Re: "dict"ionary typing?

On 12/05/2013 11:01 AM, Michael Hendricks wrote:
>    ?- writeln(append([a,b],[c,d],~)).
>    [a,b,c,d].
>
> ?- length(List, ~ is 1 + 1).
>    List = [_G1821, _G1824].

This is indeed an interesting thought for using arbitrary
predicates as a function.  The is/2 case looks nice.  The
append case is a bit less convincing with the ~ burried
somewhere deep down in a term.  Of course, typically at
least one of append's `input' arguments will be a variable.

I doubt it will work out satisfactoty for dicts.  Compare

	?- writeln(get_dict(x,D,~)).
     to  ?- writeln(D.x).

In the dict case, composition will also be quite typical
to access a field in a nested dict structure.

Also user functions on dicts play a different role as they
exploit the tag of the dict, making it as an object.

	Cheers --- Jan