Path: utzoo!utgpu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!sdd.hp.com!elroy.jpl.nasa.gov!aero!srt
From: srt@aerospace.aero.org (Scott "TCB" Turner)
Newsgroups: alt.sources
Subject: Re: Fast strcmp() wanted.
Message-ID: <87739@aerospace.AERO.ORG>
Date: 5 Oct 90 17:26:40 GMT
References: <6003@hplabsz.HPL.HP.COM> <1145@exodus.Eng.Sun.COM> <CEDMAN.90Oct4171710@lynx.ps.uci.edu>
Organization: The Aerospace Corporation, El Segundo, CA
Lines: 16

In article <CEDMAN.90Oct4171710@lynx.ps.uci.edu> cedman@lynx.ps.uci.edu (Carl Edman) writes:
>... I still hold that the CRC-method could be by far the most
>efficient under some conditions:
>
>	- each word is checked often.
>	- strings are long... [etc]

Spell-checking is an interesting problem which involves comparisons
between many strings, and which admits of many algorithmic
improvements.  There was a very good article on how the Unix "spell"
program works in a past issue of Communications, probably in a Pearls"
"column.  I can't recall enough of the article to go into
detail here, but the algorithm finally developed involves a checksum
type approach similar to what Carl suggested.

					-- Scott Turner