English dictionary searching with little extra space

This is the authors' abstract. We don't add key points for this paper.

When text is typeset using a computer-based system, it can also be checked for spelling errors automatically and efficiently. Several methods of spelling error detection have been proposed. Morris et al. [MORR75] study statistical properties of English words, and describe an algorithm to catch possible typos by examining the relative frequency of trigrams (3-letter combinations). Kernighan et al. [KERN78] report a more conventional approach which looks up each word in a machine readable spelling dictionary. 1 Knuth [KNUT73] comments on dictionary storage, and suggests an organization intended to reduce space requirements. Another, somewhat unsophisticated technique sorts all words in a document and prints those which occur infrequently. Of course, habitually misspelled words escape unnoticed in a frequency based system.