UTF-8

UCS Transformation Format—8-bit (UTF-8) is a byte-oriented Unicode character encoding. It offers good compatibility with ASCII, because codes 0–127 (00–7F hexadecimal) represent the equivalent ASCII characters, and these codes are never used in any other context.

UTF-8 is most efficient with scripts that make heavy use of the Roman alphabet. With other scripts it may not provide as efficient an encoding as UTF-16.

Format
A Unicode code point is encoded as either 1, 2, 3, or 4 bytes. (Early versions of UTF-8 defined sequences with more than 4 bytes, but they are not valid Unicode.) Code points U+0000 to U+007F use 1 byte, U+0080 to U+07FF use 2, U+0800 to U+FFFF use 3, and U+10000 to U+10FFFF use 4.

(The version specified in the previous paragraph may be called UTF-M-8, to distinguish it from UTF-G-8 which is the extension to 31-bits and the original specification (originally simply called "UTF-8"). An extension to 63-bits, which is allegedly supported in Perl, is called UTF-E-8. The name "UTF-8" by itself often refers to UTF-M-8.)

In MySQL
MySQL calls it utf8mb4, after making the unfortunate move of using the name 'utf8' to designate a limited subset that extends only to three bytes covering the BMP range (excluding characters past U+FFFF, or #65535 decimal). This continues a long computer-industry tradition of mangling character encoding standards, from PETSCII to serving Windows 1252 as ISO 8859-1.

Specifications

 * STD 63
 * RFC 3629 (2003-11)
 * RFC 2279 (1998-01)
 * RFC 2044 (1996-10)
 * Unicode 6.0, Chapter 3 (2011) – §3.9 D92, §3.10 D95
 * ISO/IEC 10646:2003 Annex D (2003)