Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!ut-sally!husc6!cmcl2!brl-adm!brl-smoke!gwyn
From: gwyn@brl-smoke.ARPA (Doug Gwyn )
Newsgroups: comp.unix.wizards
Subject: Re: symbolic links are a botch
Message-ID: <6018@brl-smoke.ARPA>
Date: Tue, 23-Jun-87 15:13:50 EDT
Article-I.D.: brl-smok.6018
Posted: Tue Jun 23 15:13:50 1987
Date-Received: Fri, 26-Jun-87 05:48:06 EDT
References: <7879@brl-adm.ARPA> <2211@bunker.UUCP> <2676@ulysses.homer.nj.att.com> <209@netxcom.UUCP> <2680@ulysses.homer.nj.att.com>
Reply-To: gwyn@brl.arpa (Doug Gwyn (VLD/VMB) <gwyn>)
Organization: Ballistic Research Lab (BRL), APG, MD.
Lines: 24

In article <2680@ulysses.homer.nj.att.com> ekrell@ulysses (Eduardo Krell) writes:
>I don't know the answer but I certainly like the idea of "cd a/b; cd .."
>being equivalent to "cd a" whether "a" is a mount point, RFS mount point
>or a symbolic link.

It seems to me we're gradually wobbling our way towards the idea that
the "current working directory" is really just a "prefix" to be used
for relative path resolution.  This approach is taken by some OSes
(Apple ProDOS being one I'm quite familiar with) and it seems to work
fine.  Indeed, it generalizes nicely so that one can have a handful of
current prefixes simultaneously available, e.g. one for libraries, one
for commands, etc.

Of course UNIX implementations of the "prefix" would remarkably
resemble what they're doing already for the cwd.  (One doesn't have
to ACTUALLY prepend a prefix text string then look up all inodes in
the path starting at the root every time one opens an object!)

With the prefix approach, the natural meaning of "cd" is "set new
prefix", and "cd .." most naturally would mean "trim off the most
local pathname component from the current prefix".

I recommend that we move UNIX toward prefixes as a better naming
concept than "working directory".