[plt-scheme] srfi-0 cond-expand

Michael Sperber [Mr.  Preprocessor] <sperber at informatik.uni-tuebingen.de> wrote:
> This is implementation-dependent---I'd say duplicate the semantics of
> the underlying LOAD.

I dislike that, because it means that the code won't be portable, despite
the configuration language.  SRFI 7 got through without much discussion and
I think it is under-specified.  The incentive isn't there to do another SRFI
to fix it, because SRFI 0 probably will work well enough.

> The big problem with SRFI 0 is that it assumes a monolithic Scheme
> implementation: it doesn't support SRFIs which are libraries, which is
> what most SRFIs are in one form or another, especially in the context
> of PLT Scheme.

Huh?  The code I posted allows this to be done for PLT Scheme, by specifying
what modules to require for a feature symbol.  However, the interface for
adding new features is totally arbitrary at present, as David Rush correctly
points out.

MJR