Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!dciem!nrcaer!cognos!brianc From: brianc@cognos.uucp (Brian Campbell) Newsgroups: comp.sys.ibm.pc Subject: Re: SQUASHED! Message-ID: <1138@cognos.UUCP> Date: Mon, 20-Jul-87 16:00:18 EDT Article-I.D.: cognos.1138 Posted: Mon Jul 20 16:00:18 1987 Date-Received: Sat, 25-Jul-87 12:11:10 EDT References: <642@cgh.UUCP> <10710@clyde.ATT.COM> <1069@cognos.UUCP> <235NU013809@NDSUVM1> Reply-To: brianc@cognos.UUCP (Brian Campbell) Organization: Cognos Incorporated, Ottawa, Canada Lines: 27 The following is a response from Phil Katz (author of PKX?ARC) outlining the "squashing" algorithm: +-- | Brian, | | Squashing is simply a 13-bit version of the standard 12 bit "crunching" | with the important exception that no non-repeat packing is performed. | | If you have a program that does "crunching", it can easily be modified | to do crunching by removing the DLE coding/encoding stuff, declaring | tables to be 8192 entries in size rather than 4096, and adding 13-bit | code sizes to any getcode() or putcode() type of routines. | | Otherwise, it uses the same dynamic code sizes, clear codes, and | everything else as the crunched files. | | >Phil> +-- Armed with this information it should be a relatively simple task to convert the existing versions of SEA's ARC that run under unix to handle squashed files. Alas, probably slower though. -- Brian Campbell uucp: decvax!utzoo!dciem!nrcaer!cognos!brianc Cognos Incorporated mail: 3755 Riverside Drive, Ottawa, Ontario, K1G 3N3 (613) 738-1440 fido: sysop@163/8