Interview Question for Financial Software Engineer at Bloomberg L.P.:

I gave a solution that involved sorting the strings and using that as elements of a new set to check for anagrams. They asked me to reduce the memory footprint of my solution without sacrificing the O(N) time (N = number of strings, we assumed that string lengths were O(1)). I suggested that if the strings represent words in a natural language, we might be able to apply the same Huffman encoding to everything in the set, which maybe could reduce space by, say, 50%. They seemed satisfied with that solution.