SOUNDEX: Knuth's algorithm for encoding names

Автор: Rex Swain

Дата: 1999

Источник: www.rexswain.com

The REXX program SOUNDEX. REX encodes a word (typically a surname) into a 4-character code. These codes may be compared to find other similar words, or detect spelling errors.

For instance, let's say you have a built a database of surnames and their soundex codes:

Name Soundex
Euler E460
Hilbert H416
Hilburt H416
Knuth K530
Philips P412
Phillips P412

If a user enters a name "Hilbart", you might first search for an exact match on the name. Not finding one, you might then compute the soundex code for the name as entered (Hilbart = H416), look that code up the the database, and suggest to the user that they might have meant "Hilbert" or "Hilburt".