The definition of the engine is nothing more than a software version of an infinity bit processor. With the aid of SQL, the key word size is BLOB, go look it up. The things all the variable operators, a small set. It manages variables o a nested semantic tree, just like the compiler. The engine itself is fast, it uses sql scans over opcodes and data naturally. It provides an excellent sibling / sequential operator properties. It has a function call
funtion(v1,v2,v3) -> convolve((v1,v2,v3),function) // where both are in G object code format.
So, hey, take Java or C#, and compile to the nested record object model. all your objects are managed under SQL, what can be better? Everything, code, data, the same form a G, linked a world of Gs.
The open source ontology engine, the only way to go. I am thinking, I will write whole thing in 200 lines of straight code, one c file with header.
Here, lets see if we have theory here.
Take all the we pages of releance, and get just the keywords, reduce the text by 30% with removal of small connectors, remove punctuation, do dictionary look up. If we call that the quant at the lowest level, the bigbest thing, and build the minimal directed graph ontology over the whole set of these, then our ontology stack is: WidthLog(Width) in total keywords. Minimum spanning tree problem?
Who cares, this entire ontology is built upon SQL, ready for rapid scans and large sizes. So this processor zooms through masses of ontology zeroing in on its end points, Gout. Inference occurs by detecting mutual entropy among directed paths, easy to find. This thing is going to think about what all the huge mass of text means..
What i the best way to write a compiler for G? Hmm, hrow the source file into SQL, then convovle it with G.compilers.c:
MyGobject = convolve('//My special c source code comment; main() {}',G.compiler.c)
I dunno, In the land of G graphs, there are subgraphs that parse and assemble text, into G graphs. Everything ultimatelf is a table store with the minimal three components of a G element.
No comments:
Post a Comment