Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!cs.utexas.edu!ginosko!aplcen!haven!adm!smoke!gwyn
From: gwyn@smoke.BRL.MIL (Doug Gwyn)
Newsgroups: comp.sys.apple
Subject: Re: ShrinkIt for the IIgs (very long, 11k)
Message-ID: <11408@smoke.BRL.MIL>
Date: 24 Oct 89 21:08:24 GMT
References: <8909300944.AA10524@trout.nosc.mil> <10743@phoenix.Princeton.EDU> <401@batman.moravian.EDU>
Reply-To: gwyn@brl.arpa (Doug Gwyn)
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 21

In article <401@batman.moravian.EDU> nicholaA@batman.moravian.EDU (Andy Nicholas) writes:
>          I choose to make GS/ShrinkIt free.

Thanks!

>Type-II LZW goes a step farther to implement a clear of the hash-table ONLY
>when the table fills up.  This allows better compression to occur on
>stuff that compresses well with LZW: text files.

There is clearly room for substantial improvement in data compression
technology.  When I was implementing a LZW decoder for a GIF-to-rational
image translator, I discovered several places in the GIF LZW algorithm
(apparently borrowed from the BSD UNIX "compress" program) where better
compression was attainable.

There is at least one good textbook on data compression (I don't remember
the authors and the book is not at hand); perhaps a concerted effort
should be made to devise the best feasible compression algorithms first
and implement them after we've completely solved the problem.  There are
several existing candidates for "good enough" compressers while this
problem is being researched.