Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10 5/3/83; site ucbcad.UUCP
Path: utzoo!linus!decvax!microsoft!uw-beaver!tektronix!ucbcad!ingres
From: ingres@ucbcad.UUCP
Newsgroups: net.cog-eng
Subject: Re: Re: thousands of invocation flags - (nf)
Message-ID: <393@ucbcad.UUCP>
Date: Thu, 1-Sep-83 21:41:08 EDT
Article-I.D.: ucbcad.393
Posted: Thu Sep  1 21:41:08 1983
Date-Received: Fri, 2-Sep-83 16:14:46 EDT
Sender: notes@ucbcad.UUCP
Organization: UC Berkeley, CAD Group
Lines: 33

#R:ihlts:-18600:ucbcad:31200001:000:1328
ucbcad!ingres    Sep  1 11:44:00 1983

	/***** ucbcad:net.cog-eng / ihlts!ryl /  7:57 am  Aug 31, 1983*/

	   One simple quasi-standard that was suggested once
	was to require at least one argument for every command.
	Typing the command with no arguments gives a three-or-four
	line use message.  This has the advantage of not requiring
	an extra flag for help, and imposes a kind of consistency across
	all commands.  Some other simple things were to put a prefix
	on each command in the application (e.g., all commands related to
	accting would start with acct), and to accept some standard
	flags on all commands (like -v for verbose).  However,
	these suggestions were entirely too sensible, and therefore
	never implemented.

		Long live ls -alphabet,
		Bob Lied	...ihnp4!ihlts!ryl
	/* ---------- */

I see.  And what argument would you like me to give to "who"?  To "ps"?
To "rogue"?  The only type of argument in your scheme would be a null
operative which only says "do not give me the option list".  I would
prefer to have a standard way to ask for the option list.  Long before
Bell programs started using "-" by itself to mean "use standard input",
I started to use that in all my programs, but you see what happened.
The idea of using "--" for that sounds reasonable.

		Ken Arnold

Of course, the standard method of asking for all the option is "man"...