A new Unix utility, join x,y,z x,t,z all text files. Keep a set of search files, constructed in a lazy JSON search language, with appropriate wild cards and partial match routines in the attachment. The Y are simple text files.. Eventually, you and your search application develop their own special grammar, your graphs become become homomorphic. Them micro payments induce data providers to make their semantic graphs all homomorphic multiples of your search graphs, the principle of minimum traversal points applies.
Ultimately we make each call to join its own thread, the machine can handle asynchronous traversals from multiple sources. Just keep the output partitioned from either of the two inputs. We are ultimately dealing with billions of searches going after massive data base graphs, distributed mostly by auto cache. Micro pricing balances the congestion. The join kernels can be chained in real time across the internet with simple GET and PUT.
Example:
I ha e a list of business expressions, in text, in a file. Looks like:
Home reversion plan
Horizontal merger
Hostile takeover
House price surveys
Hyperinflation
Income protection insurance
Income statement
Income tax
I can write spaghetti, in my plain text interface to the join machine; and present that that list as comma separated groups of Dot sequences, imparting the logic of the list and needing only a few lines of spaghetti. But, once doing so, the list is available for plenty of cross joins with all manner of plain text, vie the join system. With this simple start, we define a search system that goes way beyond Giggles.
No comments:
Post a Comment