Simhashing is a style of characterizing the similarity of data. The author begins with the idea that we can discard the first characters of { aaarock, aabjeep, aaareep } to prefer the latter two as most similar and concludes with computing the hamming distance of data.