[Top] [All Lists]

[ontolog-forum] Good ontologies without good tools are useless

To: ontolog-forum@xxxxxxxxxxxxxxxx
From: John F Sowa <sowa@xxxxxxxxxxx>
Date: Tue, 18 Feb 2014 09:11:04 -0500
Message-id: <530369F8.5030508@xxxxxxxxxxx>
John M, Henson, et al.,    (01)

Ontolog Forum is not a philosophy debating group.  All (or
at least most) of us are participating as *engineers*, whose
goal is to build something useful.  Without appropriate tools,
no ontology -- no matter how good -- can help us.    (02)

I chose the above subject line to tie together several threads
that have touched on these issues without clearly stating them.    (03)

In the thread "What is the difference re Data Dictionary,
Ontology, and Vocabulary?"  I cited some slides from an
introduction to ontology, starting with Aristotle:    (04)

    http://www.jfsowa.com/talks/aristo.pdf    (05)

John McClure responded:
> Tremendously useful/insightful slides, thank you John...
> I wish there were ontologies which track to these (age-old)
> classifications, encodings of core conceptualization(s),
> the same ones the knowledge industry is stumbling towards
> recreating (in an intensely inefficient manner).    (06)

I often said that Aristotle's ten categories are as good or
better than most of the upper-level ontologies that have been
proposed for applied ontology.  But all that philosophical
thinking and writing, by itself, didn't produce anything that
people who have a day job found useful to do their work.    (07)

But note the slides toward the end:  Slide 27 on Ramon Lull's
rotating circles, the encodings and calculating machines by
Leibniz (slides 28-30), and John Venn's diagrams as tools for
streamlining Aristotle's syllogisms (slides 31-34).    (08)

Those are tools that make ontologies more usable.  Modern computers
enable us to make even better tools.  In Ontolog Forum, I mentioned
Cyc almost as often as I mention Aristotle.  But as we know, Cyc is
still a small niche system.    (09)

My guess about why Cyc has not become as successful as Lenat & Co. had
hoped is that they focused on tools for building *ontologies*, not on
tools for building *applications*.  We need both.    (010)

Henson G in the thread "Where UML models fit" wrote:
> UML and SysML models are commonly used in engineering for both
> representation, i.e., description of some part of the physical world
> and as specifications for things to be built, as well as for reasoning.    (011)

Yes.  As Henson has shown, UML diagrams can be specified as a version
of logic (in fUML).  But even without a formal spec, they have been
integrated with widely used tools for building applications.    (012)

Medical informatics is a major application for Cyc.  The Cleveland
Clinic is their premiere example.  One of the people there said that
the biggest problem with Cyc is the "huge disconnect" between the
tools for Cyc and the tools for their software applications.    (013)

Just imagine how things might have developed if Guha -- who had been
the associate director of Cyc and later the chief designer of RDF --
had adopted UML in the mid 1990s.    (014)

Guha said that the reason why he designed RDF is that CycL was too
difficult for most users.  I agree with him.  But software developers
in the 1990s were happily using UML diagrams.  The UML notations,
tools, and methodologies can support    (015)

  1. Type hierarchies (the backbone of every ontology),    (016)

  2. ER diagrams (logical signatures and cardinality constraints),    (017)

  3. Activity diagrams (links between the logic and the procedures),    (018)

  4. Controlled natural languages (more readable than OCL for stating
     rules and constraints that go beyond #1, #2, and #3).    (019)

I admit that I'm making these criticisms with 20-20 hindsight.
In fact, I blame myself even more than I blame Guha or anybody else
-- because in the mid 1990s I was participating in ISO working
groups on standards for a conceptual schema (i.e., ontology).    (020)

At that time, I was proposing logic as the foundation.  If I had
proposed UML and points #1, #2, #3, #4 defined in FOL, an ISO standard
for ontology (AKA conceptual schema) might be mainstream IT today.    (021)

John    (022)

Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: 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 join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (023)

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