On 10/19/18 3:51 AM, Robert Ramey via Boost wrote:
On 10/18/18 5:40 PM, Andrey Semashev via Boost wrote:
A quick reply to this particular part. I'm opposed to this anonymity protocol and think that submitters should be *required* to come forward and actively participate in the review.
Of course. But is it necessary that they identify themselves with their real names?
For the sake of mailing list and similar conversations, including the review, real names probably don't matter that much. Although, frankly, on this list there aren't many regular participants who use an alias. But there are contexts where real names are required or simply appropriate. For example, in copyright notices in the source code or documentation. Or IRL communication, should one happen e.g. in C++Now or somewhere else. There is also an awkward moment when you reveal the author's name, who was previously known by an alias, at which point you only cause confusion. And probably either you or the author has to prove his authentity.
I think trying to attract more submissions by relieving the authors from the review process is terribly misguided and detrimental to both Boost and the proposed submissions.
In no way is it my intention to do that. Of course I expect submitters to participate like all authors do.
Let me be very clear about this. An author of a candidate build system solution for Boost should be willing to accept responsibility for a core component of our infrastructure. He should be willing to become part of the community and embrace the practices we take, including the reviews. The author should be willing to support the use cases we have in 100+ libraries and also provide long-term support for the solution in the future, should it be accepted. If an author is not willing to participate in technical discussions about his submission from the very start then I don't want to waste my time on reviewing it, let alone using it. If an author submits a solution with no intention to support it then I'm not interested in it. If this means no CMake submissions at all then so be it - I would rather have zero CMake support in Boost than a half-baked unsupported solution.
Absolutely.
My idea was that making it more attractive for potential submitters, we might draw more submissions than we otherwise might and hence end up with a better one.
That's part of my argument - you're trying to attract submitters that would normally not want to make the submission and potentially don't want to be part of Boost. At least, that's what the document draft says. This is bad for Boost and the potential submitters.