Path: utzoo!utgpu!water!watmath!clyde!att!osu-cis!tut.cis.ohio-state.edu!bloom-beacon!mit-eddie!rutgers!pyrnj!dasys1!tneff
From: tneff@dasys1.UUCP (Tom Neff)
Newsgroups: comp.binaries.ibm.pc.d
Subject: Re: Compression Standard Preliminary Vote (Complete)
Message-ID: <6441@dasys1.UUCP>
Date: 13 Sep 88 18:50:54 GMT
References: <9037@swan.ulowell.edu>
Reply-To: tneff@dasys1.UUCP (Tom Neff)
Organization: Independent Users Guild
Lines: 55

There must have been some confusion in the way the compression standard
vote was conducted.  The answers as tabulated seem like nonsense.  The
only existing "PK" compression [actually packaging - bundling is involved
as well as compression] standard is the ARC format.  There is no such
thing as a "fast" format or a "slow" format, only fast and slow utilities
to manipulate the format.  Hence the first two entries ("ARC" and "PK")
in the vote tabulation really refer to the same thing.  If combined, they
still don't outpoll ZOO, so the outcome would not be affected, but I
would feel better about the poll if the results looked more sensible.
It looks awfully as though the responding users confused the choice of
extracting utility with the choice of underlying format.  Only the second
issue is of importance to this debate; presumably the individual user is
welcome to use whatever he or she wants to build or extract the chosen
format.

I would further question why we even need to settle on exactly one
packaging format, either for ftp/uucp archives or binary news
distribution.  The following points suggest themselves:

 1. There are only a few formats in general use, each of them good.
    Choosing between them has become an issue of personal preference,
    and most sites already have the means to extract all of them.
    Those that don't can probably get what they need from colleagues or
    from us.

 2. The specific choice of packaging format in which a product is
    distributed is usually made by the author, and it's not unusual
    these days to see clauses in the READ.ME saying that distribution
    is free but must be "as is."   Forcibly tearing down a ZOO to
    make an ARC (or vice versa) would violate any such instructions;
    the restriction is a reasonable one and should not disqualify a
    product from Usenet distribution as long as the format is one of
    the "majors."

 3. The one exception I would make to point 2 above is with self
    extractors.  Self-extraction sounds great on paper and some authors
    go for it, but it is nonetheless a Bad Thing, because it usually
    restricts the flexibility a downloading user has to inspect or
    manipulate his own copy before extracting or running it; and most
    importantly because it leaves the user vulnerable to system damage
    from transmission errors or trojan mischief.  Shell archives are
    of course a species of self-extraction, but they lay their operation
    bare to user inspection and rely on tools you are responsible for
    providing yourself, so are inherently safer.

 4. Far more important than selecting one archive format is settling
    on a consistent pattern of ( chop + uuencode + shar ) etc. for
    posting in news articles.  Regardless of the file you end up
    with when you're done, the mechanism for creating the file should
    be foolproof, consistent and robust.  We need to put a little more
    thought into this, I think.
-- 
Tom Neff			UUCP: ...!cmcl2!phri!dasys1!tneff
	"None of your toys	CIS: 76556,2536	       MCI: TNEFF
	 will function..."	GEnie: TOMNEFF	       BIX: t.neff (no kidding)