Re: scsh 0.6.7 is *broken*, 0.7 of very uncertain value Arthur A. Gleckler 11 Jul 2019 21:16 UTC

John Cowan <xxxxxx@ccil.org> writes:

| The question is whether the existence of scsh
| 0.7 (which is in working order) plus the list of
| differences in the Implementation section of
| SRFI 170, minimally satisfies the need for a
| SRFI 170 sample implementation.  (The reference
| to scsh 0.6.7 in that section is incorrect.)  If
| not, the best we can do is to write an
| implementation for one specific Scheme, either
| scsh or another one.

The very idea behind this SRFI, which is a good
one, almost inevitably leads to its implementation
being non-portable.  There is an escape clause in
the SRFI process document for such cases.  See the
sublist in item #7 under Structure:

  https://srfi.schemers.org/srfi-process.html#Structure

Having an implementation like that of SCSH 0.7,
which is relatively close to what the SRFI
specifies, will help implementers a lot.  Anything
you can do to provide sample implementations of
the differences, too, would be greatly appreciated
by future implementers, I'm sure.

My feeling is that as long as SRFI 170 doesn't
differ too much from SCSH 0.7, a list of
differences is sufficient.  But it would be much
better if some implementation, somewhere, actually
matched SRFI 170 in detail so that it could be
properly vetted by at least one implementer before
the SRFI is finalized.

Is there any Scheme implementer reading this who
would be willing to implement SRFI 170 in their
Scheme?