Re: [stack] Concatenative versus Compositionality

On Dec 3, 2007 4:14 PM, John Nowak <[email protected]> wrote:
>  On Dec 3, 2007, at 1:52 PM, John Cowan wrote:
>
>  > Christopher Diggins scripsit:
>  >
>  >> No it is not true in the general case. But does the general case hold
>  >> for Forth or Postscript? Does it also hold for all Joy programs
>  >> without exception?
>  >
>  > Yes, though "programs" must be interpreted strictly. "1 [ 2" is not a
>  > Joy program.
>
>  There may also be type system restrictions involved. For example, I
>  believe this is a valid program in Cat:
>
>  foo == [10] dup apply
>
>  However, the due to the current lack of equirecursive types, this isn't:
>
>  foo == [10] bar
>  bar == dup apply

Cat 0.18.2 handles recursion by generating a new type ('A -> 'B) and
unifying that. So "dup apply" is now valid again.

>  Does that mean Cat isn't concatenative (or at least not fully
>  concatenative)?

The answer depends on whether an ill-typed program is valid or not.

- Christopher