Algorithms For Dummies

 

»

Encode shorter sequences using a unique series of bits. When using 

variable length bit sequences, you have to ensure that you can’t misinterpret 

a shorter sequence in place of a longer one because they are similar. For 

instance, if the letter a in binary is 110 and z is 110110, you could misinterpret 

the letter z as a series of two-letter a characters. Huffman encoding avoids 

this problem by using prefix-free codes: The algorithm never reuses shorter 

sequences as initial parts of longer sequences. For example, if a is 110, then z 

will be 101110 and not 110110.