Path: utzoo!yunexus!ists!jarvis.csri.toronto.edu!mailrus!shadooby!samsung!gem.mps.ohio-state.edu!tut.cis.ohio-state.edu!ucbvax!hplabs!hplabsz!sartin
From: sartin@hplabsz.HPL.HP.COM (Rob Sartin)
Newsgroups: news.groups
Subject: Re: The disservice of pushing for sci.aquaria
Message-ID: <4362@hplabsz.HPL.HP.COM>
Date: 15 Nov 89 00:06:35 GMT
Article-I.D.: hplabsz.4362
References: <6951@ficc.uu.net> <11414@cbnews.ATT.COM>
Reply-To: sartin@hplabs.hp.com (Rob Sartin)
Organization: Hewlett-Packard, Software Technology Lab
Lines: 51

>Richard, and the readers of .aquaria, had a very good, if selfish, reason
>to want their group in the sci. heirarchy.

1. Given the number of people who got all worked up over the arguments
   and the system administrators who have announced that they won't
   carry the group, I'm beginning to wonder if the propagation will be
   better.  I suspect rec.aquaria might have had better propagation.

2. Wanting better propagation is not a valid reason for placing a group
   in a certain hierarchy.  By this argument all groups should go under
   whatever top level gets the best propagation.  Hopefully that is not
   what will happen.

[Hit 'n' - or whatever - now to skip my rantings on hierarchies]
An aside on hierarchies (general and news hierarchies in particular):


The use of hierarchies to model various natural (and unnatural)
structures seems prevalent.  We find it in biology, library science,
Unix file system, our educational system (Cornell, at least, has
University->colleges->schools(departments)->specialties when you look at
Engineering as the college, different hierarchies for different
colleges), and the newsgroup structure among others.  We also saw it
(and thankfully discarded it) as a model for inheritance in C++.

Hierarchies are often insufficient to model the structure of the
information, but are attractive because of the simplifications they
offer.  We have seen, in the alt/sci/rec.aquaria events, that the notion
of hierarchical structuring of newsgroups is arbitrary.  Many groups out
there could well be in multiple top level hierarchies (it's "a hobby
that has a strong technical content" or "a science that happens to be my
hobby" applies to many groups other than .aquaria).

In some areas we see various modifications to systems to work around the
fact that hierarchies don't model reality well.  I don't know biology,
but I do know that there are arguments and reclassifications continually
occurring due to our forced notion of hierarchy.  Library science
(again, to the best of my understanding) sticks with a "major
classification" and uses cross referencing and multiple listings
extensively.  The Unix filesystem has symbolic links and hard links.  In
school we had "multi-discipline" groups and design-it-yourself degrees.
C++, in its evolution from 1.2 to 2.0, added multiple inheritance which
(at least) made inheritance a DAG instead of a tree.

The news software and users needs a method for jumping out of the system
(of hierarchies) when it appears to be inadequate.  Instead we have
pointless flame wars.

Rob Sartin			internet: sartin@hplabs.hp.com
Software Technology Lab 	uucp    : hplabs!sartin
Hewlett-Packard			voice	: (415) 857-7592