Re: [stack] Concatenative Research

John Carter scripsit:

> ie. In a very strong sense, parameter names and parameter order is fluff and
> irrelevant to the actual computation being performed!


Well, either is dispensible given the other, but you can't get rid of both,
as many functions are not commutative.

> However, it you were to look at the stack effect declarations for fn1, and
> fn2, it would probably be obvious from the types alone, as which stackOps
> you would need between fn1 and fn2!

Only if you don't need to support subtraction.

> Thus in principle one could create a language where you could write a
> fragment as fn1 fn2 and the compiler could infer what stackOps (or calling
> parameters and their order) are required.

Scheme compilers actually do rearrange the argument order when all calling
points are known.

-- 
So they play that [tune] on                     John Cowan
their fascist banjos, eh?                       [email protected]
        --Great-Souled Sam                      http://www.ccil.org/~cowan