support for json_decode and json_encode

[coderofsalvation]

make jq-dependency for json integration:

powscript:
               cat foo.json | json_decode items.users
                 if not value =~ 
                 echo "$keys"
                 echo "$values"
                 echo "$id $name $email"

bash: 
               cat foo.json | json_decode items.users # this will output tsv: "12\t'foo'"
                                 | json_expose
                 local keys="${*//[a-zA-Z0-9_]:/}"
                 local values="${*//:['"a-zA-Z0-9_]\t/\t}"
                 local id=12
                 local name="foo"
                 echo "$*" | grep "id:12" | awk '{print $2}' # will output name:'foo'
                 echo 
               done

why jq

first I thought about a fallback-mechanism to:

node -pe 'JSON.parse(process.argv[1]).foo' "$(cat foobar.json)"
python curl -s http://twitter.com/users/username.json | python -c "import json,sys;obj=json.load(sys.stdin);print obj['name'];"
perl
jq

why not json.sh or json.awk

• json.sh: probably slow and could break whole application.
• json.awk that would be better if it would evaluate simple 'jsonpaths' (can it?)

however, that introduces many incompatibilities concerning the jsonpath syntax-dialects.
JQ however:

• is already in many package-repositories
• could eventually easily become optional dependency of bash (native binding etc)
• is lightweight/fast/portable dependency (compared to node/python/perl)
• does support basic jsonpath evaluation