Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!apple!usc!wuarchive!gumby!wmu-coyote!campbell
From: campbell@sol.cs.wmich.edu (Paul Campbell)
Newsgroups: comp.compression
Subject: Re: Is arithmetic compression a crock?
Message-ID: <1991May31.150751.27894@sol.cs.wmich.edu>
Date: 31 May 91 15:07:51 GMT
References: <1991May30.035552.14435@alembic.acs.com>
Organization: Western Michigan Univ. Comp. Sci. Dept.
Lines: 12

To be fair, it would be a bad idea to compare this to a string compressor
because that's not what it is. It would be fair to compare it to a Huffman
encoding scheme. It does better than Huffman, and it is the same class of
data compression as the Huffman compressor. Neither were intended to beat out
string compressors except in a few cases. They both rely on the frequency
of occurance of a specific character. You could, for instance, run the output
of the compress program through the input of the arithmetic encoder to get
even better compression.

Also, as mentioned elsewhere, PLEASE don't use a fixed model unless you scan
the data first to generate a table for your specific case or do an adaptive
implementation.