May 10, 2005
ß original RDF file final reduction to Orbs à
ABC ontology:
List of active issues related to Orb – RDF transforms
1.0: A review of the effort made is sampled in the many developer’s e-forums where people are struggling with the difficulty of understanding the behaviors of various ontology editing or inference software like Jena and Protégé.
1.1: The notational system for Orbs will have the property of having a precise mapping between Orb editing and visualization software and the notation. Thus one ought to be able to always predict the behavior of the software using simple notation and set theory.
2.0: The RDF standard focuses on two things that we do not need.
2.1: One is URI pointing to the definition of the meaning of something. We hold that, in principle, meaning is rarely well defined by some distance information.
2.1.1: If this information is needed, then some means other than long and unreadable URI can make the connection to this distant resource and bring that information locally.
2.1.1.1: The use of URIs when they make sense is reasonable. However, in many cases the local enterprise knows what it wants the meaning to be of the concept representations in “it’s” ontology.
2.1.1.2: The variation of meaning form one community to another is so critical that local definition of meaning makes sense from a “communities of practice” point of view.
2.1.2: In fact, the convolution over a set of “locally defined” information is the key to developing global analysis using the Orb convolutions (see SLIP tutorials and theorems)
2.1.3: The local verse global issue has deep roots in the theory of language, and those other natural sciences that we feel supports the Second School intellectual positions.
2.2: A type of acquisition of information via an inheritance relationship can be obtained. Why not? But the degree to which the tail is waging the dog has to be realized. Inference by humans visualizing Orbs is a much more powerful means to management and project knowledge.
3.0: The Intellidimension RDF repository has an internal RDF encoding that uses four column databases, the fourth column is a “context” column. This same structure can be used to store the Orb triple.
4.0: The visualization of the Orb triple is easier, less code. This has to be demonstrated, and there are still some issues.
5.0: W3C recommendations on URI is intended to liberate each of us from having to have “information” on our computer. This is true, and in many cases this will be found to be useful.
5.1: But if I want a model of the knowledge about something specific, like what is in a container that I, as a Custom’s official, is about to open and inspect, I want this information to be localized on my machine and have labels that easily evoke in me the mental awareness that I need to understand the container.
5.2: From the neurocognitive viewpoint on human experience of knowledge, we know that a great deal of personalization occurs as part of the process of localizing information.
5.2.1: This personalization provides tacit knowledge to the experienced knowledge – and this is value added that is sometimes lost if the meaning of things is defined someplace else and delivered to the human via a URI. (Right?)
5.2.2: Personalization also adds viewpoint, and so two people sometimes experience the exact same information, and yet the result is quite different in the interpretations.