2000)
• Unicode Standard Annex #27: Unicode 3.1 (2001)

They are all the same in their general mechanics, with the main differences being on issues such as allowed range of code point values and safe handling of invalid input.

• Original UTF-8 paper (or pdf) for Plan 9 from Bell Labs
• RFC 5198 defines UTF-8 NFC for Network Interchange
• UTF-8 test pages by Andreas Prilop, Jost Gippert and the World Wide Web Consortium
• Unix/Linux: UTF-8/Unicode FAQ, Linux Unicode HOWTO, UTF-8 and Gentoo
• The Unicode/UTF-8-character table displays UTF-8 in a variety of formats (with Unicode and HTML encoding information)
• Characters, Symbols and the Unicode Miracle – Computerphile on YouTube