Improve the OOP ABI

[max haughton]

On Sunday, 1 October 2023 at 22:01:56 UTC, deadalnix wrote:
> On Sunday, 1 October 2023 at 13:07:10 UTC, Alexandru Ermicioi 
> wrote:
>> On Sunday, 1 October 2023 at 00:21:48 UTC, ryuukk_ wrote:
>>> However, you are wrong at expecting an impact for D by 
>>> improving OOP, nope, D needs to improve the non-OOP story, 
>>> more OOP is repulsive, specially for system languages
>>
>> Thats one extremely biased view on D use. OOP is useful for 
>> application development (irrelevant of your feelings), and D 
>> is also used for app development.
>
> Indeed, if the basics such as OOP are not implemented right, 
> pilling up more and more on top of it is not going to help. 
> Solid foundations matter. OOP is just a basic feature of modern 
> programing languages and widely used in the industry.

Also worth adding that there is a difference between the midwit 
bullshit enterprise OOP, the modern synthesis OOP one might read 
in a book, and merely classes.

Improvements to the latter should be a free win, other than ABI 
breakage.
One of the reasons why I think we need to massively reduce the 
surface of the D toolchain is that if we don't we massively 
increase the activation energy to actually implement these 
changes —  hand waving using Arrhenius/Boltzmann, T ~ 
O(\exp(E_a)), doesn't scale well.