Path: utzoo!censor!geac!torsqnt!jarvis.csri.toronto.edu!cs.utexas.edu!uunet!munnari.oz.au!cluster!nuts!cc!c8923075
From: c8923075@cc.nu.oz
Newsgroups: comp.sys.cbm
Subject: LZW in 64K
Message-ID: <12000.25e95ff6@cc.nu.oz>
Date: 26 Feb 90 06:57:26 GMT
Organization: University of Newcastle
Lines: 29


	I'm currently developing a Lempel-Ziv-Welch de/compression prog for a
CBM64, and would like to ask anyone who've tried the same in a simular 64K
memory limitation situtation for any information that would be of use.

	The prog is being written in 6502 Assembler (No prizes for guessing
why), & I intend it to resemble what you've probabally seen on pirated games
(Although this program is not for a Com.Sci assignment or simular, I have
more scolarly intentions). You run the load & run the prog, it deARCs itself
without accessing any external memory (Those of you familar with a 1541
d/drive would know why).

	What I specifically would like in the way of advice is how to impliment
the LZW table, the deARCed & ARCed files within the same 64k, permitting the
maximum table size (to increase compacting efficency). A proposial would be to
have the table in $0200-$0800, allowing about 510 table entries; the ARCed file
move forward up in memory as far as possible; the decompressed prog
reconstructed from $0801 onwards; and the actual decompression code kept in
$0010 to $01F0 (Stack use kept to a minimum, for a 6502 uses $0100-$01FF)

For anyone interested in compacting data, I've used the following reference:
	"The Design & Implementation of a Microproessor based data compression
	 system" --- by Demetrios A. Prountzos. UMI Dissertation service 1986.
	- This includes "C" language listings on not only LZW, but also a
	 rationalised Huffman algorithm.

-----------------    Assisting 8-bitter owners is good for the soul.
C8923075@cc.nu.oz
-----------------