dynamic function arguments or generic system

[MSMissing]

Since functions are basically macros, it would make sense to have a data type for function arguments that can be any type that works with the code.

fn sum(dyn stack, i16 r) {
  r = 0;
  cell x = 0;
  while stack.len {
    pop(stack, x);
    add(r, x);
  }
}

Or, there could be a type abstraction system (would fit with #43)

generic stack {
  cell len,
  fn push(),
  fn push_d(),
  fn unshift(),
  fn unshift_d()
  fn pop(),
  fn shift()
}

fn sum(stack s, i16 r) {
  r = 0;
  cell x = 0;
  while s.len {
    s.pop(x);
    r.add(x);
  }
}

[Heathcorp]

@MSMissing I was thinking about whether generics are needed, and it probably wouldn't hurt to have them (except in the work of maintaining it, but eh whatever this is a fun project).

This idea seems interesting, akin to rust traits. However I feel that adding a trait system would be a pretty insane amount of work for now. I was thinking that generic programming could be implemented more simply, with function signatures/overloading.

Basically the programmer and stdlib is responsible for making sure every function overload produces the same behaviour, and a generic function can work as long as all functions used inside have overloads for the types being used.

Basically, I like your first idea with the dyn keyword, I just would like to decide on the exact keyword/syntax to use, I will add it to the new language design proposal in #43 .