On Wed, Mar 5, 2014 at 9:12 AM, John Cowan <xxxxxx@mercury.ccil.org> wrote:
> Alan Manuel Gloria scripsit:
>
>> This is a rather SRFI-105-centric proposal, but I'd like to propose
>> that, if '<? and friends are given a single comparator argument, it
>> return a function predicate:
>
> I don't understand why this proposal in any way benefits SRFI 105-based
> code.  Can you explain further?

Mostly just a way of writing it shorter.

With make<?:

{x (make<? c) y (make<? c) z} ; ((make<? c) x y z)

With <? operating as make<? does now:

{x (<? c) y (<? c) z} ; ((<? c) x y z)

That's it, that's the only advantage for SRFI-105.

It's not clear to me if (<? c) would mean anything useful anyway, so
why not, but...

>
>> Admittedly, it's possible to just use 'make<? and friends if we're
>> going to use SRFI-105 anyway, but it seems clearer to use <? if
>> SRFI-105 is something to support.
>
> I don't see why.  The idea of <? and friends is that they are generalizations
> of < and friends (also string<? and friends, etc.); that is, they are
> predicates.  Rather than having the function be either a predicate or
> a predicate-maker, my idea is to separate the two purposes, so that
> in one case we always get a boolean result, in the other case a procedure.

...Well, I suppose I can agree with you here, there is such a thing as
clarity.  This is scheme, after all, not Perl, so something that is
convenient a few times for some people is not something to be
recklessly added to the language.

Sincerely,
Amkg