On 14 May 2015 at 7:48, Robert Ramey wrote:
This is a vision to the current Boost. Niall is prefectly free to propose it here, but for lot's of reasons, it's extremely unlikely that Boost could reach a consensus which would agree to such changes.
Agreed. It comes down to critical masses and swings towards a majority opinion.
However, there is an option which would be acceptable to Boost and it's members. Niall could easily create a forked version of that part of boost which meets his criteria. All the proposed ideas could be applied here. If Niall is correct and realization of his vision is inevitable, it will become clear eventually and any path forward would become clear. If Niall is incorrect, this fork will die a painless death.
I no longer think a full fat fork necessary. As almost all of the upcoming C++ 11/14 libraries are standalone, it makes no sense to force them into a monolithic distro, including the 1.x series. We also now have a mechanism to enable libraries to coexist in both C++ 03 and C++ 14 use contexts. No, I think a simple dashboard with "custom distro builder" is enough. Individual library maintainers are already making the choice to do better than before in their new libraries now they are freed of older C++ constraints. My handbook of C++ 11/14 library best practices https://svn.boost.org/trac/boost/wiki/BestPracticeHandbook is getting closer, 16 of 19 items are mostly finished. Items 17 to 19 are going to take quite a lot more work given their more speculative undefined future natures. I'm hoping that library maintainers find value in that Handbook when designing their Boost libraries.
And there is precedent for one person taking an idea, developing it with the idea that it may someday be adopted by Boost. I'm speaking of course of the Boost Library Incubator. The Incubator hasn't reached a level of success which would support a consensus for absorption into Boost, but it has had some success. The future of the Boot Library Incubator is still open - which is just fine.
The only concern I have about all this is the usage of the name "Boost". Usage of such a name has become quite valuable and users and public have come to know that, in the context of C++, it means a reputation for quality which has been built over the last 15 years with the sweat of 100's of persons. I would very much hate to see this name freely usable by anyone who want's to leverage on it for unknown and/or unendorsed purposes.
In hindsight, I also regret naming my talk "Boost 2.0" which I would not like to see confused with the current proposal.
Boost I think is the combined efforts of its participants. If we aim high and achieve high, that's where the Boost brand value comes from, even if we disagree about the particulars. Also, historically anyone could use Boost for anything as it is not trademarked, and it hasn't been a problem till now nor do I suspect it will be in the future (we are not important enough to subvert!). Niall -- ned Productions Limited Consulting http://www.nedproductions.biz/ http://ie.linkedin.com/in/nialldouglas/