On Wed, 29 Jan 2025 21:33:21 +0100 Michael Niedermayer <mich...@niedermayer.cc> wrote: > Hi all > > Heres my current "work in progress": (sending that before fosdem, so people > can discuss if they like) > > Goals: > The proposed changes aim to improve the General Assembly's structure to > ensure inclusivity, fairness, and resilience against attacks. The key goals > are as follows: > Increase the Size of the General Assembly > Inclusivity: Allow every contributor to have a vote, ensuring all > voices are heard, regardless of their role or level of involvement. > Enhanced Security: By increasing the number of voters, it becomes > significantly harder for a malicious actor or group to influence decisions. A > larger voting pool dilutes the impact of any single attack or coordinated > effort. > Make Voting Power Proportional to Contributions > Fair Representation: Allocate voting power based on contributions, > ensuring that those who dedicate substantial time and effort to the project > have a stronger voice than those with minimal involvement. This creates a > system where contribution equals influence.
This should be time-gated to count only commits in the recent, say, 3 years (to match the current GA cycle). Counting purely historical commits seems odd. To use a loose analogy to a real-world democracy, it's like suggesting that retirees should have more votes than young adults because they've lived in a country for longer. This is a non sequitur - in reality, it should be, if anything, the exact other way around. Those with a stake in the project's present and future deserve more say than those with a stake purely in its past. > Resilience Against Attacks: Attackers would need not only a large > number of people but also a comparable volume of meaningful contributions to > influence the vote, further safeguarding the project. > Motivating Participation: Encouraging higher levels of engagement by > rewarding contributors with more influence in decision-making. > Broaden the Definition of Contributions > Previously everyone was a software developer. But really there are > many people in the community, who are not software developers. > > > Shares in Alternative P > 1 release == 100 shares > 1 entry in MAINTAINERS == 100 shares > 1 commit in git master branch == 10 shares > 1 subscription in ffmpeg-devel == 10 shares > 1 subscription in ffmpeg-user == 10 shares > 1 fixed ticket in trac == 10 shares > 1 reported issue in trac == 2 shares > 1 mail in ffmpeg-devel == 2 shares > 1 mail in ffmpeg-user == 2 shares > 1 (backported) commit in release branch == 1 shares > > If the condorcet vote software fails due to the number of shares then > all shares shall be divided by 2 before all rounding and non integer > shares shall be rounded to nearest even > this shall be repeated until the vote software no longer fails due to the > number. > > Shares in Alternative F > Everyone who either has authored a commit in git master or sent a mail to > ffmpeg-devel or user > or fixed or reported an issue in trac has exactly the same vote power. > This is a true classical democracy. > It includes the nearly same people as the previous suggestion but without > the > proportionality. It is vulnerable to a group of a few thousand actual > people joining and coordinating > an attack. The proportionality raises the bar for such an attack by ~2 > orders of magnitude. > > We need a list to remap multiple addresses to the same person (this is > not needed for the proportional case) > > Any single company collective vote power is limited to 10%, associated > companies count as the same company here. > > If anyone can show that specific activities are automated then the test used > for detecting them > shall by confirmed by GA vote and then be added to a list of anti-bot tests. > This vote shall be > performed by a GA that is on the closest first january prior to the event > adding the disputed shares. > > the list of mails on ffmpeg-devel and ffmpeg-user should be filtered by the > current subscribers based on the > idea that someone who left by choice does not want to receive vote mails. If > they want to vote they can > re-subcribe > > In all cases, whenever possible decisions should be made by consensus on > ffmpeg-devel. > Voting should only be used when consensus was tried at least twice and failed > > A factor related to last activity will be in a seperate vote > A veto power may be in seperate vote > > -- > Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB > > Some Animals are More Equal Than Others. - George Orwell's book Animal Farm > _______________________________________________ > ffmpeg-devel mailing list > ffmpeg-devel@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-devel > > To unsubscribe, visit link above, or email > ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe". _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-devel To unsubscribe, visit link above, or email ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".
