New competitor to D

[Ola Fosheim Grøstad]

On Wednesday, 27 July 2022 at 03:51:28 UTC, Tejas wrote:
> I don't think they can afford to provide any automatic memory 
> management solutions, since everyone is basically looking for 
> an excuse to hate on them

They have already scared away the most opinionated C++ users by 
using an unfamiliar syntax and removing things like constructors, 
multiple inheritance and exceptions...


> At best, I can imagine them enshrining C++'s `make_shared`, 
> `shared/unique_ptr`, etc constructs for better 
> optimisations/error reporting/airtight implementations, so that

That would be great, as make_shared provides the same as you 
would get from ARC in Swift!


> I genuinely think this thing might be successful, people 
> disliked Go so much, yet look where it is today: the entire 
> foundation of the cloud is built on it;

Yeah, that is the scary part, isn't it? Carbon currently uses 
«CapitalizedFunctionNames()» just like public Go functions. Which 
is somewhat annoying as it creates a rather noisy picture in the 
editor. I prefer everything in a standard-lib being lower case as 
my own types should be more visible than the standard ones. 
(Phobos…).


> I feel Carbon will be a similar success as well, if they really 
> get C++ interop, tooling, and job opportunities right

I don't think they will start implementing the compiler itself 
until late 2023, so I guess we have to wait and see. Maybe ready 
for adoption in 5 years? So they effectively compete with what 
people expect from C++29?

I would feel better about Carbon if they had to compete with C++, 
as then they would listen more to the community, but maybe they 
only "compete" with their own management? In which case it will 
end up with whatever Chandler Carruth fancies.

Anyway, they are open for breaking changes for another year, and 
that is a breath of fresh air :-)