P1655 LEWG Omnibus Design Policy Paper

[wg21bot]

P1655R0 LEWG Omnibus Design Policy Paper (Zach Laine)

[tituswinters]

Revise, proposing the addition of a Standing Document which will be edited by LEWG via the regular paper-mailing process. Changes to that SD will be made as Straw Poll motions in plenary during each meeting.
SF F N A SA
21 5 0 0 0

Someone should write a paper describing policy on ranges x parallel algorithms. (We need SG1 plus Ranges experts to volunteer.)
SF F N A SA
13 7 1 0 0
David H, Matthias, Nevin

A constructor callable with a single argument or a conversion operator should be declared explicit unless it:
— is a conversion between two types that are essentially the same;
— preserves all data during conversion;
— imposes little or no performance penalty;
— does not throw; and
— results in a memory-safe value or is designed to be a reference/view type.
SF F N A SA
20 5 1 0 0

Add to the above “should not implicitly convert both ways.”
SF F N A SA
2 3 8 9 3

Standard library functions that return bool should not be prefixed with is_.
SF F N A SA
1 2 9 5 3

Include discussion of nodiscard in a revision.
SF F N A SA
7 6 8 0 0

Predicate-like type traits should be prefixed with is_/has_.
SF F N A SA
5 7 5 0 0

Ask Jonathan Mueller for his notes from concept renaming for rules on concept naming in the revision.
Unanimous consent

An type-erased class template should be named any_${C}, where $C is the name of the concept that it represents.
SF F N A SA
11 7 5 1 1

Incorporate the above revisions in a proposed initial-draft for a future standing document.
Unanimous consent

(Revised paper should include proposed initial wording for such a document.)

[jensmaurer]

If it's still on LEWG's plate, we should leave the LEWG label on it. (Otherwise, the proposal looks orphaned.) I've removed the milestone pending a revised paper.