Simpler usage of named arguments

[tpluscode]

I'd like to add a simpler way of defining named arguments. Currently, each named argument requires a blank node with name/value pairs

[
  code:arguments [
    code:name "foo" ; code:value "foo" ;
  ] , [
    code:name "bar" ; code:value "bar" ;
  ] , [
    code:name "baz" ; code:value "baz" ;
  ] ;
] .

This is verbose and makes complex scenarios harder. I propose to flatten to a single blank node which represents an object and have properties mapped to code

[
  code:arguments [
    arg:foo "foo" ;
    arg:bar "bar" ;
    arg:baz "baz" ;
  ] ;
] .

To be decided what arg prefix means. Could be a shorthand for https://code.described.at/argument/, where we extract the last segment and use is as code arg name.

This should allow deep object structures and mixing positional and named arguments