Saturday, December 17, 2011

Handling binary JSON and graphs

We can define the binary JSON schema be be the set of triplets inside the folloring:

BinaryJSON:(Set of triplets inside this operator)

The documentation should be easy. I use the schema here because the operator delivers proper triples and does not suggest opening keywords. (Attributes want to open them up/ schema wants to coral them in)

Anyway, what are Binary JSON triplets? The same triplet as always except they, Binary JSON committee members, define the semantics and syntax of pointer, and they get a whole byte of opcode in the link field, and they decide when to open up keyword. And they, being fairly smart, will by default, define the syntax we need to get variable expressions and arithmetic,(straight from triplets) in our graph convolutions. The underlying sql can deliver a whole set of BJSON triplets to the BJSON method. This machine should deliver a set of triplets in a millisecond. This is win/win.

How might that work?
Well, the BJSON format could leave the machine with an object structure, right there, on a descending branch of nested store. Then any operations later down the chain can reference them. We could get little methods. We could use bjson fundamental types as a basis for built in patterns, fill in the form data collection. Build a bit of a pattern matching events from bjson encoded triplets. I warm to the idea.

No comments: