That's optimal for 100% accuracy. You can get much lower memory usage using Bloom filters (basically, hashing words into a small range of hashcodes, expecting c

That's optimal for 100% accuracy. You can get much lower memory usage using Bloom filters (basically, hashing words into a small range of hashcodes, expecting collisions, and storing a mask of all the hashcodes that correspond to words), at the expense of false positives. (thinking a mispelled word is correct.) This was a common solution in the 80s.