On Thu, Jul 27, 2017 at 8:28 AM, Gary Furnish via Boost
How many of the latest proposals for C++ have actually been in boost? Ranges? Polymorphic allocators? I can not remember the last time I watched a CPPCON/etc video and someone was talking about bringing something from boost into the standard library besides filesystem (which is legacy) or asio (which technically is standalone).
That's only because it is possible for a proposal to be accepted having just a paper, and sometimes no implementation (sized delete anyone?) One individual, who currently has a proposal far along, said to the effect that they skipped submitting to Boost first because the formal review process was too demanding and time consuming, and they wanted to get their proposal in to C++20. I disagree with this practice. A TS should be backed by a library that is already in use, which appears in an actual released product whether it is Boost or a standalone library, that has a version number, quality control, tests, and existing users. That is the route that I have chosen for Beast, so that the library may age in the Boost cask. To summarize, fewer proposals are coming from Boost because the standards for proposals have not been raised to meet at least the level of excellence required of a Boost library. Thanks