C++ Logo

std-proposals

Advanced search

Re: nullptr_t and nullopt_t should both have operator<=> and operator== to enable the *_with concepts

From: Arthur O'Dwyer <arthur.j.odwyer_at_[hidden]>
Date: Fri, 9 Jul 2021 13:33:44 -0400
On Fri, Jul 9, 2021 at 10:36 AM Ville Voutilainen <
ville.voutilainen_at_[hidden]> wrote:

> On Fri, 9 Jul 2021 at 16:33, Arthur O'Dwyer via Std-Proposals
> <std-proposals_at_[hidden]> wrote:
> >
> > On Fri, Jul 9, 2021 at 7:18 AM Giuseppe D'Angelo via Std-Proposals <
> std-proposals_at_[hidden]> wrote:
> >>
> >> Il 09/07/21 08:49, Justin Bassett ha scritto:
> >> > Thanks for the insight. I don't know much about the different working
> >> > groups. What are the different groups, and what does it mean if I
> target
> >> > them (other than putting them under "Audience")? How do I know which
> >> > groups to target?
> >>
> >> Please see here:
> >> https://isocpp.org/std/the-committee
> >>
> >> I'm far from an expert in the area, but you're proposing design changes
> >> to both language and library, so you should target the respective
> >> Evolution groups, and possibly their Incubators study groups.
> >
> > I'm also not an expert. :) My understanding is that for the purposes of
> the "Audience" field, literally all you have to do is categorize your paper
> as one of three kinds: "CWG," "LWG," or both. Once you've marked your paper
> appropriately as "CWG" and/or "LWG", then that specific paper's progress
> through the EWGI->EWG->CWG and/or LEWGI/LEWG/LWG funnels will be tracked in
> real time by
> > https://github.com/cplusplus/papers/issues
>
> I am an expert, and that's wholly incorrect. This paper should have
> EWG as its target audience; while the nullopt part
> of it is library-only, it's proposing the general extension that the
> special null types for nullable handles become comparable,
> and doing so across library types and built-in types.
>

I agree that this paper is targeting both the EWG->CWG funnel and the
LEWG->LWG funnel. According to my understanding, that means its "Audience:"
should list both EWG and LEWG (or both CWG and LWG; or both EWG and LWG; I
claim it doesn't *particularly* matter which stage(s) of the pipeline you
name in the paper, because it's going to change over time anyway).

Could you elaborate on your reasons for describing this model of
audience-selection as "wholly incorrect," and suggest a different model?

If I had to take a second approximation, I'd add that perhaps it'd be
clearer to target behavior-changing papers like this one to their
respective evolution groups ("EWG and LEWG" in this case). If an author is
targeting a paper at "CWG" and/or "LWG", that's a signal that the author
intends their paper to be (1) not-behavior-changing, at least not in the
"evolutionary" sense; and (2) perhaps fixing a specific documented defect
report in that group, where the evolutionary direction has already been
agreed on and documented. For example, Tim Song's P2259R0 targeted "LWG",
not "LEWG".
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2020/p2259r0.html

Some authors do update their targets to match the paper's progress through
the pipeline, e.g. Jean-Heyd Meneide's P1682R0 said "LEWG" but P1682R1 said
"LWG"
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2019/p1682r0.html
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2019/p1682r1.html
However, my mental model says that that kind of updating is unnecessary
(certainly not harmful, but not positively helpful to scheduling) because
at best it's redundant with the source of truth
https://github.com/cplusplus/papers/issues and at worst it *conflicts* with
the source of truth https://github.com/cplusplus/papers/issues .

I don't immediately see any papers in the current mailing that list "EWGI"
or "LEWGI" in their target "Audience"s (but I probably missed some,
especially if they gave the SG number instead).
Some list a whole breadcrumb trail, e.g. "SG14, LEWG, LWG, WG21"; again my
understanding is that that's not harmful but AFAIK not particularly helpful
either. (Although I can see how "SG14, LEWG" would be a way of
communicating "I think this is a behavior-changing library proposal *and* I
*want* it to be scheduled in SG14 first instead of LEWG," whereas if you
listed only "LEWG" you'd be leaving it more up to the various chairs'
discretion about whether it should be diverted to SG14 or not.)

–Arthur

Received on 2021-07-09 12:33:58