[bitc-dev] Re-thinking currying

[David-Sarah Hopwood]

Jonathan S. Shapiro wrote:
> Specifically, I am considering a change to the apply rule where, given:
> 
>   (f a b c)
>   f: (fn 'a 'b -> (fn 'c -> 'd))
> 
> we would canonicalize this apply based on the function type of f into:
> 
>   (apply (apply f a b) c)
> 
> This is "curry like"; the main difference being that it supports N-ary
> functions.

Does this mean that the dynamic semantics depends on types, i.e. a BitC
program cannot be executed by erasing all types, and then running the
resulting program in a corresponding untyped language?

That seems like a property that should not be sacrificed without
some thought. (I don't know whether there are any other reasons why
it already doesn't apply to BitC.)

> The main use case for this, by the way, is variadic functions that
> accept an argument of type DYNAMIC. Namely: printf-like functions.

When calling a printf-like function, you know at the call site which
arguments are variadic. As long as the syntax is still fairly concise,
I don't think it would matter if calling such a function required a
slightly different syntax than calling a non-variadic function with
the same total number of arguments.

-- 
David-Sarah Hopwood ⚥