Sunday, May 22, 2005
Notational Standards for Ontology referential
bases (Orbs)
One of the things that Protege does allow one to do is to develop a simple ontological model with concepts, sub concepts (having certain types of inheritance between concept and subconcept), concept properties have domains and ranges which means that properties are often used to define a relationship between two concepts.
Given the use of the Protégé editor, one can create a RDF file. This file will list the concepts and subconcepts first and then list properties, as in the rdf file I attach here.
The Orb ontology specification appears to us to be complete, without the complications that Protégé, OWL and RDF have.
For example, differential ontology framework (DOF) assumes a set of ontological model element having the Orb form:
< a, r, b >
where a and b are “locations” in a semantic or syntactic space and r is some relationship between the two locations.
An attribute, for the RDF standard, is often something that relates a property or concept to some computer data type, like the "integer data type". The Orb notation would represent an attribute simply as:
< something , is a attribute of, something else >
The practical difficulty with RDF and OWL is that the "natural concept" of type is needed in the pure ontological model and great confusion occurs because the end-user of an ontological model might naturally think about an attribute as being something like a property. The RDF standard straightjackets the common user into thinking about ontological modeling in the way that the W3C has determined. This authority has problems simply because humans are often uncomfortable with absolute authority.
In addition the authority standardization issue, we have a deeper problem.
Humans are attempting to learn about ontological modeling as a means to understand complex system in the real world. Our efforts at developing the knowledge sciences are made obscure by technical demands related to the RDF and OWL standards. These technical demands may in fact be incorrect if evaluated from a complex systems point of view. The key example has to do with establishing standard meaning for all concepts.
The notion of occurrence is important in the Topic Map standard. This can be addressed by “saying” in Orb language,
< topic q , occurs at, location r >
so that topic q is a location in meaning space and so is “location r”. Humans can understand this language.
Machines can compute with the vocabulary used in the Orb triples. This ability is completely open. Each ASCII string can be uniquely identified and each ASCII strong can be associated with a hash table were the “container” associated with the hash element has metadata used to mark issues with the interpretation and context of the string.
Using the encoding methods described in the Orb Notational Paper we have three lists for each set of Orb triples.
( ACSII string | metadata )
These three lists can be encoded into the Intellidimension four column table and with the metadata being placed into the first of the four columns, reserved by Intellidimension for “context”.