Hi all, On Thu, 03 Sep 2026 12:29:05 +0200 Simon Tournier <[email protected]> wrote: > However, sadly it cannot be so simple. :-) The GCD deliberation isn’t > the expression of some mood or feelings at the deliberation time, > instead it’s the result of a deep thought about (1) “Can I live with my > concerns and the flaws I identify” and (2) “Aside my personal > preferences, is the outcome of the GCD at hand a good direction for the > project to move forward?”.
As an outside observer, I find it really odd that the GCD process is set up in a way that allows a single person to completely torpedo a ~2 month process. I understand and support the want to avoid "tyranny of the majority", but I think similarly there is also issues with granting an unchecked veto, even if there is some filtering in place of only maintainers being able to vote. To me this seems like an easy way to potentially deadlock anything, while staying de facto blameless, since you're just exercising your rights after all. The author of the GCD has a ton of responsibility during the process, while a process like this should be as cooperative as possible. If it were up to me, I'd have two non-supporting outcomes: - A sizable minority vetos the GCD. Can be 5% of all maintainers, can be 10%, or some other ratio. I'm not good with such numbers and the exact doesn't really matter for this discussion anyway. The point is, if every tenth or even twentieth maintainer thinks "This would make the project markedly worse and I genuinely don't see any way to change the GCD into a version that I could support with a good conscience", then it's probably not a viable route. If a single person thinks so, then that is still very unfortunate, but if GCD not passing would make even more people leave, then you're between a rock and an an even harder place. Sadly, I don't think there is a solution here that avoids some disappointment. Obviously, for such a veto to be valid, people would need to provide justification, because otherwise would only lead to follow-up questions and potential annoyance. It'd also be their responsibility to adequately prove, that the author hasn't addressed their concern(s) or that the GCD is fundamentally incapable of aligning with what they consider to be the project's best interests. Vetoing could only happen in the deliberation stage after the text is finalized, because until that point is reached, the GCD could still change and people could still change their minds about it. A vetoed GCD could not be resubmitted without substantial changes and it'd be up to the author to argue why their new attempt is different enough to perhaps sway those who blocked it. - A quorum of maintainers could request changes and postpone deliberation. I guess this could happen a week or two before deliberation is supposed to start, people could object by saying "it needs changes because X Y Z" and if the quorum is reached, then the deadline is extended and the community focuses on negotiating these topics, with other conversation discouraged / moderated away. Votes without justification would once again be discarded and it is on the voters to provide adequate evidence that the author hasn't addressed their claims yet. Change-voters are expected to participate in the negotiations, otherwise they're disqualified from the next vote. Then once this new deadline nears, the vote restarts and people can either object again or assent (by staying silent). This could then potentially continue for multiple rounds (with a reasonable limit to stop deadlocks). To not drag out things for several months, the discussion rounds (after the first one which would be longer) could be shortened to ~2 weeks or similar, allowing already good proposals to be voted on quickly, while providing multiple opportunities of tight feedback loops for those that need work. Then at the end the final vote happens, where the GCD is either accepted (enough voters for, no veto), withdrawn to be presented again later(not enough voters either way), or vetoed (veto limit is hit). To make this process work (but frankly even if you disregard everything from before, because it's just the ramblings of someone not involved all that closely with the project, this might still prove useful on its own), I think GCDs should have an additional role in the form of moderators. This would be someone who swears impartiality and works to keep discussions in the right direction, warning off-topic conversations, conducting votes. In short, this person would act as a quasi-"secretary" for the author and allowing them to focus on only the actual negotiations. I hope some of these thoughts might prove helpful, if nothing else, as one data point out of many. Take care and thank you for reading, Nemin
