On Wed, Nov 29, 2023 at 2:56 PM Robert Ramey via Boost < boost@lists.boost.org> wrote:
On 11/28/23 4:54 PM, Jeff Garland via Boost wrote:
On Tue, Nov 28, 2023 at 3:35 PM Robert Ramey via Boost < boost@lists.boost.org> wrote:
And maybe it should be more than some repos -- like an actual distribution that could overlay a boost release with the libraries that are part of the incubator.
This conflicts with my vision of a "modular" boost. I would hope that if our norms about boost rules were a little more formal (Concepts) we could create libraries which didn't have to be part of boost to function. The we could just a a library at will and be on our way. This would work for nich libraries and questionable libraries like those in the incubator. The library could be dropped just by deleting it. This seems like it should be possible since no other library will depend on the inclubator/nitch library. Maybe were already there. But our build/test/documentation infrastructure doesn't encourage this type of modularity.
I think we can do both -- we can distribute collections while encouraging modular libraries that can be used on their own. I think in most cases that might still mean a tie back to boost.core, but that's small.
It turns out that many/most of the submissions were not even close to boost quality. And those that were, didn't really require the incubator to get accepted. It also required altering the test/build/documentation infrastructure to work well. That is, it sort of required that "modular boost" be more of a reality than it currently is.
Interesting -- it's almost as if the authors that are gonna succeed in boost can cut thru everything by themselves. Jeff