On Saturday, 11 February 2023 19:32:32 PST Phil Bouchard via Std-Proposals
wrote:
> Anyway my point is if Microsoft don't wanna be ISO certified, including low
> level memory management then it's their choice.
>
> It's the choice also for the industry to disregard non-ISO certified
> companies.

Indeed. You can vote with your feet or, more precisely, with your wallet, and
choose other options. There are a lot of OSes out there, from tiny, real-time
and near real-time ones that are much easier to certify for a purpose (like
Zephyr and myNewt and FreeRTOS) to full server and consumer OSes.

I know a few large companies that don't trust Linux for their server needs or
have legal qualms about it, and therefore use FreeBSD. It's their choice.

> > That's not paranoia. They already did killed Netscapa and Wordperfect
> > similarly.

And got in legal trouble for it. Same as when Intel[*] got caught having code
that detected competitor processors and chose unoptimised code paths.

None of this is in-scope for the C++ Standard or this mailing list. There's
nothing we can or should do about it. What's more, the Standard is not going
to make recommendations that go against the legitimate objections of one of
the three main compiler vendors, especially if the objection is "this is not
implementable in the OS"..

[*] disclaimer; I work for Intel, but I have no knowledge of the details of
the case at hand.

-- 
Thiago Macieira - thiago (AT) macieira.info - thiago (AT) kde.org
   Software Architect - Intel DCAI Cloud Engineering