UTF-8

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 0010 0000 1010 1100. 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 the encoding will be three bytes long, its leading byte starts with three 1s, then a 0 (1110...)
5. The four most significant bits of the code point are stored in the remaining low order four bits of this byte (1110 0010), leaving 12 bits of the code point yet to be encoded (...0000 1010 1100).
6. All continuation bytes contain exactly six bits from the code point. So the next six bits of the code point are stored in the low order six bits of the next byte, and 10 is stored in the high order two bits to mark it as a continuation byte (so 1000 0010).
7. Finally the last six bits of the code point are stored in the low order six bits of the final byte, and again 10 is stored in the high order two bits (1010 1100).

The three bytes 1110 0010 1000 0010 1010 1100 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.

Since UTF-8 uses groups of six bits, it is sometimes useful to use octal notation which uses 3-bit groups. With a calculator which can convert between hexadecimal and octal it can be easier to manually create or interpret UTF-8 compared with using binary.

• Octal 0–177 (hex 0–7F) is coded with an unchanged single byte.
• Octal 0200–3777 (hex 80–7FF) shall be coded with two bytes. xxyy will be 3xx 2yy.
• Octal 4000–77777 (hex 800–7FFF) shall be coded with three bytes. xyyzz will be 34x 2yy 2zz.
• Octal 100000–177777 (hex 8000–FFFF) shall be coded with three bytes. 1xyyzz will be 35x 2yy 2zz.
• Octal 200000–4177777 (hex 10000–10FFFF) shall be coded with four bytes. wxxyyzz will be 36w 2xx 2yy 2zz.

When converting UTF-8 into code points, the following rules apply:

• Octal 302–337 is the first of two bytes. 3xx 2yy will be xxyy in octal.
• Octal 340–347 is the first of three bytes. 34x 2yy 2zz will be xyyzz.
• Octal 350–357 is the first of three bytes. 35x 2yy 2zz will be 1xyyzz.
• Octal 360–364 is the first of four bytes. 36w 2xx 2yy 2zz will be wxxyyzz.
• Octal 200–277 are continuation bytes, and others beginning with 3 are invalid.

The following table summarizes usage of UTF-8 code units (individual bytes or octets) in a code page format. The upper half (0_ to 7_) is for bytes used only in single-byte codes, so it looks like a normal code page; the lower half is for continuation bytes (8_ to B_) and leading bytes (C_ to F_), and is explained further in the legend below.

     Blue cells are 7-bit (single-byte) sequences. They must not be followed by a continuation byte.

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

     White cells are the leading 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 leading byte.

     Red cells must never appear in a valid UTF-8 sequence. The first two red cells (C0 and C1) could be used only for a 2-byte encoding of a 7-bit ASCII character which should be encoded in 1 byte; as described below, such "overlong" sequences are disallowed. The red cells in the F_ row (F5 to FD) indicate leading bytes of 4-byte or longer sequences that cannot be valid because they would encode code points larger than the U+10FFFF limit of Unicode (a limit derived from the maximum code point encodable in UTF-16), and FE and FF were never defined for any purpose in UTF-8.

     Pink cells are the leading bytes for a sequence of multiple bytes, of which some, but not all, possible continuation sequences are valid. E0 and F0 could start overlong encodings, in this case the lowest non-overlong-encoded code point is shown. F4 can start code points greater than U+10FFFF which are invalid. ED can start the encoding of a code point in the range U+D800–U+DFFF; these are invalid since they are reserved for UTF-16 surrogate halves.

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 ensures that string comparisons and searches are well-defined.

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

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 non-continuation byte before the end of the character
• the string ending before the end of the character (which can happen in simple string truncation)
• an overlong encoding as described above
• a sequence that decodes to an invalid code point as described below

Many of the first UTF-8 decoders would decode these, ignoring incorrect bits and accepting overlong results. 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[14] and Apache's Tomcat servlet container.[15] RFC 3629 states "Implementations of the decoding algorithm MUST protect against decoding invalid sequences."[16] 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."

Some implementations of decoders throw exceptions on errors.[17] This has the disadvantage that it can turn what would otherwise be harmless errors (such as a "no such file" error) into a denial of service. For instance early versions of Python 3.0 would exit immediately if the command line or environment variables contained invalid UTF-8.[18]

An alternative practice is to replace errors with a replacement character. Replacement requires defining how many bytes are in the error:

• Using the length defined by the lead byte (so that E1,A0,C0 is one error) is a bad idea, as a missing byte will cause the next character to become an error as well.
• Since Unicode 6[19] (October 2010), the standard (chapter 3) has “recommended” a "best practice" where the error ends as soon as a disallowed byte is encountered. In these decoders E1,A0,C0 is two errors (2 bytes in the first one). This means an error is no more than three bytes long and never contains the start of a valid character, and there are 21,952 different possible errors.[20] The standard also recommends replacing each error with the replacement character "�" (U+FFFD).
• Another popular practice is to turn each byte into an error. In this case E1,A0,C0 is three errors. This requires the UTF-8 parser to back up 2 bytes when an error is detected. The primary advantage is that there are now only 128 different errors. This allows the decoder to define 128 different error replacements such as:
  • The invalid Unicode code points U+DC80–U+DCFF where the low eight bits are the byte's value.[21] Sometimes it is called UTF-8B.[22] This has the nice property that it makes invalid UTF-8 into invalid UTF-16.
  • The Unicode code points U+0080–U+00FF with the same value as the byte, thus interpreting the bytes according to ISO-8859-1 Care must be taken so that the C1 control codes such as NEL 0x0085 do not cause further code to misbehave.
  • The Unicode code point for the character represented by the byte in CP1252, which is similar to using ISO-8859-1, except most bytes in the range 0x80–0x9F turn into printing characters. For example, 0x80 becomes the Euro sign, U+20AC. This makes text where legacy encodings are mixed with UTF-8 readable, and thus it is commonly done in browsers.

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. 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 "£".

Since RFC 3629 (November 2003), the high and low surrogate halves used by UTF-16 (U+D800 through U+DFFF) and code points not encodable by UTF-16 (those after U+10FFFF) are not legal Unicode values, and their UTF-8 encoding must be treated as an invalid byte sequence.

Not decoding unpaired surrogate halves makes it impossible to store invalid UTF-16 (such as Windows filenames or UTF-16 that has been split between the surrogates) as UTF-8. To preserve these invalid UTF-16 sequences, their corresponding UTF-8 encodings are sometimes allowed by implementations despite the above rule. There are attempts to define this behavior formally (see WTF-8 and CESU below).

Many Windows programs (including older versions of 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 byte order is irrelevant to UTF-8. While ASCII text encoded using UTF-8 normally is backwards compatible with ASCII, this is not true when Unicode Standard recommendations are ignored and a BOM is added. Non-UTF-8 software may show the BOM as three garbage characters, e.g. "" in software interpreting the document as ISO 8859-1 or Windows-1252, and "" if interpreted as code page 437. This is an example of mojibake, the output of garbled text when text is decoded using an unintended character encoding.

The Unicode Standard neither requires nor recommends the use of the BOM for UTF-8, but warns that it may be encountered at the start of a file transcoded from another encoding.[23] The presence of the UTF-8 BOM may cause problems with existing software that can 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 UTF-8 BOM at the start of the file.
• Programs that identify file types by leading characters may fail to identify the file if a UTF-8 BOM is present even if the user of the file can handle the BOM. An example is the Unix shebang syntax, another 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).

Nonetheless, some software such as text editors will refuse to correctly display or interpret UTF-8 unless the text starts with a byte order mark or it contains only ASCII.

The official Internet Assigned Numbers Authority (IANA) code for the encoding is "UTF-8".[24] 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 IANA list (which include CSS, HTML, XML, and HTTP headers),[25] as the declaration is case insensitive.[24]

Other descriptions, such as those that omit the hyphen or replace it with a space, i.e. "utf8" or "UTF 8", are not accepted as correct by the governing standards.[16] 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.[26]

Unofficially, UTF-8-BOM and UTF-8-NOBOM are sometimes used to refer to text files which respectively contain and lack a byte order mark (BOM). In Japan especially, UTF-8 encoding without BOM is sometimes called "UTF-8N".[27][28]

Supported Windows versions, i.e. Windows 7 and later, have codepage 65001, as a synonym for UTF-8 (with better support than in older Windows),[29] and Microsoft has a script for Windows 10, to enable it by default for its program Microsoft Notepad.[30]

In PCL, UTF-8 is called Symbol-ID "18N" (PCL supports 183 character encodings, called Symbol Sets, which potentially could be reduced to one, 18N, that is UTF-8).[31]

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
• RFC 5198 defines UTF-8 NFC for Network Interchange (2008)
• ISO/IEC 10646:2014 §9.1 (2014)[46]
• The Unicode Standard, Version 11.0 (2018)[47]

They supersede the definitions given in the following obsolete works:

• The Unicode Standard, Version 2.0, Appendix A (1996)
• ISO/IEC 10646-1:1993 Amendment 2 / Annex R (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)[48]
• The Unicode Standard, Version 5.0 (2006)[49]
• The Unicode Standard, Version 6.0 (2010)[50]

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.

• 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).
• 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 and Greek alphabet code pages), characters in UTF-8 will be double the size. For some scripts, such as Thai and Devanagari (which is used by various South Asian languages), characters will triple in size. There are even examples where a single byte turns into a composite character in Unicode and is thus six times larger in UTF-8. This has caused objections in India and other countries.
• It is possible in UTF-8 (or any other variable-length encoding) to split or truncate a string in the middle of a character. If the two pieces are not re-appended later before interpretation as characters, this can introduce an invalid sequence at both the end of the previous section and the start of the next, and some decoders will not preserve these bytes and result in data loss. Because UTF-8 is self-synchronizing this will however never introduce a different valid character, and it is also fairly easy to move the truncation point backwards to the start of a character.
• 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. Searching for a string within a long string can for example be done byte by byte; the self-synchronization property prevents false positives.

• 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 written in the same file without specialised markup or manual settings that specify an 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 next valid character and resume processing. If there is a need to shorten a string to fit a specified field, the previous valid character can easily be found. Many multi-byte encodings such as Shift JIS are much harder to resynchronize. This also means that byte-oriented string-searching algorithms can be used with UTF-8 (as a character is the same as a "word" made up of that many bytes), optimized versions of byte searches can be much faster due to hardware support and lookup tables that have only 256 entries.
• Efficient to encode using simple bitwise operations. UTF-8 does not require slower mathematical operations such as multiplication or division (unlike Shift JIS, GB 2312 and other encodings).
• 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. Self-synchronization also takes more space.

• 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 ASCII-based 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.
• 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 for all modern European languages.
  • Text in (for example) Chinese, Japanese or Devanagari will take more space in UTF-8 if there are more of these characters than there are ASCII characters. This is likely when data mainly consist of pure prose, but is lessened by the degree to which the context uses ASCII whitespace, digits, and punctuation.[nb 2]
  • Most of the rich text formats (including HTML) contain a large proportion of ASCII characters for the sake of formatting, thus the size usually will be reduced significantly compared with UTF-16, even when the language mostly uses 3-byte long characters in UTF-8.[nb 3]
• Most communication (e.g. HTML and IP) and storage (e.g. for Unix) 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.

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 three-byte UTF-8 encoding, resulting in six-byte sequences rather than four 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.

In Modified UTF-8 (MUTF-8),[51] the null character (U+0000) uses the two-byte overlong encoding 11000000 10000000 (hexadecimal C0 80), instead of 00000000 (hexadecimal 00). Modified UTF-8 strings never contain any actual null bytes but can contain all Unicode code points including U+0000,[52] 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 (if it is the platform's default character set or as requested by the program). However it uses Modified UTF-8 for object serialization[53] among other applications of DataInput and DataOutput, for the Java Native Interface,[54] and for embedding constant strings in class files.[55] The dex format defined by Dalvik also uses the same modified UTF-8 to represent string values.[56] Tcl also uses the same modified UTF-8[57] as Java for internal representation of Unicode data, but uses strict CESU-8 for external data.

WTF-8 (Wobbly Transformation Format – 8-bit) is an extension of UTF-8 where the encodings of unpaired surrogate halves (U+D800 through U+DFFF) are allowed.[58] This is necessary to store possibly-invalid UTF-16, such as Windows filenames. Many systems that deal with UTF-8 work this way without considering it a different encoding, as it is simpler.

The term "WTF-8" has also been used humorously to refer to erroneously doubly-encoded UTF-8[59][60] sometimes with the implication that CP1252 bytes are the only ones encoded.[61]

Extensions have been created to store invalid byte sequences as UTF-16. They include MirBSD OPTU-8/16[62] which converts them to U+EF80…U+EFFF in a Private Use Area, and Python "surrogateescape"[63] which converts them to U+DC80…U+DCFF which are low (trailing) surrogate values and thus "invalid" UTF-16.

For the encoding to be reversible, the UTF-8 encoding of the error code points must be disallowed. For the Python version this makes the encoding incompatible with WTF-8 or CESU-8 (though only for 128 code points). Re-encoding must also be careful of a malicious actor arranging the errors to convert back to valid UTF-8 encoding. For these reasons is far easier to store invalid UTF-16 as WTF-8 than to store invalid UTF-8 as some form of UTF-16, so using byte strings is best if you must work with both.