tl;dr: Relax the requirement that sponsors be directly involved in the package review process.
Sponsors are responsible (but not solely responsible) for shepherding people through the packaging process. They should know how to do reviews, but there is nothing so special about a packager's first review that it cannot be handled by the regular packager community. We trust packagers to do every other package review, after all. We also allow new packagers to be sponsored without actually going through the package review process at all via the comaintainer process so what we appear to be emphasizing is that someone is there to assist and monitor the new contributor and not that the contributor make it through the arduous process of a package review with a highly restricted pool of reviewers.
Proposal: Decouple sponsorship from the review process.
Allow the community to do reviews as normal. Remove the requirement that the first review be done by a sponsor.
Emphasize that sponsors can sponsor anyone separate from the review process. They can sponsor them before the review has been done, after it has been done, in the middle of the process, whatever. (This is all currently true in any case, but the process documents link most everything to the completion of a review.)
Notes: Obviously sponsorship should still be tied to package maintenance in some way; sponsoring someone without any intention of having them work on a package in some way is pointless.
Note that I do not intend to imply that sponsors need not know how to do proper package reviews. The guidelines for becoming a sponsor currently and should continue to specify that having done some package reviews is important to the process. The same goes for actually maintaining packages. Sponsors should know both the mechanics of maintaining packages and the standards for package quality.
Hopefully this will open up the actual reviewing to the community as a whole, eliminating one bottleneck.
We could potentially end up with people who have completed package reviews but who cannot yet actually import their packages. This would be worse than having people waiting in the sponsorship queue, because they actually did more work and someone from the community actually did some work as well. This could be mitigated through vigilance coupled with some scripting, or additional process in the packager-sponsor trac for requests that happen to fall through the cracks.
Searching bugzilla for NEEDSPONSOR tickets still open with fedora-review+ set should be a reasonable first-pass report for those waiting. Mailing a filtered version of that to the sponsors would probably be effective but annoying.
I have little to add here but my +1. I believe that anything we can do to reduce the initial packaging hurdles is a net benefit to Fedora. Even if the first pass of a package is less-than-perfect, people learn over time and are far more likely to stick with a process that lets them do that than one that simply demands perfection out of the gate.
Oh, I have plenty of other ideas for making things better, but this seemed to be an easily solved issue that does occasionally get in the way. And when it does get in the way (i.e. at the end of the process everyone realizes that the review was done by the wrong person) it really sucks for both the new contributor and the person who wasted their time doing a review.
In the future I really would like to see reviews out of bugzilla and into some bespoke software with everything in a pagure repo, in order to make things work with a swarm of drive-by reviewers firing off pull requests and dropping karma. That more fits the "I have ten minutes" state that many potential reviewers find themselves in. I would also like to see a procedure for packages maintained in copr to be nominated for inclusion into the main distro. But all of that really is going to have to wait until some software is written.
tl;dr: Using copr is also feasible if sponsor is dispensable, there is no obligation that everyone needs to become a packager. I've seen several people got sponsored in the past 3 years, then disappeared after 1 or 2 years.
IMO There is no bottleneck here, I'm now thinking Fedora Community is loosing strict policies and standards to stop losing the market quotient. If you believe most of sponsors are just running out of time, and they need help from others, then I believe there are vacancies of sponsors.
I did a quick look through over recent reviews, some got sponsored quickly, for example 'haxe', while some others are not. And, there is a little amount of packages queued for sponsors recently are just slags, eithor from years ago even without homepage, or nonsensical.
First review is significant for newcomers, not all people are packaging gurus, most of them just know how to make things work, but not make things exquisite or brilliant, sponsors should perform first review with help from others if possible, but not solely from others.
Well, all I can say is that I've worked in this space for a very long time (in Fedora terms, at least) and that my experience leads me to disagree. The packager community is just as able to provide a reasonable review for someone's first package as their fortieth. I do believe that often the general standard of review quality isn't as high as I would like for it to be, but I don't believe that there's all that much differentiation between sponsors and non-sponsors there.
I did not intend my proposal to have any bearing on people leaving the community. I believe that is a completely orthogonal issue.
And this has nothing to do with "market quotient" or anything like that; those who know me would understand that I wouldn't really be motivated by such a thing. This was instead motivated by specific experiences I have had, and by my general state where I have time to assist people and provide sponsorship but not always the large block of time required to complete a proper package review.
In any case, I don't intend to put up much more of an argument here, and submit to the will of the committee.
+1
If the proposal is just that not to have mandatory requirement of having Sponsor to review the first package of new contributor then I am +1 to this proposal.
But I don't like to see the Sponsor sponsoring before finishing that first package review. It should be mandatory that whenever first package review and some peer package reviews gets completed then only that new contributor should receive sponsorship. Are we thinking that we should also not emphasize on peer reviews done by new contributor?
One case here I can think is that if the new contributor has pro-actively done some peer package reviews and submitted few new packages for review then Sponsor can sponsor him before his any package gets reviewed.
The things you believe should be mandatory aren't currently mandatory, or even required at all. Sponsors can sponsor who they like, whenever they like. That has always been the case.
I would personally be against what you propose, but it's really unrelated to the proposal I put forward. My proposal is only related to who must do the first review.
Replying to [comment:7 tibbs]:
Well, I have been following this since many years to make sure new contributor is knowing packaging guidelines, has submitted packages which he is going to maintain, done peer reviews to help with our "New packages awaiting for their reviews" queue and then sponsor him.
I see I can't take this point for further discussion as mandatory word is already not written in [https://fedoraproject.org/wiki/How_to_get_sponsored_into_the_packager_group#Convincing_someone_to_sponsor_you wiki page].
My proposal is only related to who must do the first review.
I thought so but then you have written more information below the proposal line which confused me.
+1 to your proposal.
From today's FESCo meeting:
I have written information about this change into the following wiki pages:
I have also added a bit of process to the packager-sponsors trac to assist with the case where someone has accepted packages but has not yet been sponsor.
Please let me know if I have missed any other pages which reference the removed requirement.
Login to comment on this ticket.