From:	CRDGW2::CRDGW2::MRGATE::"SMTP::CRVAX.SRI.COM::RELAY-INFO-VAX"  1-AUG-1989 09:54
To:	MRGATE::"ARISIA::EVERHART"
Subj:	Re: What's ZOO

Received: From KL.SRI.COM by CRVAX.SRI.COM with TCP; Tue,  1 AUG 89 06:18:59 PDT
Received: from RELAY.CS.NET by KL.SRI.COM with TCP; Tue, 1 Aug 89 06:13:44 PDT
Received: from crdgw1.ge.com by RELAY.CS.NET id aa16274; 1 Aug 89 9:13 EDT
Received:  by crdgw1.ge.com (5.57/Ultrix 3.0 (1.49))
	 id AA27079; Tue, 1 Aug 89 09:13:00 EDT
Message-Id:  <8908011313.AA27079@crdgw1.ge.com>
Date:  1 Aug 89 09:09 EST
From: EVERHART%ARISIA.DECNET@crdgw1
Subject: Re: What's ZOO
To: info-vax <@RELAY.CS.NET:info-vax@KL.SRI.COM>

ZOO for VMS was on the F88 and S89 VAX SIG tapes...sources
included.
   ZOO is a multifile archiver. Unlike old DECUS C ARC, which just
stored files, or ARC in the Software Tools, which did likewise,
ZOO compresses files before storing them. The result is that a ZOO
file is a SINGLE file which contains many files, all in compressed
form (or not compressed where the compression failed to reduce the
size). Compression is done via 12 bit LZW compress algorithm.
   One can also make VMS Backup savesets and use LZCOMP to compress
them. This is significantly faster and produces smaller files, for
a set of tests I ran recently. It also preserves RMS attributes
for files, which ZOO does not (everything becomes stream_lf).
   However, ZOO archives can have their directories listed without
decompressing the whole thing, and you can extract individual pieces
without decompressing the whole. This is a major advantage. By using
the FILE utility to record RMS attributes, and including that output
with the ZOO, you can also create a ZOO file from which the original
files WITH attributes can be recreated.

   The following considerations apply to various methods:
1. Backup to a saveset (/group=0/nocrc) followed by LZCOMP:
   Advantages: Saves all files and attributes
   Faster than ZOO, ARC, or commercial compressors, even though
	there are two steps
   Smallest available compress output for multifile saves, for
	all input types tested.

   Disadvantages: Must decompress entire saveset to extract any
	part. Requires significant disk space and time.
   Cannot get directory of contents without first decompressing
	saveset.

2. ARC (another PC based file compressor)
   Advantages: Can compress with 12 bit LZW or Huffman.
   Directory of all files possible
   Can extract one part without extracting all.

   Disadvantages: Slowest of all compression tools tested.
   filenames cannot be longer than 8 chars, 3 char extensions
   Attributes lost (use FILE as with ZOO)
   Originator of the format given to lawsuits for what appear to
	me frivolous and destructive purposes
   
3. ZOO
   Advantages: Faster than ARC by significant amounts
   Filenames and directory paths stored even where long.
   Can store entire directory trees and extract creating directory
	tree intact.
   Generally does as well as ARC at compression; average about
	even between them.
  
   Disadvantages: Must use FILE utility to reconstruct RMS
	attributes if stream_lf output not adequate.
   
4. Compressing virtual disks (in the works)
   Advantages: Stores RMS attributes.
   Files can be used in place (decompression on the fly as pieces
	are read)
   Directory of compressed data possible without decompressing all.
   (Possible with ZOO also...forgot to put that above)
   Simplest to use once set up.

   Disadvantages: Not done yet, so no idea how well it will perform
	in space or time. (First cut will do 32 blocks at a time)
   Requires privileges to set up...no other technique does.
   Inherently nonportable to non VMS systems.

  We used ZOO to compress a number of source directories on recent
symposium tapes because it is fairly convenient to use and we needed
to make some room. Ideally a format that might store RMS attributes
as well as full filenames and directories, and restore them
automatically (rather than requiring use of a tool like FILE as
one needs with ZOO) would be preferable. It needs also to be
freely distributable. The ZIP format is under development by Phil
Katz of PKware (target of one of the frivolous lawsuits I mentioned)
and looks like it may fill the bill. The format has reserved space
for attribute records and the intention of the person porting it to
VMS was to save/restore RMS attributes.
   Unfortunately, a compressing disk is not useful for symposium tapes,
because too many people using the material would have problems getting
the necessary privileges to load the driver and start the host processes
detached which would be needed to use that technology.
   Given the current situation, I regard ZOO and FILE as the best
general purpose combination of tools for file archiving now available.
By using FILE to record RMSO