I wonder if it wouldn't be valuable to identify some target
applications for the FFI?

Between character issues, string and string index issues,
no-reasonable-way-to-support-writable-shared-strings/vectors, and so
on -- it seems clear to me that a portable FFI is never going to be
able to compete with a native FFI for some tasks.

At the same time, the bar can be pretty low for what a portable FFI
has to be able to do to be very useful.

Maybe we could pick a couple of libraries that exhibit common
requirements (conversion to and from basic C types, common cases of
flow of control, error handling and reporting) and aim to include in
the SRFI reference bindings for those libraries using the FFI.

-t