Date: Thu, 15 Oct 2026 13:03:53 -0400
On Thu, Oct 15, 2026 at 12:56 PM Ville Voutilainen via Std-Proposals <
std-proposals_at_[hidden]> wrote:
> On Thu, 15 Oct 2026 at 19:19, Brian Bi via Std-Proposals
> <std-proposals_at_[hidden]> wrote:
> >
> > I don't think segment trees are a good candidate for standardization.
> There are multiple different implementation strategies and the choice of
> implementation strategy affects performance in a meaningful way.
>
> Which part of that is not solvable by specifying complexity guarantees?
>
Ville, IIUC you're saying "LWG can force the STL to provide a particular
implementation by careful wording" (like we did for `unordered_map`).
But Brian, IIUC, is saying "Different kinds of users need *multiple
different* particular implementations, and the STL can't reasonably provide
*all* of them."
(I agree with Brian's point, as I understand it to be. However, I also
admit that that point didn't stop WG21 from standardizing unordered_map, or
regex, or hive, or linalg, or...)
–Arthur
std-proposals_at_[hidden]> wrote:
> On Thu, 15 Oct 2026 at 19:19, Brian Bi via Std-Proposals
> <std-proposals_at_[hidden]> wrote:
> >
> > I don't think segment trees are a good candidate for standardization.
> There are multiple different implementation strategies and the choice of
> implementation strategy affects performance in a meaningful way.
>
> Which part of that is not solvable by specifying complexity guarantees?
>
Ville, IIUC you're saying "LWG can force the STL to provide a particular
implementation by careful wording" (like we did for `unordered_map`).
But Brian, IIUC, is saying "Different kinds of users need *multiple
different* particular implementations, and the STL can't reasonably provide
*all* of them."
(I agree with Brian's point, as I understand it to be. However, I also
admit that that point didn't stop WG21 from standardizing unordered_map, or
regex, or hive, or linalg, or...)
–Arthur
Received on 2026-10-15 17:04:07
