(define (fix p b) (define f (p (lambda i (apply f i)) b)) f)
(define (mix c p) (lambda (f b) (c f (p f b))))

It can be made even shorter. Or to avoid exponentially bad performance
without having to think, it can use force and delay everywhere rather
than sparingly.

Explanations:
https://github.com/metareflection/poof/

Looks at util/coop.scm there (with a shorter prefix in coop.rkt or
coop.ss) for a complete example with plenty of helper functions, that
I now use for all my (pure functional) slides.

—♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org
“Think you can, or think you can’t — either way, you’ll be right.”  — Henry Ford

On Wed, May 21, 2025 at 9:49 PM John Cowan <xxxxxx@ccil.org> wrote:
>
> This SRFI is about fine-tuning the existing Prometheus design, which is based on Self, not about something completely minimal. It's not like we are going to wind up with a single standardized object system, after all. It's too easy to build whatever object system suits you.
>
> What are your two lines?
>
> On Wed, May 21, 2025, 7:42 AM Faré <xxxxxx@gmail.com> wrote:
>>
>> Strong disagree. Why need a root object at all?
>>
>> My prototype object system Gerbil-POO has no notion of "root object",
>> no need for "system-defined slots". You get the slots you ask for, no
>> more no less.
>>
>> My other prototype object system, COOP, has no notion of object
>> either. Just mixin inheritance of prototypes that you can instantiate
>> into records with a simple fixed-point operation. OOP without objects,
>> in two lines of code—plus a few helpers to make it more usable.
>>
>> —♯ƒ • François-René ÐVB Rideau •Reflection&Cybernethics• http://fare.tunes.org
>> “One has not only a legal but a moral responsibility to obey just laws.
>> Conversely, one has a moral responsibility to disobey unjust laws.”
>>         — Martin Luther King, Letters From Birmingham Jail
>>
>>
>> On Mon, May 19, 2025 at 10:20 PM John Cowan <xxxxxx@ccil.org> wrote:
>> >
>> > Looking good! However, I don't see the benefit of having multiple root objects. If by convention *the-root-object* has no slots except the system-defined ones, then any alternative roots can be direct children of the unique root with no loss of function. Removing everything in that section except *the-root-object* would achieve a useful simplification.
>> >
>> > On Mon, May 19, 2025, 11:08 AM Daniel Ziltener - dziltener at lyrion.ch (via srfi-discuss list) <xxxxxx@srfi.schemers.org> wrote:
>> >>
>> >> I updated the proposal. I added a mirror interface for additional
>> >> reflection functionality. Please have a look, feedback, advice and
>> >> additions welcome; unless further time for discussion is needed, this
>> >> will be handed in on the 26th.
>> >>
>> >> The proposal can be found here:
>> >> https://forgejo.lyrion.ch/zilti/srfi-object-system-proposal/src/branch/master/spec.org
>> >>