> On Nov 22, 2016, at 6:53 PM, John Cowan <xxxxxx@ccil.org> wrote:
>
> What's the Right Thing to do here?  Provide portable but inexact implementations, or just say that the implementation is inherently specific to a given Scheme?

I suggested before that the functions that are just a renaming and a thin FFI layer over C library functions should be put in a table and the others described separately.

I don’t believe any Scheme implementor is going to reimplement a significant chunk of libc to support a SRFI.  If this SRFI is designed to be a standard way to refer to libc functions, perhaps it should be written that way.

It may be that a FFI is necessary to implement this SRFI; that would be different, but OK with me.

Brad