Boost logo

Boost :

From: Ville Voutilainen (ville.voutilainen_at_[hidden])
Date: 2020-06-28 18:55:19


On Sun, 28 Jun 2020 at 19:40, Glen Fernandes via Boost
<boost_at_[hidden]> wrote:
>
> On Sun, Jun 28, 2020 at 12:24 PM Vinnie Falco wrote:
> >
> > On Sat, Jun 27, 2020 at 4:36 PM Jeff Garland wrote:
> > > Being in std:: really is the ultimate c++ distro.
> >
> > Yes well this attitude needs to change as it is obviously
> > unsustainable. Lowering Boost's acceptance bar doesn't sound like a
> > good way to solve this.
>
> +1.
>
> Boost is a peer reviewed collection of libraries. Boost's success has
> come from that peer review.
>
> Just because Boost.X was peer reviewed and has built user trust,
> sticking a "Boost." in front of "Y" to form "Boost.Y" just because Y
> is proposed for the standard isn't right.
>
> If "Y" wants to be part of Boost and be "Boost.Y", it should be
> formally reviewed.
>
> Now if you want to have "BoostExperimental.Y" then make a new project
> called BoostExperimental but let's not shovel anything into the Boost
> release distribution that wasn't peer reviewed.
>
> If the problem being solved is unacceptable things being put into the
> C++ standard library, and the idea is having them go through Boost
> first, I like that. But that should mean going through Boost's formal
> review process. Not being exempt just because they were proposed for
> the standard.
>
> The pathological case scenario is Boost's reputation and quality being
> diminished and consequently Boost not being a viable means for those
> proposed libraries to get user experience anyway.

Fully agreed. Lowering the bar of getting something into Boost doesn't
help anyone.

Some standard proposals come with the statement "this has been
accepted into Boost",
and some with "this has been shipping in Boost for N years". Those are
useful bits
of information. They become useless bits of information if Boost
starts accepting libraries
without peer review. They become worse than useless if Boost starts
accepting libraries
just because they are in the standard pipe, and I don't think anyone
has suggested that.

Boost can be a step on the way to include a library into the standard.
But it should be such
a step if and only if the expected-quality peer review that we've come
to expect from Boost takes place.
Boost is also not only such a step, it's more than that, so it
shouldn't be subservient to the purposes
of standardizing library proposals.


Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk