Date: Tue, 28 Jul 2026 12:55:22 -0700
On Tue, Jul 28, 2026 at 12:49 PM Jeremy Ong <jeremycong_at_[hidden]> wrote:
> The test as worded should stand, since "we should include X in the
> standard because Y isn't allowed to use non-standard code" is a pretty weak
> argument.
>
Worse than weak, it is dangerous, because it is effectively a rhetorical
device which can be used to avoid providing any substantive justification
for standardization. It works because the WG21 rules have not stated
explicitly that this not an acceptable reason.
In my opinion, corporations who can only use library components from an
international standard, should come together and form their own Working
Group under SC22, and develop their own separate draft for a "C++ Library
Attachmenté."
They can put whatever they want in it. They can evolve it separately
without being tied to WG21's work program. And then when anyone says "but I
can only use standard library components" we can point them there and get
on with our important work.
Best
> The test as worded should stand, since "we should include X in the
> standard because Y isn't allowed to use non-standard code" is a pretty weak
> argument.
>
Worse than weak, it is dangerous, because it is effectively a rhetorical
device which can be used to avoid providing any substantive justification
for standardization. It works because the WG21 rules have not stated
explicitly that this not an acceptable reason.
In my opinion, corporations who can only use library components from an
international standard, should come together and form their own Working
Group under SC22, and develop their own separate draft for a "C++ Library
Attachmenté."
They can put whatever they want in it. They can evolve it separately
without being tied to WG21's work program. And then when anyone says "but I
can only use standard library components" we can point them there and get
on with our important work.
Best
Received on 2026-07-28 20:03:51
