sqlite lets you register a collating function, hence making is possible for the graph layer to construct a match on the matchable portion of the subject value in a triple. So you get this bounding index, the object pointer based on rowid, and the subject matcher, and the two together determine a pass or reject between the corresponding nodes on either of the commutative sub graphs.
So, here I am, thinking. Just as soon as the complete JSON syntax is complete, all issues and bus removed, etc etc, then hey, the code is a few months away from completion, built on sqlite. The wide spread use of sqlite means, by default, it becomes the basis of the semantic machine.
Where have I gone wrong?
Another topic:
Further, follow the analogy a bit more. What is a DOM tree? Just a portion of the giant graph, it is BSON, kind of, by definition. The browser display is nothing more than a convolution of some local subgraph of windowing methods with another subgraph of data.
My guess is their is an XML bias going on, hence the weakening of IndexDB, like a conspiracy of some sort! Really, it is No-SQL/Yes_XML. The thing is going to blow up in their faces. This g machine, already in the hands aof a dozen superior programmers, wadya gonna do?
Some day I will get over my phobia of looping, complex Unix typeface and make a code repository.
No comments:
Post a Comment