JSON Query Languages

// Bind to variable
$log = read(hdfs("log"));
$user = read(hdfs("user"));
// Query 1: filter and transform
$log
-> filter $.from == 101
-> transform { mandatory: $.msg };

// result ...
[
{
"mandatory": "Hello, world!"
}
]
// from http://code.google.com/p/jaql/wiki/JaqlOverview
Looks a lot like a sequence you might find in graph traversal. Bind phase, subquery, output. I guess my point here is that the query syntax and the nested form syntax should all be mushed into one group of syntax which describes a self directed graphs that convolve with any other graph. Notice they grabbed the dollar sign for us!!! But they brought back the parenthesis, I thought we dumped that.

Bind to variable should be transparent to the user. "user" and "log" are not files, they are name/graph pairs.
The convolution is: @{101.msg,log} = result, though I almost certainly got the syntax wrong (note I grabbed the = since they offered it!!. However, bottom line, JSON people are working this issue, working it enough that I am interested in writing a line of code or two. I think the inner join of the two syntax JSON sets (describe and query) gets what graph traversal needs.

I am still unsure of brackets. Does the graph layer need a hint that things might be kinda square? No, not really. The concept of squareness is likely the outcome of a slider bar somewhere. Squareness is a signal to noise ratio, it tells the bots how compact is the information storage.

Anyway, my inner join of JSON gets me the following Ugly Set:
{}=:$., From other places we got _*?!

Regarding the dollar sign. My understanding of dollar is that it means, do a defined method right now. I need reserved words or a dollar sign, I think that is the deal here. Maybe I can dump the dollar sign and get its functionality from named pairs.

What about string quotes? At the graph layer they are meaningless, at the parser, quotes really mean, include white space and ugly characters in the key value. However, we still have to work the regex for the key value='graph of bytes' problem. We have ugly set conflicts with regex. In theory our string version of graphs is TE itself, which should be regex identically. Notice graphs and regex both have the graph convolution model.