Joining separate databases

I have two attachments for mym join machine. which seems to work well.

One is my lazy JSON interface, with its wildcard, and more intelligent grouping.

The other is my plain text data base.  This attachment crawls through arbitrary text and produces the isolated words in sequence.A very simple interface, collects the ext set of alph character, and appear like a long omma sequence of key words.

So I can join:

join_cursor(JSON_cursor,Text_cursor,out_cursor);

This is all very little code, not hard to maintain.  Adding more search opcodes is easily done, the join machine does not change, all the spaghetti is in the enhanced search grammar, in the attachments. It is a great separation of interfaces.

Up and running. I can feed arbitrary text to the join machine, my opposing search graph can be enhanced with its own sophistacet  match.  At the same time my 'plain text' random word database van be enhanced, give it capability to group by sentence and paragraph, and trim off the words that add little to search..

I will be able to run two word streams together, feed that to a search graph; chain the joins.  In the attachment write spaghetti code, just follow the rule and have your call backs. Powerful stuff.