C++ Logo


Advanced search

Re: On "transparently replaceable" in std::vector operations

From: Ville Voutilainen <ville.voutilainen_at_[hidden]>
Date: Fri, 30 Apr 2021 21:39:12 +0300
On Fri, 30 Apr 2021 at 21:23, Hyman Rosen via Std-Discussion
<std-discussion_at_[hidden]> wrote:
> Since when is the ability to run your code in the compiler the driving force behind what C++ should be? Seems like a silly goal.

I may be wrong about this, but all symbolic evaluation arises out of
the same modeled principles. This includes
compile-time evaluation, static checking, proof-of-correctness when
optimizing, and "no false positives" when
instrumenting run-time checks that are too expensive or infeasible to
prove at compile-time.

Your suggestion of "scrap the lifetimes and treat everything as bags
of bits with types that are determined by
the glvalue accessing the storage" would need to prove that it doesn't
disturb the foundation of any of that,
not just that it turns off optimizations that you don't like. And even
when it turns off optimizations you don't like,
it would need to prove that that's a net win, which is of course not a
question of pure logic and math.

Write a paper that we can analyze, Hyman. Your vague suggestions of
where the model is wrong and how
it should be fixed can't be analyzed without one. The picture you
paint is too inaccurate to be anything
but noise. Stop claiming that we would never look at it because we're
an "optimizationalist coven". I'm not
a member of any cult, nor any church, but I know where to get feedback
on such a proposal, meaning
feedback in the analytical sense, not in the sense that it would just
be rejected because we're all blind idiots.

Or, if you prefer, let's just continue this fun like we have so far,
for a couple of years. You keep telling everybody
who cares that C++ is broken, and devil may care. :) I occasionally,
when I bother, ask you to elaborate,
and you keep telling me that it's pointless because it wouldn't be
seriously considered anyway, even if
correct. I keep redacting what I really think about that sort of
nonsense. :P For what it's worth, the word
on the streets is that your analytical capabilities are far beyond
what your cop-outs about this matter

Received on 2021-04-30 13:39:25