Just to be clear. In my last post, I wasn't talking about conformance
to the SRFI process (as John and Arthur, I don't see a problem here),
I was talking about implications of inclusion into R7RS-large.

Am Mi., 28. Okt. 2020 um 16:09 Uhr schrieb John Cowan <xxxxxx@ccil.org>:
>
> There seem to be several senses of "portability" involved here:
>
> 1) Portability of the Scheme code between implementations.  I think it is clear that this has to be abandoned in favor of something else.
>
> 2) Portability to difficult ISAs.  If they don't do IEEE floats at all, either they have to emulate them in software, or it's hopeless.  If (like a few ARM designs) they have different endianness for double and uint64 types, it can be done, but it's not clear that it's worth doing for such an edge case.
>
> 3) Portability *of* a particular Scheme implementation to various machine-vendor-OS triplets.  I don't understand the relevance of this, except insofar as it is relevant to 2.
>
> 4) Portability to different C standards.  It is certainly possible to construct a well-chosen double (not a NaN) and put it in a double/uint64 union and see what integer you get, so that makes it possible in principle to overcome implementation-specific mapping problems.  (I did this long ago, unioning a string "\x01\x02\x03\x04" and seeing what came out as a uint32 in order to distinguish between big-, little-, and mixed-endianness.)  But overall I think it is safe to treat the behavior as reliable.
>
> In the end, the responsibility for correctness falls on the implementer of a particular Scheme.  The design using C code + FFI that we've been discussing is something between a 7b solution ("A mostly-portable solution that uses some kind of hooks provided in some Scheme interpreter/compiler. In this case, a detailed specification of the hooks must be included so that the SRFI is self-contained.") and a 7c solution ("An implementation-specific solution. Ideally, tricky issues that had to be dealt with in the implementation will be identified.").  Both of those are perfectly acceptable per <https://srfi.schemers.org/srfi-process.html> when a fully portable pure Scheme implementation (7a) is not possible.
>
>
>
> John Cowan          http://vrici.lojban.org/~cowan        xxxxxx@ccil.org
> And they pack their lyrics till they're so damn dense
> You could put 'em in your yard and you could use 'em for a fence.
>       --Alan Chapman, "Everybody Wants to Be Sondheim"