Python Essential Reference, Second Edition

Unicode I/O

Internally, Unicode strings are represented as sequences of 16-bit integer character values.As in 8-bit strings, all characters are the same size, and most common string operations are simply extended to handle strings with a larger range of character values. However, whenever Unicode strings are converted to a stream of bytes, a number of issues arise. First, to preserve compatibility with existing software, it may be desirable to convert Unicode to an 8-bit representation compatible with software that expects to receive ASCII or other 8-bit data. Second, the use of 16-bit characters introduces problems related to byte ordering. For a Unicode character U+HHLL, little endian encoding places the low-order byte first, as in LL HH