Where to start (was: Ontology for Dummies)

[Tony Linde]

Welcome Steve.

I don't want to spark an OntoWar, but could people (participants and
lurkers alike) say if there is any significant difference between a
topic map approach to ontologies and an OWL (or its ancestors') language
approach.

Are there other approaches of significance?

Cheers,
Tony. 

> -----Original Message-----
> From: owner-semantics at us-vo.org 
> [mailto:owner-semantics at us-vo.org] On Behalf Of Steve Pepper
> Sent: 02 October 2002 12:47
> To: semantics at us-vo.org
> Subject: Where to start (was: Ontology for Dummies)
> 
> 
> Thanks to Ashish Mahabal for bringing this forum to my attention.
> 
> Tony Linde wrote (about "Ontology for Dummies"):
> >I know, there isn't such a book, so what can we (the 
> non-experts) have 
> >to get us up to speed. Can we come up with suggested texts 
> that would 
> >take a non-expert from basics through to understanding why 
> and how to 
> >encode a subject domain ontology and why and how to use it in one or 
> >more applications.
> 
> Here are a couple of papers that I've found useful:
> 
> * Deborah McGuinness' chapter in the forthcoming "Spinning 
> the Semantic
>    Web", called "Ontologies Come of Age":
> 
>    
> http://www.ksl.stanford.edu/people/dlm/papers/ontologies-come-
> of-age-mit-press-(with-citation).htm
> 
> * "Ontology Development 101: A Guide to Creating Your First Ontology"
>    by Natalya F. Noy  and Deborah L. McGuinness
> 
>    
> http://protege.stanford.edu/publications/ontology_development/
> ontology101-noy-mcguinness.html
> 
> I don't know of anything that discusses ontology design with 
> specific reference to topic maps (apart from internal course 
> material here at Ontopia), but the principles are pretty 
> simple. If you've read "The TAO of Topic Maps" [1] you know 
> that the core concepts in topic maps are Topics, 
> Associations, and Occurrences (hence "TAO"). Each of these 
> can be classified by type (topic types, association types, 
> occurrence types)
> - and types are themselves also topics...
> 
> Thus "Puccini" is a topic of type "composer", and "composer" 
> is also a topic. The association (relationship) between the 
> topic "Puccini" and the topic "Tosca" (which is of type 
> "opera") is of type "composed by", which is also a topic. An 
> information resource containing the libretto of Tosca would 
> be an occurrence (of type "libretto") of the topic "Tosca".
> 
> In this example, the (typing) topics "composer", "opera", 
> "composed by" and "libretto" constitute the ontology -- or 
> what some people term the "upper ontology". The "lower 
> ontology" consists of the individuals, "Puccini", "Tosca" etc.
> 
> (Apologies for not being able to give you examples from your 
> own domain: The only "stars" I am familiar with are ones like 
> Maria Callas and Pavarotti :-)
> 
> It seems to me (correct me if I am wrong!) that what this 
> initiative needs is
> 
> (1) agreement, as far as possible, on a common upper ontology and how
>      to *identify* the classes it consists of
> (2) agreement on how to identify a useful common subset of known
>      individuals, but without (necessarily) assigning those 
> individuals
>      to classes (since this might be controversial).
> 
> As soon as you have that, everyone can start making their own 
> assertions about any of the subjects in question in such a 
> way that knowledge about them can be aggregated, shared and reused.
> 
> I would advise you to apply the KISS principle. You will 
> experience an almost irresistible temptation to bite off more 
> than you can chew, to "overmodel", to get embroiled in 
> endless discussions about almost unresolvable details. Resist 
> that temptation! Agree on the minimal common subset that is 
> actually useful and extend it later as necessary. In 
> particular, don't bother at this stage to try and express 
> complex constraints on the classes you identify for the upper 
> ontology.
> 
> Finally, I would advise you to look seriously at the 
> Published Subject initiatives taking place under the auspices 
> of OASIS:
> 
    http://www.oasis-open.org/committees/tm-pubsubj/

You will achieve a lot in a short time if you concentrate on defining
published subjects for your upper and lower ontologies. Among other
things, it will make the choice of knowledge representation formalism
less important, since published subjects allow multiple formalisms to
interoperate. If anyone is interested in pursuing such a course, say so
(in a new thread) and Bernard Vatant and I can explain how to go about
it.

I wish you luck in your endeavours and will continue to lurk here in
case I can be of use.

Steve

[1] http://www.ontopia.net/topicmaps/materials/tao.html

--
Steve Pepper, Chief Executive Officer <pepper at ontopia.net> Convenor,
ISO/IEC JTC1/SC34/WG3  Editor, XTM (XML Topic Maps) Ontopia AS, Waldemar
Thranes gt. 98, N-0175 Oslo, Norway. http://www.ontopia.net/ phone:
+47-23233080 GSM: +47-90827246