Path: utzoo!attcan!uunet!aplcen!uakari.primate.wisc.edu!sdd.hp.com!ucsd!ucbvax!hplabs!hpcc01!hpcuhb!hpindda!kmont
From: kmont@hpindda.HP.COM (Kevin Montgomery)
Newsgroups: comp.cog-eng
Subject: Re: Formalizing location and distance in file retrieval systems
Message-ID: <3500013@hpindda.HP.COM>
Date: 13 Aug 90 22:22:34 GMT
References: <1801@nvuxr.UUCP>
Organization: Bill and Dave's
Lines: 24

/ hpindda:comp.cog-eng / jim@se-sd.SanDiego.NCR.COM (Jim Ruehlin, Cognitologist domesticus) / 10:33 am  Aug 10, 1990 /
> This could be a subjective issue, depending on the user.  To some, distance
> could mean conceptual distance, where a deeply nested subdirectory is
> considered closer to it's ancestor than it's ancestor is to its sister.
> - Jim Ruehlin

I think this is more to the point- it's the perceived "distance" that
really matters, since the filesystem is just an abstraction of an 
information storage device.   

I think the correct term for the Unix filesystem would be an undirected 
multigraph. Undirected, because I can go up ("../") or down ("/etc"), 
and multigraph (as opposed to a simple graph), because I can link a 
file/directory to itself.  A tree has 1 path to each node.  A simple 
graph has >1 path to the same node, but no node points to itself.  A 
multigraph releases this restriction.

(God, I hope this is all right, or I *REALLY* didn't get anything out
of that old Discrete Math class!  :)

							kevin

ps: alot of disjoint work is being done on naming at the moment- it's
    a "global" problem  (yes, pun intended).