Subject: Re: DEC uses NetBSD
To: Scott Reynolds <scottr@plexus.com>
From: Jason Thorpe <thorpej@nas.nasa.gov>
List: current-users
Date: 03/25/1997 12:24:54
On Tue, 25 Mar 1997 14:05:09 -0600 (CST) 
 Scott Reynolds <scottr@Plexus.COM> wrote:

 >     contrib/  large packages contributed to Berkeley by outside parties
 > 
 > Even if you s/Berkeley/NetBSD/, how does it make sense to put ported
 > software from a given third party under /usr/contrib, especially given
 > that it probably wasn't contributed?  Is /usr/contrib being suggested
 > simply because it's the best fit in hier(7), even if it gives an
 > inaccurate impression?  (This is a philosophical question, and I will not
 > argue the point to death; I just hope it's not ignored.)

...given the definition of "contrib/" in hier(7), I would actually prefer
that it went away completely.  Software that is contributed to the
NetBSD Operating System is integrated directly into the tree.

 > In any event, I would have to agree with the folk who think /usr/local is
 > a bad place to put things.  I'd much rather see a /usr/pkg or /usr/opt,
 > which among other things gives the admin the flexibility to install common
 > packages on an NFS server, without stomping on /usr/local in the process.
 > One could always put symlinks under /usr/local to point to the packages
 > that lived elsewhere...

Okay, so I suppose, for the purpose of binary backages, create a new
directory?  Call it "/usr/opt"?  I get the vibe that Too Many people
just don't like /usr/local.  Of course, for packages like XFree86,
they may install into /usr/opt, but would still require a symlink
like:

	/usr/X11R6 -> /usr/opt/X11R6

...or something.

Jason R. Thorpe                                       thorpej@nas.nasa.gov
NASA Ames Research Center                               Home: 408.866.1912
NAS: M/S 258-6                                          Work: 415.604.0935
Moffett Field, CA 94035                                Pager: 415.428.6939