ZIP was developed based on the LZ77 compression method (LZW). In addition to the LZ77 method, the "Shannon Fano Coding" - which, like the Huffman coding, is an entropy coding - was also used, which essentially improved three things:

• Data in which no pattern is repeated is no longer compressed. In the case of LZW, this state has meant that LZW-compressed data can also be larger than the original data. This is not the case with ZIP.
• Each character has a unique pattern in the library, which means that smaller units than one byte are used for encoding frequently occurring characters such as "e".
• The pattern length depends on the character frequency. This results in fewer entries in the library, which in turn reduces memory requirements.