Below is an image that represents recursion of construction and destruction in the world of poly amino acids. This is the core of the operations of a cell. In the cell, the selection of a protein from a list is achieved through regulation of the DNA to mRNA conversion. The code generated modifies the universe and in the process produces itself as a final act. When I first modeled this process many years ago it gave me an interesting answer not pictured here ( because I am lazy and it is visually complex and difficult to represent in 2 dimensions ). When everybody has WebGL as a standard in their browser , I will be able to include 3 dimensional relationships as examples.
The statement that I wish to select one thing from a list and apply that thing is equivalent to writing a function that defines `int ArrayList[n];` `int thing;` `int select;` and then performs `select = rand()%n` `thing = ArrayList[select];` `return thing;`.
So, methods that a person might wish to program are items with a greater selection. This is where the computer shines. It has virtually perfect memory of what might be in a list of possible solutions. I can make IO functions that serve to allow the program to get information from audio, video , position information and other data that is "sense like" to define the choice matrix.
I think that I have the correct structure to allow this process to grow correctly and it is mainly the motive file, or the problem to be solved that will define what types of programs are generated and what data they are tested against. I can manipulate the process in any aspect, and I will likely place obscure methods in the method file to save some time and thrashing to determine the best algorithms for specific purposes in each context.
I will link an artificial imagination into the programs with generic interface. It is essentially an openGL interface that allows the construction of objects from lists of known objects to move, scale, shade, and rotate them. It incorporates a mechanism to take the frame buffer and compare it an image from any source. I already have many scripts that allow the conversion and scaling of images to the method that the program expects ( ImageMagick mostly ).
I feel there will be some interesting surprises to come from the generated code. If the method file is properly spec'd it would seem that it would always produce safe code. I will try and see what happens today. I plan to have a run which goes from description, to code generation from code design file , to compile ( chosen from language select ) , to splint, to gdb ( with XFakeKeys), to a virtual machine using Xfake also.
It seems that the program generator would be complex and messy, but I hope that at some point it will recurse to write that code also. If the process is coherent and performs in general the way I expect, I will then give it the equation that defines matter and allow it to generate the analogous sequence as a continuous stream of DNA, mRNA or proteins. It should be an easy conversion and I can then have a physicality to the computation matrix that is self generating and recursive in matter.
From there it is a matter of stretching the limits of its complexity and analyzing the results for useful methods. Since it would function from request to implementation, I should be able to define an objective and observe the solution.
0 comments:
Post a Comment