Re: [Asset Stewardship] About the Beman Project

From: Vinnie Falco (vinnie.falco_at_[hidden])
Date: 2024-09-15 20:19:38

On Sun, Sep 15, 2024 at 12:21 PM Zach Laine via Boost
<boost_at_[hidden]> wrote:
> ...the Boost review
> process and the WG21 review process have fundamentally different aims.
> Boost wants the best C++ libraries that appeal to its reviewers.  WG21
> wants to make sure something is appropriate for the standard.  Those
> are similar, and even overlapping, but fundamentally different goals.

This might be how it is today, but it is not how it should be. The
ongoing cost of an accepted Boost library is considerably less than
the ongoing cost of accepting a new standard library component. When a
library is accepted into the standard, it makes subsequent additions
more expensive, as the new addition must be considered in the context
of an expanded set of existing facilities. There was a brief
discussion earlier this year about which libraries should be accepted
into Boost. Peter's answer was "a library that is useful" (among other
things of course).

Being useful should be a necessary but insufficient quality of
standard library proposals. Not only must they be useful, but they
should appeal to a broad audience, and the paper must include the
rationale for why the library should be incorporated into standard
instead of remaining as a downloadable dependency from elsewhere. To
my knowledge, no paper has ever provided a quantitative analysis for
why it needs to be in the standard instead of existing as a third
party library. I think they should.

Thanks