2023-07-13 UTC
# 
[KevinMarks] it broadly applies to the class of compressors that are squeezing repetition out by finding longer and longer byte sequence matches. We know text is very redundant there. It also works quite well for language identification, in that different languages have different byte sequence distributions.