Information handling in high dimensions

I am studying how to apply new techniques in information handling. I think that te methods that exist are weak and start with certainty and hope to extend certainty into new areas and I have found that I have better luck by starting with uncertainty and focusing inward. This video at Google Talks is about low dimension manifolds and though it might be considered complex, I also see it as too simplistic to resolve the problems. I have found that by mixing technology and concepts it can be easier to come to a result that is simply a weighted answer, not complete or decisive, but just adequate. I use Fourier analysis in images with generative techniques. For recognition of dimensional relationships it serves well. The base object is the vector distance relationships and whatever projected manifold has a degree of certainty associated with it based on the use cases that it encounters. Put simply, in context it is possible to resolve images and relationships, but without a limiting context, the information is ambiguous. It does relate to the puzzle and decision trees and high space dimensions. I continue to collect the data portals that can be analyzed and here is one interesting place to get add-ons for `celestia` to observe the entire universe or even imaginary worlds of science fiction.

The process of integration of all this technology from WebGL, `java`, OpenGL, models, `makehuman`, `blender`, Python, Javascript, game theory, trees, DNA HOX, genotyping, protein pathway analysis, physics of the atom and many more, to antfarmgl is proceeding, but the coherence is the key. Once the coherent framework becomes evident it is simply a matter of getting all the pieces in the right place to be usable.

I am sticking with C as the base code and interfacing to shell, java, and other necessary modules. I need to understand the interface as well as the material that moves through it. That Java code tutorial was really scary as the innards of JLS need to be used with a knowledge of what the underlying method actually implements. With C, I already do that. When I create a new module for myself I go down to the ASM language level and model it for myself. I know what the hardware does and there are no surprises on integer add, or floating point operation. I suppose the state of the FPU across several concurrent tasks can be a bit spooky, but I have dealt with that also.