Tuesday, February 21, 2006
Generative Methodology Glass Bead Games
On the limits of the OWL standard à [184]
Reading material [1]
Reading material [2]
Reading material [3]
Summary of the discussion up to this point à [186]
On Edelman’s notion of degeneracy, and n-articulate ontology
Related discussion on ontologyMapping thread à [17]
Question to
the Protégé forum
I was wondering where one draws the boundary between ontology and
knowledge-base. In the database world the demarcation between schema and
data is clear. However, the situation seems less clear in
the knowledge engineering world. My current thinking is that the
boundary (if any) is related to how the knowledge is being used and not so much
about specific artifacts. That is, an ontology represents a (fairly)
static interpretation about some domain, whereas a knowledge-base system may be
comprised of an ontology and makes inferences at runtime to reason about the
world(?). Enlightenment is appreciated.
Additional comment on the OWL-Protégé forum à [202]
Cheers,
Cameron Ross.
I am also dealing with where this
distinction is. In a, so called, knowledge base like one
constructed using the OWL standard; the data is (IN FACT) separated from the
organizational model.
However, in the presentation of OWL by the core OWL community; the issue of fidelity of the model is confused by an assertion that descriptive logic has an inherent validity because it is based on logic. The various flavors of description logic are categorical, in nature, and allow an additional "assertion" of categorical information, such as X is a class and Y is a property of that class.
The set of all assertions are called "axioms". But there are really two types of axioms, those asserted by the standard (at what ever flavor) and those asserted by the designer of a specific application.
The set of assertions in an application then provides an organizational schema to the data. In this way there is a perfect analogy between "data" in a relational database and "data" in an OWL informational model.
The wonderful and proper discussions about a need to separate the data layer and the program/logic layer in the relational and object oriented paradigms are as valid in OWL as it was in the "three-tier" and "n-tier" information architecture discussions.
But this issue, or separation, is made complex by the claim that OWL is based on an open world assumption whereas RDBMS and Object Oriented Design is based on a closed world assumption. I prefer to call the OWL paradigm "partially open" because individual (data) can not on their own "give rise to" new categories which are then added to the OWL schema (the set of assertions, also known as the set of axioms). This ability to “induct” new or modified categories is not part of the OWL paradigm, but is part of natural ontology.
The RDBMS restrictions that are placed on adding data is opened up due to the nature of XML. Since OWL uses XML, there is additional flexibility in regards to when new data is accepted. (Data can be incomplete or having elements that is not modeled by the OWL information model).
"Ontology", the word, in its historical setting needs to mean "reality". This meaning is over written by the W3C presentation of its standard. So one form of confusion comes if one has the background to challenge the W3C presentation.
This communication is a "dissenting" voice in regards to the voices normally expressed in this forum, and is protected speech. (This is just to make things clear.) . It is often that I make technical mistakes, and hope that when I do there will be a kind pointing out of this technical mistake.
Paul S Prueitt
The Taos Knowledge Institute
Taos, New Mexico