Path: utzoo!attcan!uunet!mcsun!hp4nl!sci.kun.nl!phoibos!ge
From: ge@phoibos.cs.kun.nl (Ge Weijers)
Newsgroups: comp.binaries.ibm.pc.d
Subject: Re: PKZIP 1.10 vs. LHARC
Keywords: PKZIP
Message-ID: <1727@wn1.sci.kun.nl>
Date: 27 Jun 90 16:16:36 GMT
References: <6782@amelia.nas.nasa.gov>
Sender: root@sci.kun.nl
Lines: 28

roelofs@amelia.nas.nasa.gov (Ender Wiggin) writes:

>And the reason you haven't seen a zip program for Unix or VMS or whatever
>is that it doesn't exist.  You're welcome to write it, of course. :)  I
>took one look at Katz's description of the Implosion format and decided
>cloning it was not for the faint of heart.  But I suspect the literature
>contains the basic algorithm for compression using Shannon-Fano trees; if
>Katz himself had invented it, they'd probably be called "Katz trees"  or
>something similar, n'est-ce pas?  Perhaps in Communications of the ACM...
>any computer scientists/compression specialists out there know for sure?

A good book (still not for the faintharted) is
"Data Compression" by James Storer. published by Computer Science Press,
I believe.

I faintly remember something about Shannon-Fano trees in there. After reading
it you'll certainly know why LHARC is slow. It uses 2 compression schemes,
first a Lempel-Ziv variant which is much more complex than LZW, followed
by dynamic Huffman coding on the high-order 6 bits of the 12-bit code.

Where could I find the Implosion format, by the way?
I'd like to have a peek at it.

Ge' Weijers.
Ge' Weijers                                    Internet/UUCP: ge@cs.kun.nl
Faculty of Mathematics and Computer Science,   (uunet.uu.net!cs.kun.nl!ge)
University of Nijmegen, Toernooiveld 1         tel. +3180612483 (UTC+1,
6525 ED Nijmegen, the Netherlands               UTC+2 march/september