Path: utzoo!utgpu!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!mailrus!ames!killer!texbell!sugar!peter
From: peter@sugar.uu.net (Peter da Silva)
Newsgroups: comp.sys.amiga
Subject: Re: Devices in general
Message-ID: <3066@sugar.uu.net>
Date: 7 Dec 88 13:49:34 GMT
References: <5816@louie.udel.EDU> <3061@sugar.uu.net> <4972@bsu-cs.UUCP>
Organization: Sugar Land Unix - Houston, TX
Lines: 40

In article <4972@bsu-cs.UUCP>, dhesi@bsu-cs.UUCP (Rahul Dhesi) writes:
> In article <3061@sugar.uu.net> peter@sugar.uu.net (Peter da Silva) writes:
> [some name-calling because zoo won't write its archives to standard output]

...and won't read files or archives from standard input, etc...

> The reason zoo archives aren't written to standard output is because
> doing so requires sequential writes.  If they were, updating an archive
> would require copying the entire archives. This would mean that you
> could not manipulate large archives in limited disk space.

> Database handling programs in general don't write purely sequentially.

This is a bit of a straw man. Consider...

	You can maintain the archive any way you like it, but having an
option to write it out as a stream may well save on disk space. Just ask
anyone who's used cpio. How about:

	1> run zoo -stream files >PIPE:x

[switch windows]

	Kermit> send pipe:x

	Yes, when you just want to add a file to an archive, it's more
efficient to do it in place. On the other hand, the most common mode of
using zoo is all-or-nothing. It's more often used as an archiver than a
database.

	Finally, since zoo is mainly a PC product... calling it pc-ware
is accurate. I apologise for the word "bozo", though. It was uncalled-for.

PS: is there a program that'll generate a list of filenames suitable for
input to 'zoo I'? This is a very cpio-ish option, you know.
-- 
		    Peter da Silva  `-_-'  peter@sugar.uu.net
		     Have you hugged  U  your wolf today?

	          Disclaimer: My typos are my own damn busines#!rne