Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!amdahl!oliveb!jerry
From: jerry@oliveb.UUCP (Jerry F Aguirre)
Newsgroups: comp.bugs.4bsd,comp.text
Subject: Re: bug in spell
Message-ID: <628@oliveb.UUCP>
Date: Tue, 10-Mar-87 14:13:51 EST
Article-I.D.: oliveb.628
Posted: Tue Mar 10 14:13:51 1987
Date-Received: Wed, 11-Mar-87 04:36:42 EST
References: <13292@cmcl2.UUCP> <1943@utah-gr.UUCP>
Reply-To: jerry@oliveb.UUCP (Jerry F Aguirre)
Organization: Olivetti ATC; Cupertino, Ca
Lines: 24
Keywords: spell spellin aniversery
Xref: mnetor comp.bugs.4bsd:213 comp.text:541

In article <1943@utah-gr.UUCP> thomas%spline.UUCP@utah-gr.UUCP (Spencer W. Thomas) writes:
>that spell would run *much* slower if it attempted to look each word
>up in a dictionary.

Not true!  I have a version of spell that I wrote myself.  It uses a
long word list (>90,000 words).  An index based on the first two letters
provides a seek pointer into the list and the program "guestimates" the
final position based on letter frequency.

I like it because it displays the entire input file with unfound words
in reverse video.  This allows me to see the exceptions in context and
verify whether I really wanted that funny character string.  Also it is
not subject to random acceptance of funny strings.

I have run timing test comparing my version to the standard spell and it
runs about 10% slower.  Acceptable to me given the greater
functionality.

I also tried out a version using "dbm" to access the word list and the
speed was comparable.  This could add some improved functionality as the
entry could contain additional information about afixes and such.  The
speed was comparable.

					Jerry Aguirre