On Sun, Oct 8, 2023 at 4:42 PM Peter Dimov via Boost
Vinnie Falco wrote:
I want to be clear here, the practice of "we need to accept it into
Boost
to get users and find out if its any good" to be a harmful idea that carries significant reputational risk. There is no stated policy which describes whether or not this practice is valid for determining acceptance in a review, which is itself a problem (lack of a clear written instructions on how reviewers should evaluate libraries), and the reason that I am having to make these posts.
Boost has never required a library to have users in order to be accepted. If the review process says it's good, it's good.
Reviews from users count more than reviews from non-users, of course, but that's the extent of it.
As for "reputational risk", the reputation we have has been built while using the above policy.
I'm also thinking that we apparently need to have a (written) rule about people questioning the review result without having submitted a review. If you can't be bothered to invest the time to review the library, your opinion on whether it belongs in Boost doesn't count.
Or stated differently, the proper and only way to express your opinion on whether the library belongs in Boost or not is by submitting a review.
That said, it is still valid to express an opinion in principle. I don't feel qualified to write a review, but in principle I think it's valid to vote REJECT due to major design disagreements, as in "this is the wrong solution to the problem", however this requires at least outlining an alternative design. Other than that, if a library is well documented, solves problems reasonably well, and is written with competence and with domain knowledge, then it should be accepted.