Then, in 1978, they came up with a better algorithm (dubbed [LZ78 | http://www.rasip.fer.hr/research/compress/algorithms/fund/lz/lz78.html], that only required the compressor to output a phrase number and a mismatching character. The phrase extended with the mismatching character would then be added to the dictionary. 

  

 LZ77/78 encoding remained the de-facto standard for compression of data (along with some cool stuff like HuffmanCoding, that's how PKZIP works) until 1984, where Terry __W__elch published "A Technique for High-Performance Data Compression" in Computer magazine. He realised that with a dictionary that had all the possible characters in it already, you wouldn't have to output the mismatching character like in LZ78; you could output a phrase number that matched it. This algorithm is called __LZW encoding__ (Lempel, Ziv, Welch). 

  

 !!! How it works 

  

 The algorithm is really quite simple, and will be described here.