[Top] [All Lists]

Re: [ontolog-forum] Sharing and IntegratingOntologies

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
From: "John F. Sowa" <sowa@xxxxxxxxxxx>
Date: Wed, 14 Apr 2010 12:06:42 -0400
Message-id: <4BC5E812.9040902@xxxxxxxxxxx>
The following note addresses some issues that are common to both
the SIO project and the OMG.    (01)

John Sowa    (02)

-------- Original Message --------
Subject: Re: Are there shared concepts?
Date: Wed, 14 Apr 2010 11:20:25 -0400
From: John F. Sowa
To: AESIG <architecture-ecosystem@xxxxxxx>    (03)

Jim,    (04)

JA>  However, it seems possible that you are describing requirements
> for SIO, not necessarily AE-SIG.    (05)

The requirements I was describing are for a common subset of both.    (06)

JA> That is, could AE-SIG pick an approach (say for example RDFS++
> or OWL-- or something), and then exploit SIO to link that in with
> other approaches to FL?    (07)

Yes.  In fact, OWL is one of many different logics that is covered
by the tool set that John Bateman described in his slides.    (08)

JA> I want to keep AE-SIG focused on enabling the integration of
> modeling languages, leveraging what others have done and are doing
> to solve the overall knowledge sharing problem.    (09)

That is an excellent focus.  But I would just make the observation
that a modeling language is essentially a logic with a built-in
ontology.  The SIO project processes and relates both logics and
ontologies and any combinations of them.    (010)

For example, the underlying logic of STEP has the expressive power
of first-order logic.  Therefore, any tools that relate STEP models
to other modeling languages must be capable of processing and
relating full FOL.  Just that requirement by itself goes beyond
what RDFS or OWL can do.    (011)

I think that we should learn from the lessons of other projects
that started with "simple" subsets:    (012)

  1. UML began as a merger of the diagrams by the Three Amigos.
     But most models also have constraints that go beyond what can
     be expressed in those diagrams.  Therefore, UML also includes
     OCL, which has the full expressive power of first-order logic.    (013)

  2. R. V. Guha had been the associate director of the Cyc project,
     which had a very rich and expressive logic-based language.
     But he thought it was too complex for most people, so he
     designed a very simple triple-based notation.  Then he teamed
     up with Tim Bray to represent it in XML.  The result was RDF.    (014)

  3. As time went on, the W3C discovered that they needed to state
     constraints on those triples, so they adapted an earlier
     language called DAML to XML notation and called it OWL.    (015)

  4. But OWL started to diverge into multiple versions with different
     levels of expressive power, each with a slightly different
     semantics.    (016)

  5. Then many people decided that they didn't need all the power
     of OWL for their application, so they designed RDFS.    (017)

  6. Then other people wanted query languages, rule languages,
     and other variations.  So they developed different notations
     that could all process RDF triples in various ways.    (018)

  7. But other people decided that they wanted to extend the
     triples to quads, and all the people who had designed logics
     that could handle triples had to extend them to handle quads.    (019)

There is nothing wrong with having multiple notations and diagrams
specialized for different purposes.  I would encourage that trend.    (020)

But it is essential to ensure that the semantics of each special
case is compatible with all the others.  The simplest way to do that
is to specify the common semantics of *all* of them right at the
beginning.  Then each of the specialized languages must be defined
by its mapping to and from the common semantics.    (021)

Compare that to first-order logic.  In 1879, Gottlob Frege
developed the first complete system of FOL in a notation that
nobody else ever adopted.  In 1880 and 1885, C. S. Peirce
independently developed a notation that evolved into the
modern version of predicate calculus.    (022)

The important point is that those notations were *independently*
developed, but their semantics was *identical*.  Furthermore,
that semantics is still the common core of every version of
full FOL that has been developed over the past 130 years.    (023)

How many of those modeling languages would you expect to remain
unchanged for 130 years?  For 20 years?  For 5 years?    (024)

John    (025)

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
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx    (026)

<Prev in Thread] Current Thread [Next in Thread>
  • Re: [ontolog-forum] Sharing and IntegratingOntologies, John F. Sowa <=