[Top] [All Lists]

Re: [ontolog-forum] Curation view of Ontologies

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Sun, 22 Jun 2008 12:07:58 -0400
Message-id: <485E78DE.5010805@xxxxxxxxxxx>
Ravi,    (01)

Ranganathan's colon classification system is of fundamental importance.
I would strongly urge everybody with any interest in ontology to get
a basic acquaintance with the system.  For starters, see the appendix
to the following article about Ranganathan:    (02)

    http://www.garfield.library.upenn.edu/essays/v7p045y1984.pdf    (03)

For Ranganathan's own outline and examples, see    (04)

    http://www.iskoi.org/doc/colon.htm    (05)

A few of his detailed examples:    (06)

    2;45:6      circulation of newspapers
    B331,1,2:1  numerical solution of ordinary linear differential
                 second order equations
    B952:72     orbits of comets
    C3:11;5     velocity of sound in water
    D5153,8:5   specification for the brake of railway carriage
    H191:16     genesis of diamond
    H4115.53    volcanoes of France
    H7.438      mineral resources of Burma
    H7113:16.8  origins of the copper deposits of Australia
    KZ351:421:5 prevention of tuberculosis in poultry
    NQ44,3;3:6  Indian distemper fresco painting of landscape
    O-,2J64,51:g(S)  psychology of Hamlet
    P111,9D56175,1:1  pronunciation in Yorkshire dialect
    T15:3(B2),1 audio-visual method of teaching algebra in elementary
    W691:58(Q)  freedom of conscience in communist state
    X:53.440r56'N5  the influence of British tariff on Indian tariff
                 brought up to 1950's    (07)

I would challenge anyone who prefers some other system to give any
evidence that their own system is more general, more principled,
or more systematic than Ranganathan's.  I wouldn't claim that R's
system is perfect, but I would say that anyone who hasn't studied
it is in serious danger of reinventing a square wheel.    (08)

RS> I have tried many times to talk around the factors that lead
 > Ranganathan (*reference19 in the Curation PDF) toward his colon
 > classification (similarly many multitudes of such concepts in
 > Sanskrit texts) but have not yet had many responses from our
 > participants and I thought that concepts from other ways of
 > thinking about Reasoning and Logic were not very much encouraged!    (09)

First of all, colon classification is not one of the "other ways
of thinking".  It is based on a fundamental principle that must
be taught to every newbie and every so-called expert in ontology:    (010)

    Single inheritance is hopelessly inadequate for classification.    (011)

Librarians give several reasons why colon classification has not
been more widely used:    (012)

  1. The detailed codes can become too long to fit on the back of
     a book spine.    (013)

  2. The methods for deriving a code are hard to learn.    (014)

  3. Tree structures, such as the Dewey Decimal System or the Library
     of Congress codes, can be mapped to a linear string of cards in
     a catalog or bookshelves in a library.  But multiple-inheritance
     systems have no linearization that preserves the ordering.    (015)

The first point can be met by the obvious way of truncating the
code in order to fit it on the book spine and pasting the full code
inside the front cover.    (016)

The second point can be met in the usual way:  most librarians
today just look up a book in a directory to see how the "experts"
have classified it.  That is just as easy to do with R's system
as with Dewey's.    (017)

But the third point was a major stumbling block in the pre-computer
days.  The solution to that problem is to replace the linear card
catalogs with computers (and that has been done for every large
library and most small libraries in the world).  Then the computer
can map the multiple-inheritance system to any coordinate system
used for locating books.    (018)

RS> Of course, it would require thinking outside the notion that all
 > we know about these subjects came largely from Greek Civilization...    (019)

Not true.  I have often recommended that people learn something
about the history of world civilizations:    (020)

  (a) The Egyptian, Sumerian-Babylonian, Minoan, Greek, Persian,
      Indian, and Chinese civilizations were trading ideas, religions,
      mathematics, inventions, spices, silk, seeds, livestock, and
      armies along the silk road since about 1500 BC.    (021)

  (b) As examples of the trade of ideas, all the alphabets of the
      world (except the relatively recent Korean alphabet) are based
      on the Phoenician alphabet, which is based on a simplified
      version of Egyptian hieroglyphics.  In the opposite direction,
      the Arabs transmitted the Indian invention of "Arabic numerals"
      to Europe.  And Leibniz invented binary arithmetic by studying
      the encoding of the 64 hexagrams of the Chinese I Ching.    (022)

  (c) Aristotle's system of syllogisms fully supports multiple
      inheritance.  Leibniz exploited that fact for his own lattice-
      based system (Universal Characteristic), which uses the same
      principle as Ranganathan's facets.    (023)

Summary:  Every single-inheritance ontology is obsolete.  Any
single-inheritance system that is currently in use should be
replaced or updated to a multiple-inheritance system in order
to make it suitable for further development and extension.    (024)

John    (025)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Subscribe/Config: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (026)

<Prev in Thread] Current Thread [Next in Thread>