BOM（Byte-order Mark）

Byte-order Mark

Always prefix a Unicode plain text file with a byte-order mark. Because Unicode plain text is a sequence of 16-bit code values, it is sensitive to the byte ordering used when the text was written.

A byte-order mark is not a control character that selects the byte order of the text; it simply informs an application receiving the file that the file is byte ordered.

Ideally, all Unicode text would follow only one set of byte-ordering rules. This is not possible, however, because microprocessors differ in the placement of the least significant byte: Intel® and MIPS® processors position the least significant byte first, whereas Motorola processors (and all byte-reversed Unicode files) position it last. With only a single set of byte-ordering rules, users of one type of microprocessor would be forced to swap the byte order every time a plain text file is read from or written to, even if the file is never transferred to another system based on a different microprocessor.

The preferred place to specify byte order is in a file header, but text files do not have headers. Therefore, Unicode has defined a character (0xFEFF) and a noncharacter (0xFFFE) as byte-order marks. They are mirror byte-images of each other.

Since the sequence 0xFEFF is exceedingly rare at the outset of regular non-Unicode text files, it can serve as an implicit marker or signature to identify the file as a Unicode file. Applications that read both Unicode and non-Unicode text files should use the presence of this sequence as an indicator that the file is most likely a Unicode file. (Compare this technique to using the MS-DOS EOF marker to terminate text files.)

When an application finds 0xFEFF at the beginning of a text file, it typically processes the file as though it were a Unicode file, although it may also perform further heuristic checks to verify that this is true. Such a check could be as simple as testing whether the variation in the low-order bytes is much higher than the variation in the high-order bytes. For example, if ASCII text is converted to Unicode text, every second byte is zero. Also, checking both for the linefeed and carriage-return characters (0x000A and 0x000D) and for even or odd file size can provide a strong indicator of the nature of the file.

When an application finds 0xFFFE at the beginning of a text file, it interprets it to mean the file is a byte-reversed Unicode file. The application can either swap the order of the bytes or alert the user that an error has occurred.

The Unicode byte-order mark character is not found in any code page, so it disappears if data is converted to ANSI. Unlike other Unicode characters, it is not replaced by a default character when it is converted. If a byte-order mark is found in the middle of a file, it is not interpreted as a Unicode character and has no effect on text output.

The Unicode value 0xFFFF is illegal in plain text files and cannot be passed between Win32 functions. The value 0xFFFF is reserved for an application's private use.