Question
有很大一个电子图书馆，里面每本书的每一页都是 OCR 转换出来的 text，所有每页大
约有 5%的 error（转换错误，错误分割单词，跳脱。。。）。设计一个方法判定图书馆
里是否有完全一样的书（duplicate），或者将来有书进来时判定同样的书是否已存在。
Solution
Very large string matching, we mush use hashing or similar technique. Since books have error, we need to do fuzzy search/matching. Bloom filter is designed to do this!
Refer to [Design] Big Data - Fuzzy Search Url (Bloom Filter).