Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!seismo!husc6!mit-eddie!genrad!decvax!decwrl!pyramid!prls!philabs!ttidca!redden
From: redden@ttidca.TTI.COM (John Redden)
Newsgroups: comp.unix.wizards
Subject: Re: symbolic links are a botch
Message-ID: <876@ttidca.TTI.COM>
Date: Wed, 24-Jun-87 11:23:16 EDT
Article-I.D.: ttidca.876
Posted: Wed Jun 24 11:23:16 1987
Date-Received: Sat, 27-Jun-87 04:46:41 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: redden@ttidca.UUCP (John Redden)
Organization: Citicorp/TTI, Santa Monica
Lines: 30
Keywords: .[.]* schema

I mentioned this as a proposal to our systems group and it generated a bit
controversy.

We run a number of flavors of Unix at TTI.  The most popular are BSD based
thus symbolic links (as opposed to hard links) are used often in our work.
Typically we use symbolic links to create a view (or schema if you will)
of one of our file systems to conform to another organizations (equally
valid) view of a file system.  This is particularly useful in a *heavily*
cross mounted NFS environment.

Very early in the game we discovered the semantic anomally of cd .. 
after a cd to a directory that was an object of a symbolic link.  This
is not a great problem as long as the semantic behavior is understood.  And
actually its easy to create an alternate "cd .." to the subject directory
of the symbolic link.

It would be nice to have a consistant syntax for cd:

cd .. takes you the real parent (as it currently does)
cd ... takes you to the first symbolic link
cd .... takes you to the second symbolic link

cd ..[.]*.  takes you to the ith symbolic link

One bad side effect (no considering implementation problems) would be
that ... and the like would become illegal file names.

Comments?

These are solely my opinions and not the people whom I work with.