>> I wondered if maybe it was practical to create a new language by removing stuff from C++.

What would be removed and why?

The issue is that whatever is left would not be C++ and it would be difficult to get everyone to agree about what should be removed.

Hyrum's law (https://www.hyrumslaw.com/) also applies, so you can't remove anything without breaking things for someone.

If you want a simpler, yet quite capable modern language, I would recommend Go (https://go.dev/). It might not replace C++ depending on your use case (for example: embedded programming or writing operating system drivers), but it is a good general purpose language that is easy to learn and easy to use.

I respectfully disagree. And as i pointed out in my other comment cpp_front and google carbon are trying to do just that.

Bjarne is quoted to have said that there is a smaller C++ inside the current C++. As everything that was build buy gradual evolution, a clean reboot can converge on a "better" solution

Cpp_front and Carbon are not C++.

Parts of Cpp_front might become part of C++, but adding more stuff to C++ (usually increasing the overall complexity) has not been an issue for C++. That could be seen as part of the problem why C++ is so large and complex.

By way of comparison, Carbon calls itself: "fundamentally a successor language approach, rather than an attempt to incrementally evolve C++." (https://github.com/carbon-language/carbon-lang) It aims to be a successor to C++, not the evolution of C++.

>> Bjarne is quoted to have said that there is a smaller C++ inside the current C++.

To officially be C++, the Standards Committee must agree to deprecate the features that give a smaller C++ in a future release of the C++ Standard.

Good luck getting the C++ Standards Committee to agree on what to remove.

>> a clean reboot can converge on a "better" solution

Yes, but would it be backward compatible with existing C++ codebases? If the "clean reboot" is too different or not backward compatible enough would it be adopted by existing C++ users / the community?

If it required significant effort to port old C++ code to the rebooted version, why would the rebooted version be chosen over another language?(for example: Rust, Go, Ada, Zig, etc.)

Perl tried an ambitious rewrite from Perl 5 to Perl 6 and it did not work very well: it took too long, did not see wide adoption, and ultimately Perl 6 was renamed Raku.

Perl did not have to push Perl 6 changes through a standards committee or try to get a dozen or more compiler implementors to agree. (https://en.m.wikipedia.org/wiki/List_of_compilers#C++_compil...)

While it is possible that C++ could evolve into a smaller, cleaner version of itself, I do not think it is likely given all of the forces involved.