Problem/Motivation
- Core has many subsystems, including many that have too few maintainers.
- We can support subsystem maintainers by responding quickly when they volunteer to take on responsibility for a subsystem, and respect their efforts and time commitments by responding quickly when they choose to step down.
- The BDFL cannot always commit subsystem maintainer patches in a timely fashion.
- Core committers can generally assess a potential subsystem maintainer's contributions, agreement to the governance policy, and whether they have community support.
Proposed resolution
- Allow any committer to add subsystem maintainers, so long as the other maintainers for the subsystem and the community agree.
- Allow any committer to remove a subsystem maintainer at that maintainer's own request when they choose to step down.
- Per Dries, maintainer additions that do not have consensus still should be escalated to the BDFL, as should adding topic maintainers, initiative coordinators, or committers, as well as removing maintainers when it is not at their specific request.
- As with all areas of governance, the BDFL may also still overrule any decision.
Proposed additions to the core governance
If the change is adding additional maintainers, other maintainers for that area must be given opportunity to sign off.
If the change involves adding a new subsystem maintainer, a product, framework, or release manager must approve or refuse it (depending on the subsystem). (Any may commit a change to remove a subsystem maintainer at that maintainer's own request.)
If the change is adding new full or provisional core committers, initiative coordinators, or topic maintainers, the BDFL must approve or refuse it. The same is true if there is any cause for concern with adding a maintainer, or if the change involves removing a maintainer except at their specific request.
Comment | File | Size | Author |
---|---|---|---|
#14 | interdiff.txt | 1.83 KB | xjm |
#14 | governance-subsystem-maintainer-14.patch | 2.91 KB | xjm |
#9 | interdiff.txt | 2.27 KB | xjm |
#9 | governance-subsystem-maintainer-8.patch | 2.95 KB | xjm |
#7 | interdiff.txt | 2.27 KB | xjm |
Comments
Comment #2
xjmNow with less run-on.
Comment #3
xjm(Adding the proposed additions as text for easier review.)
Comment #4
xjmAnd fixing a typo ("maintianers").
Comment #5
xjmComment #6
xjmPer @webchick's suggestion, trying to split the latter bullet into two.
Comment #7
xjmJust fixing a closing bullet.
Comment #8
xjmNote that I also removed the repetition that the BDFL may overrule any particular decision, since the BDFL section already states that clearly.
Comment #9
xjmFixing another typo.
Comment #10
webchickLooks good to me! Would love at least one other person who wasn't part of that conversation to eyeball this. :)
Comment #11
jhodgdon+1. I have not been involved in this conversation, but the changes seem like a Good Idea (TM) to me -- we want systems to be Maintained (TM) and waiting for Dries to have time to review/commit changes (especially if the person has already been acting like a maintainer for a while and other committers know it) seems silly. The provision for existing maintainers to sign off also seems sane, since you wouldn't want to add a maintainer if it would cause the existing ones to run away screaming or something.
I also think that the updated text is very clear in the latest patch.
Comment #12
jhodgdonSorry about all the TMs. Well, not really very sorry. ;)
Comment #13
tim.plunkettAs the only non-committer to comment so far, I'm +1 on this change.
Just thought someone else should chime in, otherwise it's like Congress voting to give themselves a raise :)
Comment #14
xjm(/me files a patch for all committers to be given chocolates for New Year's.)
I actually realized we can probably just say:
...since in all cases it's potentially kind of the same as a "significant" change for that area and Dries takes into consideration the good functioning of each team. Plus, fewer words.
Leaving at RTBC since I think this is just eliminating a potential misunderstanding/concern about other roles not having opportunity to sign off.
Comment #15
Crell CreditAttribution: Crell at Palantir.net commentedI guess that's the other Routing subsystem maintainer giving +1. :-)
Comment #16
Dries CreditAttribution: Dries commentedThis looks great (per our discussion). Thanks @xjm. Committed!