Experience with a space efficient way to store a dictionary
The paper, “Computer Programs for Detecting and Correcting Spelling Errors” by James L. Peterson [3], listed methods for checking and correcting spelling errors. One significant method, however, was not included: a probabilistic technique suggested by Carter, Floyd, Gill, Markovsky, and Wegman [1]. The present note discusses aspects of these practical space efficient algorithms for testing set membership—a simple abstraction of looking a word up in a dictionary. An implementation of one of these algorithms uses only 20 percent of the space used by the Stanford SPELL program described by Peterson.