Path: utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!nstn.ns.ca!news.cs.indiana.edu!widener!ukma!dftsrv!mimsy!nrl-cmf!Ra.nrl.navy.mil!ricks From: ricks@nrl.navy.mil (Richard Schumeyer) Newsgroups: comp.compression Subject: Re: Is arithmetic compression a crock? Message-ID: Date: 31 May 91 13:46:32 GMT References: <1991May30.035552.14435@alembic.acs.com> Sender: usenet@nrl-cmf.UUCP Organization: NRL Lines: 19 In-reply-to: csu@alembic.acs.com's message of 30 May 91 03:55:52 GMT In article <1991May30.035552.14435@alembic.acs.com> csu@alembic.acs.com (Dave Mack) writes: [stuff deleted] > Is this a really bad implementation, a really bad algorithm, or > is arithmetic compression basically useless for text compression? > > -- > Dave Mack Arithmetic compression gives the best compression for a given model. If your model stinks, so will your compression. Did you use the static model given in the paper? Much better text models exist, like the string-based model in compress. -- ------------------------------------------------------------ Rick Schumeyer ricks@luke.nrl.navy.mil Naval Research Lab (202)404-7966 Space Science Division ------------------------------------------------------------