Re: Remaining work on SRFI 178
Marc Nieper-Wißkirchen 17 Aug 2020 15:42 UTC
Am Mo., 17. Aug. 2020 um 17:10 Uhr schrieb Wolfgang Corcoran-Mathe
<xxxxxx@sigwinch.xyz>:
>
> On 2020-08-17 10:03 -0400, John Cowan wrote:
> > On Mon, Aug 17, 2020 at 1:44 AM Marc Nieper-Wißkirchen <
> > xxxxxx@nieper-wisskirchen.de> wrote:
> >
> > > bitvector-unfold does not have a /bool or /int suffix in John's draft.
> >
> > It doesn't actually need one: proc can be required to accept a bit.
>
> If this were why the unfolds lacked /bool and /int flavors, we could
> drop them with -map and -fold, as well. But this is not the reason.
> The procedure argument of bitvector-unfold(-right) is only applied
> to index and seed values; the latter can be anything. The procedure
> is require to return a bit as its first value. Since the bit value
> can be easily unwrapped, it's unnecessary (and something of an
> abstraction leak) to have two versions.
Thank you for this explanation, Wolfgang. So the suffixes "/bool" and
"/int" refer to how values are communicated _to_ user-provided
procedures and continuations.