Simplify Json builtin routines

1) We tried the json commands in PL/I, but sorry, they are programmer unfriendly.

If I want to use them well I have to write a program around them and this is really not what I want.

2) a variable type json would be interesting ...

sample json get:

{ "passes" : 3,
"data" :
[
{ "name" : "Mather", "elevation" : 12100 }
, { "name" : "Pinchot", "elevation" : 12130 }
, { "name" : "Glenn", "elevation" : 11940 }
]
}

dcl name char(20);
name = jsonget(jsonstring,"data->1->name")

--> if this is not matchable, then return a null value

dcl namearray[5] char(20);
name = jsonget(jsonstring,"data->name")

--> if this is not matchable, then return a null value, empty elements can be filled with a null value; too many elements should result in an exception

dcl passes bin fixed(15);
passes = jsonget(jsonstring,"passes")

-> if passes not in json, then return a null value

sample json put:

jsonput(json, variable, path*) -> * path is the way to the jsonelement, where the new tag should be inserted, if it is not in the current json, it should be inserted. If not there, then root is used.

sample:

dcl jsonstring char(2000); -> here maybe a varialbe type json would be better, to optimize pl/1 working with json.

dcl name char(20) init('peter');
dcl number bin fixed(15) init (200);

jsonput (json, name);
jsonput (json, number);
jsonput(json, name, "data"

result:

{ "name" : "peter",
"number" : 200,
"data" : { "name" : "peter",
"number" : 200
}
}

Hope I described it in a good way. Main thing is: as simple as possible to use as much optional as possible.