Recognizing & grouping anagrams (Homework for Data Structures)
Implemented by mapping letters to prime numbers, to map words to composite numbers.
output.txt was generated from input.txt, which was designed to cover cases described in the assignment.
All Anagrams.sort *.txt were generated from my /usr/share/dict/words at the time.
To generate numbers, each of the 26 letters of our alphabet is assigned a unique prime number; the number representing a word or phrase is the product of those prime numbers raised to the power of the number of occurrences of the corresponding letter in the word or phrase. If I were to take this further, generating phrases which are anagrams of a given phrase is reduced to a factoring problem, which is well understood and I suspect more efficient than string searching.
(from the Algorithms section of the lengthy descriptive comments in anagram.java)