On Fri, 18 Sep 2020 at 20:56, Roger Orr via Std-Discussion <
std-discussion_at_[hidden]> wrote:

> There are some thoughts about the choice of language versus library in
>
> P2000R2 section 5.2 "Library and Language":
>
>
>
> "If we can supply a feature as a library, we should do so because it is
>
> easier to validate a design through
>
> experimentation, the facility will be available to users earlier, and a
>
> library is usually easier to specify in
>
> isolation."
>
>
Yes I’m aware of this position. However to be fair, we have had std::vector
(almost) in its current form since the dawn of the STL.

How much is enough time before a library feature proves itself so
fundamentally useful to a language that it is deemed validated?

Many (almost all?) other successful languages build in such fundamentals
very early in their lives. And it is not uncommon for libraries to become
core elsewhere.

Ones position on this will of course depend on one’s vision of what C++
wants to be, and I appreciate that there is a spectrum of view here.

Do we want C++ to be essentially C with a few twists and a slightly more
restrictive memory model, or do we want it to be a language in its own
right?

At present it seems to me that it is neither, with language features
depending on library types and library features depending on compiler
intrinsics.

I’ll stop here before I inflame too many tempers. A good jester knows when
the court is about to lose its sense of humour. But I don’t think it does
us any harm to think the unthinkable from time to time.

Thank you everyone for your attention.


>
> Roger.
>
>
>
> --
>
> Std-Discussion mailing list
>
> Std-Discussion_at_[hidden]
>
> https://lists.isocpp.org/mailman/listinfo.cgi/std-discussion
>
> --
Richard Hodges
hodges.r_at_[hidden]
office: +442032898513
home: +376841522
mobile: +376380212