views:

3822

answers:

5

what is the difference between Big Endian byte order and little Endian Byte order.

These both are related to Unicode and UTF16

where we use this?

+16  A: 

Big Endian (BE) / Little Endian (LE) are two ways to organize multi-byte words. For example, when using two bytes to represent a character in UTF-16, there are two ways to represent the character 0x1234:

BE:  12 34
LE:  34 12

In order to decide if a text uses UTF-16BE or UTF-16LE, the specification recommends to prepend a Byte Order Mark (BOM) to the string, representing the character U+FEFF. So, if the first two bytes of a UTF-16 encoded text file are FE, FF, the encoding is UTF-16BE. For FF, FE, it is UTF-16LE.

A visual example: The word "Example" in different encodings (UTF-16 with BOM):

ASCII:     45 78 61 6d 70 6c 65
UTF-16BE:  FE FF 00 45 00 78 00 61 00 6d 00 70 00 6c 00 65
UTF-16LE:  FF FE 45 00 78 00 61 00 6d 00 70 00 6c 00 65 00

For further information, please read the Wikipedia page of Endianness and/or UTF-16.

Ferdinand Beyer
+4  A: 

UTF-16 encodes Unicode into 16-bit values. Most modern filesystems operate on 8-bit bytes. So, to save a UTF-16 encoded file to disk, for example, you have to decide which part of the 16-bit value goes in the first byte, and which goes into the second byte.

Wikipedia has a more complete explanation.

joev
this answer is incorrect. endianess is related to the underlying hardware architecture
Mitch Wheat
You can store a UTF-16 encoded file in either byte order regardless of the underlying hardware.
joev
Given in the context of the question, this answer is perfectly acceptable IMHO
Binary Worrier
@joev: Exactly. It often *is* related to hardware architecture, but needn't necessarily be. For cross-platform compatibility, Unicode encoders/decoders should therefore be able to use either endianness.
Noldorin
A: 

Byte endianness (big or little) needs to be specified for Unicode/UTF-16 encoding because for character codes that use more than a single byte, there is a choice of whether to read/write the most significant byte first or last. Unicode/UTF-16, since they are variable-length encodings (i.e. each char can be represented by one or several bytes) require this to be specified. (Note however that UTF-8 "words" are always 8-bits/one byte in length [though characters can be multiple points], therefore there is no problem with endianness.) If the encoder of a stream of bytes representing Unicode text and the decoder aren't agreed on which convention is being used, the wrong character code can be interpreted. For this reason, either the convention of endianness is known beforehand or more commonly a byte order mark is usually specified at the beginning of any Unicode text file/stream to indicate whethere big or little endian order is being used.

Noldorin
this answer is incorrect. endianess is related to the underlying hardware architecture
Mitch Wheat
UTF-8 is a variable-length encoding, using 1-6 bytes per character and is thus not fixed to a single byte as stated here!
Ferdinand Beyer
Right, so I haven't stated that endianness depends on hardware architecture, but I don't see how my answer is explicitly incorrect. Consider that text files written/read on different architectures must have their endianness known.
Noldorin
@Ferdinand: You are correct - I should mention that *some* variants of UTF-8 do not require it...
Noldorin
Sorry, you still don't got it right. There are no variants of UTF-8 that don't require multiple bytes. If you only use ASCII characters, UTF-8 will represent them using single bytes. All characters with character code >127 will be encoded using multiple bytes!
Ferdinand Beyer
Just for completeness - UTF-8 requires between 1 and 4 bytes. Valid UTF-8 cannot contain more than 4 bytes.
Nemanja Trifunovic
@Nemanja: You are right!
Ferdinand Beyer
@Noldorin: As I said, ASCII characters use a single byte. This is a *property* of UTF-8, not a variant! Using single-bytes, you cannot encode Non-ASCII unicode values.
Ferdinand Beyer
@Ferdinand: Yes, I've realised that since your original correction. Post has been clarified again, as I see your point... though I think I somewhat confused myself in the process of correcting myself. :P
Noldorin
(contd.) I think I'm right in saying that because "words" in UTF-8 are 8-bits/one byte long (invariably, despite the variable length of char codes), then there is no problem with endianness at least.
Noldorin
http://unicode.org/faq/utf_bom.html seems to agree, though again correct me if I'm wrong...
Noldorin
+1  A: 

Ferdinand's answer (and others) are correct, but incomplete.

Big Endian (BE) / Little Endian (LE) have nothing to do with UTF-16 or UTF-32. They existed way before Unicode, and affect how the bytes of numbers get stored in the computer's memory. They depend on the processor.

If you have a number with the value 0x12345678 then in memory it will be represented as 12 34 56 78 (BE) or 78 56 32 12 (LE).

UTF-16 and UTF-32 happen to be represented on 2 respectively 4 bytes, so the order of the bytes respects the ordering that any number follows on that platform.

Mihai Nita
A: 

little-endian: adj.

Describes a computer architecture in which, within a given 16- or 32-bit word, bytes at lower addresses have lower significance (the word is stored ‘little-end-first’). The PDP-11 and VAX families of computers and Intel microprocessors and a lot of communications and networking hardware are little-endian. The term is sometimes used to describe the ordering of units other than bytes; most often, bits within a byte.

big-endian: adj.

[common; From Swift's Gulliver's Travels via the famous paper On Holy Wars and a Plea for Peace by Danny Cohen, USC/ISI IEN 137, dated April 1, 1980]

Describes a computer architecture in which, within a given multi-byte numeric representation, the most significant byte has the lowest address (the word is stored ‘big-end-first’). Most processors, including the IBM 370 family, the PDP-10, the Motorola microprocessor families, and most of the various RISC designs are big-endian. Big-endian byte order is also sometimes called network order.

---from the Jargon File: http://catb.org/~esr/jargon/html/index.html

Paul Reiners