On Fri, Sep 24, 2010 at 2:00 PM, R. Kent Dybvig <xxxxxx@cs.indiana.edu> wrote:

> Before this is put to bed, I wonder if a rationale should be given for
> why, in the end, there's no encoding for identifiers that don't map
> cleanly to file names.

Adding a rationale sounds good, but maybe challenging to convey
succinctly.  Based on the reasons I'm remembering, I'm thinking of
adding something like the following to the Design section:

------------------------------------------------
This SRFI does not specify any special handling of library names which
do not map cleanly to file names, so that characters with special
file-system meaning may be intentionally used for such meaning in
library names, and because, in addition to the characters disallowed
in Windows file names, the case-insensitivity of some file systems
requires an excessively complex design to achieve comprehensive
special handling of all the issues pertaining to mapping all possible
library names to portable file names.
------------------------------------------------

Feedback is welcome.