Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!snorkelwacker.mit.edu!bloom-beacon!eru!hagbard!sunic!sics.se!fuug!funic!nntp.hut.fi!usenet
From: Tapani.Lindgren@HUT.Fi (Tapani Lindgren)
Newsgroups: comp.compression
Subject: Re: Is arithmetic compression a crock?
Message-ID: <1991Jun9.191353.10030@nntp.hut.fi>
Date: 9 Jun 91 19:13:53 GMT
References: <1991May30.035552.14435@alembic.acs.com> <6285@ns-mx.uiowa.edu>
Sender: t31694c@kaira.hut.fi (Tapani Lindgren)
Organization: Helsinki University of Technology, FINLAND
Lines: 12
Nntp-Posting-Host: kaira.hut.fi

Douglas W. Jones <jones@pyrite.cs.uiowa.edu> writes:
>In a multiple-state source model, many characters will be so common in
                                   ^^^^
>some states that they will code to less than a bit when arithmetic
>codes are used.

May I rephrase this statement so noone will get the wrong idea:
There may be some states, where _one_ character (or any symbol)
is so common (probability > 50%) that it will code
to less than a bit.

Tapani Lindgren