C++ Logo

std-proposals

Advanced search

Re: [std-proposals] Function overload set type information loss

From: Marcin Jaczewski <marcinjaczewski86_at_[hidden]>
Date: Thu, 1 Aug 2024 13:32:00 +0200
czw., 1 sie 2024 o 12:59 organicoman <organicoman_at_[hidden]> napisał(a):
>
>
>
>
>
> Sent from my Galaxy
>
>
>
>
> czw., 1 sie 2024 o 12:39 organicoman <organicoman_at_[hidden]> napisał(a):
> >
> > Please watch some of Leslie Lamport's videos.
> > He is a Turing award holder.
> > He explains exactly the opposite of what you say.
> >
>
> What video? and how opposite? And opposite to what statement?
> You do not address any of my concerns, only again avoid hard questions.
> If I am wrong you should be able to quote me and exactly show where I am wrong.
>
> That's true and fair.
>
> So allow me answering it.
>
> In your analogy, about not approaching operations table unless someone checks you are qualified.
> Given that logic, who checked the first surgeon in the world? And against what?

By row of dead bodies after failed operations, first surgeons were
simply people with sharp knives,
but now after centuries of trial and error we can ask new ones if they
do their homework and will
not repeat mistakes made by the first generation of surgeons that in
many cases harm more than help.

C++ is older than me, many brilliant people work on it but even they
did fail and make blunders during
standardization. And we all need to learn from this, understand and
try not to repeat it again.
Without a close understanding of the standard and its history we will
repeat these blunders or even create bigger ones.

This is why people ask you about your understanding of C++, because if
you try to build something
on misunderstand of some of its parts, then whole construct will be
misaligned and fail to reach its goals.
Like building a home, some failure in foundations will cause the whole
home to crack or even collapse.


>
> Only trust, help, discussion and good intentions can encourage that first surgeon to break through.
>

Trust based on what? If I meet someone and he use correct language and
he show that he know what he saying
then I can risk trusting him to allow him to do first aid in car crash
but if some things that he say is bizarre and not match
exactly common knowledge then how I could risk giving him access to a
car crash victims?

> Yes, we have an accumalation of knowledge of experts in C++ language in a form of a standard, but that is always a subject of scrutiny. What was advocate for today, can be advocate against tomorrow, it is just a matter of someone bold enough to contest the status-quo.
>
> Did this address you concerns?

This is not if someone is "bold" but if he is "correct" or even more
if its "useful".
We are still waiting for real life code that show usefuls of this
change and proves that current C++ can't solve the same problem
easily.
We already had `std::fucnction` (and its flavors) and `std::any` why
C++ need more?

Received on 2024-08-01 11:32:13