A flawless model is all the basics.

https://sourceforge.net/projects/cscall/files/MisFiles/ClassGuidelines.txt/download
... OO can have many meaning. I took the object to mean the basic entity of the
programming model. The concept of object, as a foundmental concept, has to be
solid, practical and easily usable. Otherwise, more codes and efforts will be
needed latter to fix it making the original goal, practically, a lie.
IOW, (nearly) a flawless model is all the basics...
>
After 25 years, I feel the idea of OO is lost. Something not right with OO?
I'd rather believe the idea of OO being not followed or addressed is the culprit.
>
>

General rule of the element of a programming language:
If too much effort is spent in figuring out what keywords/pseudo-directive
should be in the program. Something is wrong. E.g.

T* a=b;

The target is just "mov a,b", anything more may be considered redundant and
inefficient. Now we have const, constexpr,
volatile/atomic,promotion/narrowing,mutable,modifier/attr..., and maybe isnull,
ownership,...) to consider in programming "mov a,b" for "T a=b".

Why not using assembly? Well, after not much tries, we found what we wanted was
not much different from C, but with something compromised (e.g, stack,...),
and mostly, some classified as hardware 'specifics' (ironically, differences
are the reason different hardwares exist. we use new hardware because it has something new).

Does it have to be C, afterall hardware can be redesigned to align with high
level languages? Answer: At least we don't have idea of any machine that can
exceed Turing Machine (quantum computing is just faster in some aspect).
Assembly,C is pretty much close to the description of TM...

In this stage, I would say C is desirable is because it is close to human
natural language of which the feature is recursion,modifiers in 'sentence' (not list of instructions):
https://study.com/academy/lesson/noam-chomsky-on-language-theories-lesson-quiz.html
... all languages have formal universals and principles in common, with specific options and limits
for variation in grammar and features between languages...

But too close to natural language should not be the goal because uncertainty
should and must exist in human language... So the language of a model.

I had a initial test: https://sourceforge.net/projects/symbcomp/
I feel programming language (C/C++) can do more, and can even be better
targeting as the foundation of math/logic. Reason: it models the real thing,
better than math. and logic.