Re: Catalogues for WWW

fritz@rodin.wustl.edu (Fritz Lehmann)

Mail folder: SRKB Mail
Next message: John F. Sowa: "Proposed standards for KIF and CGs"
Previous message: Timothy Finin: "GIS Ontology -- FGDC Approves Metadata Standard"

Date: Tue, 28 Jun 94 09:44:56 CDT
From: fritz@rodin.wustl.edu (Fritz Lehmann)
Message-id: <9406281444.AA08742@rodin.wustl.edu>
Newsgroups: comp.infosystems.www.misc
Subject: Re: Catalogues for WWW
References: <Cs38Az.6qv@latcs1.lat.oz.au>
Organization: Center for Optimization and Semantic Control, Washington University
Apparently-To: srkb@cs.umbc.edu
Sender: srkb-owner@cs.umbc.edu
Precedence: bulk

    There was a discussion of the subject catalogue for the
the whole of WWW about a month ago.  I hope your question re-starts 
that discussion.  I supported someone who had recommended
a sophisticated "faceted" classification system, rather than
a simple tree-structured hierarchy like Dewey Decimal.  In
the COLON system used in India, there are five "facets":
Personality, Matter, Energy, Space, and Time, and any work
is classified in its proper location in all five facets.

     There are several personal or otherwise specialized
catalogues out there.  Somehow integrating them with one
another is a great challenge -- requiring a catalogue of
catalogues.  In addition to being integrated with one
another, they should be integrated with outside systems
such as Dewey and various specialized Thesauri and also
Knowledge-Base classifications used in Artificial Intelligence.
No one system will serve everyone's purposes, but if
many systems are compatible then a user could benefit
from many of them acting in cooperation.

     Right now this is a hot research area in AI called
"knowledge sharing" in which the concept categories
in one system must somehow mapped mapped to those in 
another.  I don't know whether anyone is now attempting
such a system for WWW sources.   It should be recalled
that WWW sources will be extremely diverse, from an
encyclopaedia, to a political mailing list, to a TV
picture of a coffee-maker, to the collection of the Louvre.
Any classification system(s) will need to be very 
general and versatile at the top level at least.

     A subject-class-based system should be combined with
a good word-based indexing system.  That way, using key
words, you could quickly hop to a fairly relevant collection
of sources in a likely subject class. A further and more
challenging addition would be to classify potential 
purposes to be served by a document, and use that as
a facet.  [Let's see, PURPOSE("Cindy by the Coffee-Maker")=?]

                          Yours truly,   Fritz Lehmann
GRANDAI Software, 4282 Sandburg Way, Irvine, CA 92715, U.S.A.
Tel:(714)-733-0566  Fax:(714)-733-0506  fritz@rodin.wustl.edu
=============================================================