In a semantic network, mainly of key word ontologies, something has to externally enforce schema the way sql internally enforces it. Web bots do that, their main measuring tool is the schema. Find a schema thet best fits a table with variable length rows. That task is one of the major tasks for web bots, get most of the ontology in mostly square tables mostly where the client keeps asking for it. That's the semantic intelligence.
So, the : (schema) and its associated $ (attribute) are special operators. The machine can match on these, it can replay the schema pattern over and over, but schema still inherits the basic laws of graphs, it can count, translate, etc. It will have special match events, it will have various ways it can be overloaded in an ascension of filters. It is important, and it is important that schema bots can also do some minimal type and constraint checking on the keywords, using bson definitions. This is big since schema is one of the security tools also.
Example, this should mean something to everyone:
UseSchema:((string$_.int$,binary$_),g1,g2,g3,g4,g5)
where the g are conformal subgraphs. Whatever this means, it should make clients and geeks say, OK, that makes sense. What I want is that the first argument is a patterns, and go collecting data over the little gs, fill in by partial ordering, operator match and matching caste pairings from bson. The schema pattern in argument one applies to all the enclosed set elements. Take that meaning and specialize it to search pattern or enforcement pattern, or most likely pattern or constant pattern by look up table. Make it specialize in context specific syntax, just keep the ugly count small. Sweat the corner cases out of band.
No comments:
Post a Comment