I have been spending some time reading Synergetics by R. Buckminster Fuller and it has me thinking about the structures of dimensional databases.
My work as a data modeler has exposed me to many ontologies. Every data model you create is a self contained ontological framework. And for a long time I did not think about what those frameworks had in common. Every project was unique.
In the course of my recreational reading I learned about Pattern Language from Christopher Alexander, saw its adaptation in object oriented design and over time learned how patterns could exist in relational models to address particular applications most effectively. It became obvious that there were optimal patterns for competitive advantage and the seemingly limitless array of patterns for databases revealed optimals and families of optimals.
In studying dimensional data model design, based on Ralph Kimball’s work, I found myself also looking for patterns. The only case of patterning I found at first were called the Basic Interrogatives. Further research corroborated this pattern by way of the Zachman Framework. There was a six dimensional pattern. But at that point I stalled.
Several years later I began to read the works of R. Buckminster Fuller and a new world of possibility opened up to me. What Fuller led me to do was visualize dimensional models not in the traditional two dimensional star or snowflake, but as polyhedra with the vertices representing each of the dimensions and the fact being the center of gravity of the solid.
The fit of the basic interrogatives into an octahedral structure lead me to wonder about polyhedra further. Fuller’s work lead me to think about the platonic solids and other families of polyhedra. I began to see patterns in dimensional data models similar to the six interrogatives.
My search for generic dimensional models lead me to the world of library science. I began to look at classification and the work of S. Raganathan came to light. Colon Classification presented another generic structure and this pointed to Bliss Classification 2. Colon Classification presented four dimensions, the Basic Interrogatives presented six dimensions and BC2 presented twelve dimensions. These generic classification systems correlated with the tetrahedron, octahedron and icosahedron–the triangulated Platonic solids.
This is as far as I have come. I am now looking for more dimensional models to examine for patterns that correspond to polyhedra.