Character Representation

Character Representation

Western languages usually require only one byte for each character. Asian languages, however, often require two or even four bytes per character, and some Asian encodings allow a variable number of bytes per character.

The two kinds of encodings that allow more than one byte per character are

multibyte (MB) characters are of variable size
wide characters (WC or wchar characters) are a fixed number of bytes long)

The application developer must decide where to use WC and MB characters and strings:

Multibyte strings are almost the default: string I/O uses MB, MB code works for ASCII and ISO 8859, and MB characters use less space than do wide characters. However, manipulating individual characters within a multibyte string is difficult.
Note: Traditional strings are merely a special case of multibyte strings, where every character happens to be one byte long and there is only one codeset. All MB code, including conversion to and from wchars, works for traditional ASCII, or ISO 8859, strings.
Applications that do heavy string manipulation typically use WC strings for such activity, because manipulating individual WC characters in a string is much simpler than doing the same thing with MB characters. So wide characters are used as necessary to provide programming ease or runtime speed; however, they take up more space than MB characters.
Note: WC is system dependent--applications should not use it for I/O strings or communication.