Doing sophisticated join

To start join, you compile four files and run them. 

 The join machine comes with it own personal 'database', the offline methods, like setting up dlls on the run. This 'database' is the dll manager, it thinks the world is a directed graph of dlls It responds to key words, matches up the necessary graph of reqjuired Dlls, and installing; all on command of lazy json script.. So running is simple, just get your favorite editor and write up a lazy json configure files. Add anything, from a windows interface, to a clerk looking through the paper file cabinet. As long as you thing can look like a finite directed graph.

Set up multi level and massive text search and classify.  Do much better than giggles.  Enhance it to keep a list f asvsilsablr attachments, pick gand drop them on he fly.  Take existing, and complex text analyzers, let their ootput look like a graph.

From semantics to priced networks, this is a very straight forward fundamental, tool; like grep.  Interfaces can compute weights, price adjustments, on the fly. All attachments are guaranteed a complete "binary Dot product" decomposition..

Queue management in the bid ask is a good example. We use cascaded joins to find groupings of the disorder, a refinement; we have priced nodes, weighted.  At any given point and third party can extract the structure, at a chosen coarseness. The genius of it all is, just look like a graph.