Re: [boost] [Async] Naming land grab?

On 09/10/2023 10:12, Hans Dembinski via Boost wrote:

...

On 9. Oct 2023, at 09:27, oliver.kowalke--- via Boost wrote:

...

This is an alpha library and should come back when its >place in the ecosystem has solidified. I disagree because this restriction was not required in the past. It seems to me unfair to make it mandatory now. I am using Oliver's comment to add my own.

Klemens obviously has a good amount of stars and forks, which signals interest from the community, so he does not need me to defend his proposal. But I also want to give my option on "should libraries proposed for adoption have many users already?".

Whether or not a library has a lot of users before submission is fairly irrelevant. It can be a point in favour, but does not have to be. It may address the question "is this library useful"? but does not tell whether the library is well designed. Some people need a lot of user feedback to refine the design, others are good internal critics and can improve without external feedback. Both are valid approaches to design and suit different people. Obviously, everyone needs some external feedback, but it is quality and not quantity that matters.

Also, we all know from politics and examples in tech, that popularity is not always a good measure of quality.

Usefulness, design, quality, etc are judged by the reviewers. I have high confidence in the Boost reviewers to make this assessment. Users do not get to vote on acceptance, Boost members do. Boost is a meritocracy, not a democracy.

+1.  Also just to add to that, existing libraries with wide user bases come with a legacy that may may it harder for them to make radical improvements to the user interface should that be required.  On the other hand brilliant interface design with no user experience also carries it's risks.  This is what reviews are for, to balance all this up, and it's good to see some discussion being generated again. It's also worth pointing out that Boost when founded was a place for radical bleeding edge design and experimentation, I see no reason for it not still be so as long as such libraries are well documented as such. With apologies, I'll now go back to lurking again, as I have no knowledge of the problem domain... John.