Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!linus!philabs!cmcl2!harvard!caip!princeton!allegra!ulysses!burl!clyde!cbosgd!ihnp4!inuxc!pur-ee!uiucdcs!okstate.UUCP!uokvax.UUCP!emjej
From: emjej@uokvax.UUCP
Newsgroups: net.micro
Subject: Re: Orphaned Response
Message-ID: <3400082@uokvax.UUCP>
Date: Fri, 30-May-86 10:46:00 EDT
Article-I.D.: uokvax.3400082
Posted: Fri May 30 10:46:00 1986
Date-Received: Tue, 3-Jun-86 23:33:30 EDT
References: <690@jade.BERKELEY.EDU>
Lines: 35
Nf-ID: #R:jade.BERKELEY.EDU:690:uokvax.UUCP:3400082:000:1781
Nf-From: uokvax.UUCP!emjej    May 30 09:46:00 1986


/* Written  4:17 pm  May 10, 1986 by mwm@ucbopal.berkeley.edu in net.micro */
I feel that the minimum set of features [for a generic archiver] should be:

	1) PD versions available for most major OS's, both for micros
		and Internet hosts.
	2) The headers have no non-ASCII/EBCDIC characters, or TABS.
	3) A checksum of some kind is included on each file.
	4) The format include provisions for creating directories.
	5) It shouldn't choke on binary data in the archive.
/* End of text from net.micro */

Agreed with wholeheartedly, <mike...Some slight further comments:

re (3): That checksum should *not* involve whatever is used for the EOL
	marker, since that varies from system to system, and one would
	ideally like to let existing tools (Kermit, mainly) handle that
	conversion without fritzing the checksum.  Similar reasoning
	implies that the length of the file shouldn't be involved in
	whatever checksum is used.  (While it won't matter for archives
	with binary data, for shipping sources it would make life a
	good bit easier.)

Speaking of EOL...some places would, I'm sure, love to be able to compress
archives, and a person downloading them (or, if they're sent via UUCP,
everybody on the intervening route :-) would love to have them decompressed
after they arrive.  Again worrying about EOL markers, it strikes me that
we should also consider a compression routine that doesn't mess with EOL,
however it is represented on the local software.  (Doesn't mean that it
can't run-length encode it or whatever; just that EOL should be left as
whatever it is locally, again so that transmission methods that do
conversion can do it so we don't have to.)  I wonder how much compression
you lose if you hack, say, Lempel-Ziv to let EOL pass unscathed? 

						James Jones