Compaction Algorithm (pack vs compress)

[tweten at ames-nas.arpa]

From: Guy Harris <guy%sun.uucp at BRL.ARPA>

	The only disadvantage is that it tends to eat virtual (and physical)
	memory; as Peter Honeyman once put it, "more than two compresses makes
	disks dance!"  I don't care, since my machine is a one-user machine,
	but on a multi-user machine this may make a difference.  I'm also not
	sure whether the latest "compress" uses memory that freely.

Compress version 4.0 reduced its memory requirement to the point that I
was able to port the full-blown (16-bit compressed codes) version to PC/MS-DOS,
using the Microsoft C compiler.  Compiled for 16-bit codes, it requires about
460 K bytes of user memory on a PC.  Naturally, the 12-bit code compilation
takes much less.

To anticipate any inquiries, source code is available from the Info-IBMPC
public domain source library, in directory <info-ibmpc> at USC-ISIB.ARPA.  By
manipulating compile-time definitions one can compile all the standard
(4.2, System V, etc.) versions from it, in addition to being able to compile
PC/MS-DOS big and small versions.