Search time for software is dropping fast. I found this neat little c code with some great tree data structures. A Stanford group did this code for public license, good stuff. I wonder if another Stanford student has connected this to a file system.
How do I look at this code? I want to take two trees, and convolve them, defining match and proceed grammar over them. I want to match my Dot and Comma operators. That is, start the graph layer down there, right out of the database file, using a tree structure optimized for convolutions. Get a node pointer definition and arithmetic that takes us across trees nifty fast. Then, get the BSON expression and data definitions. Use one of the ECMA byte code standards for compiled micro sequences. Web bots, the next big thing.
This tree structure, it is optimized for linear chain of descent and also linked element sets. And the code to get key matches, short, then push the node. This thing would skim through reams of disks. Get a BSON,graph triplet definition as the atom. Build the indexing system out of the named graph concept, also expand named graph to include URL jump. and drive that form onto the tree. Make the tree definition a combo set with the BSON definition in the standard. Then, you have this bare semantic race track, Cray can pick it up, Mango, Wiki, Cambridge Semantics.
TODO list. Look deeper into what a named graph is and how its grammar works.
Must resist urge.
By the way, Wiki is doing a great job, anybody working for Wiki has some great credentials.
No comments:
Post a Comment