Re: Informal Discussion: Identities of Voters Casting a Particular Ballot are No Longer Public
I hear where people are coming from, when they talk about not wanting to bundle things, but do not plan to conduct multiple votes. Fortunately, especially under the constitutional amendment we just passed, others who want us to act differently have the flexibility to argue for that. One of the things I've learned being in Debian for over 20 years is that agreeing on the question is sometimes harder than agreeing on the answer. Whether something is "bundled" or not depends on how you view the problem. I think the best example of this was the init systems discussion within the TC, although it was clear that during several GRs we never did come to agreement on what question we were voting on. IN this instance, I consider the secretary changes sufficiently related to the secret vote changes that I don't consider them bundled. Also, given the DPL's concerns about the number of GRs that are queued, I'd rather not have more votes than we need. I also believe that what I'm is consistent with what we've done in the past. Russ's proposal, which we just passed, included changes both to the TC voting process and to the GR voting process. We chose to vote on them all at once because they were related. In my mind the changes are related enough that it might affect how I rank them. It's also a reasonable position to view the secretary changes as seperable and even to argue about whether the secretary changes or the secret ballot changes should happen first. It's even reasonable to argue about whether removing the requirement that votes be conducted via email is a third separable option. And you could even disagree on the order of all three of these potentially independent votes. If you would like to see things unbundled, you have a few options: Once there is a formal GR on the table, you could: 1) propose and unbundled option. For example, if you think we should vote on the secretary changes first and you like them, you could propose an option that includes the secretary changes without the secret ballot changes. That option would also be appealing to people who like the secretary changes but who never want to see the secret ballot changes pass. You might think that's great. Or you might want to explicitly add text to your option saying that you think we should vote on secret ballots later, so that if your option wins, people don't think we'vedecided against secret ballots. 2) If you don't want to see things intermingled on the same ballot, you could propose an option explaining what order you think we should vote on. Something like "The Debian project believes these issues should be decided in separate votes. We should first decide on whether to have a mechanism for overriding the secretary and then decide on whether to have secret ballots." Voters will then get to choose whether they want to get it all over with at wonce or whether they want to handle things separately. I think that's the best way we can do given that we have historically found it next to impossible to agree on what question we are asking or what order to ask them in. --Sam
Re: Informal Discussion: Identities of Voters Casting a Particular Ballot are No Longer Public
> > I've also got concerns about batching up unrelated changes, with > > potentially controversial ones. And even if minor I'd prefer to see > > those debundled, even at the cost of additional GRs. > > If the only contentious point is the secrecy of votes, we could have an > amendment that includes all the other proposed changes, minus that one. I also agree with those, who speak against bundling. I'd like to be able to vote for something without needing to additionally vote for something else that is unrelated and I might even not want. If something is important enough to change the constitution for it, it's also important enough to have a separate vote on it. signature.asc Description: This is a digitally signed message part.
Re: Informal Discussion: Identities of Voters Casting a Particular Ballot are No Longer Public
On Fri, Feb 18, 2022 at 01:13:50PM +0100, Guillem Jover wrote: > On Sun, 2022-02-13 at 14:28:44 -0700, Sam Hartman wrote: > > This starts informal discussion of a proposed general resolution to > > amend the constitution. I am not seeking sponsors at this time. > > Comments including support or alternatives are welcome. I think this is > > mature enough to seek review from the secretary. > > Since the idea of general secret votes has started floating around > I've felt a sense of uneasiness, but I've not been able to clearly > put my finger on why. After some of the replies here, I think it's > starting to become clear. Thanks for this, you raise interesting points. > I think the current secrecy for the DPL votes makes absolute sense, > and I think there's no contention about that one, because these are > about voting "for/against" people, which have clear and understood > social dynamics when it applies to colleagues/friends or people we do > work with etc. I think, thus, extending secrecy to any vote related > to "people" would also be equally uncontroversial. > > Then, there's the secrecy for technical votes, which I think is where > the push back might be coming from. There's been mentions of mailing > lists being way more revealing than a vote in GR, and counters to that > mentioning that you do not need to participate in mailing lists. Both > true. The problem I think, is that to participate in Debian in any > technical role, you most definitely need to eventually make your > opinion on technical matters public, because we operate on the open. > Be that on bug reports, on changelogs, on VCS commits, or even on > mailing lists. It also feels like closing up technical votes would go > counter to the general tenets of the project and how we operate. > > And then, there's the secrecy for "political" votes. I think this > might also be problematic, depending on the subject at hand. Because > as mentioned in the thread, it might make public positions that people > otherwise would not need to make so on their daily routines in Debian. > > I think the RMS vote, was a mix of personal + political, which is what > made people uncomfortable with. The problem I see is that this is now > being lumped into a general direction to close everything up, which > seems excessive, TBH. > > I also think the DPL votes are different to any other votes, because > the DPL has limited power, and even though a DPL can certainly disrupt > or damage the project, in the end it's bound by a time limit. Compared > to a GR where the consequences might live long, and where once settled > people do not tend to try to overturn these every subsequent year. By the time something comes to a vote, be it about a technical issue or anything else, all the relevant arguments were already made in public, and sometimes to exhaustion. The voting itself at the end of the process is "just" a poll to decide which side (if any) of the argument convinced the most people. It doesn't seem to me that having the votes be secret means we are hiding anything, or compromising on our principles of transparency in any way. > I've also got concerns about batching up unrelated changes, with > potentially controversial ones. And even if minor I'd prefer to see > those debundled, even at the cost of additional GRs. If the only contentious point is the secrecy of votes, we could have an amendment that includes all the other proposed changes, minus that one. signature.asc Description: PGP signature
Re: FYI, Secret Ballots Proposal is Likely to Die for Lack of Support
On Wed, Feb 16, 2022 at 03:34:09PM -0700, Sam Hartman wrote: > > My take is there's not currently enough support on debian-vote to bring > it to a vote. > I'd want to see several additional people express support on debian-vote > before I'd feel comfortable proposing a GR. Hi Sam, thank you very much for your time spent on this issue. I really would like all votes to be anonymous in principle. Last year I blogged about the possibility to crunch our vote data and cluster DDs by affinity. While I did not manage to do it myself, I am uncomfortable with the idea that we provide machine-readable data that makes this kind of approach likely to succeed one day. http://charles.plessy.org/Debian/debi%C3%A2neries/DebianAnalytica/ Maybe the GR does not need to go deeply on the technical details and can simply state the principles, define who choses the implementation, and to what extend the final process can deviate from perfection. I do not mind making a couple of concessions with practical challenges, and if a few people have need to transiently access deanonymised data to prove that there was no fraud, no problem. I think that after enough time for the results to be accepted, the deanonymised data should be just deleted. I think that we can also remove the tally sheets from the past votes from our public servers. Sure, they will stay forever in the Internet Archive, but the point is that the amount, kind and scope of the data that we publish ourselves should match our view about pricacy. Not sure if we really have a consensus but I think that we should refrain from publishing personal data that does not serve a significant purpose. Have a nice week-end, -- Charles
Re: Informal Discussion: Identities of Voters Casting a Particular Ballot are No Longer Public
On Sun, 2022-02-13 at 14:28:44 -0700, Sam Hartman wrote: > This starts informal discussion of a proposed general resolution to > amend the constitution. I am not seeking sponsors at this time. > Comments including support or alternatives are welcome. I think this is > mature enough to seek review from the secretary. Since the idea of general secret votes has started floating around I've felt a sense of uneasiness, but I've not been able to clearly put my finger on why. After some of the replies here, I think it's starting to become clear. I think the current secrecy for the DPL votes makes absolute sense, and I think there's no contention about that one, because these are about voting "for/against" people, which have clear and understood social dynamics when it applies to colleagues/friends or people we do work with etc. I think, thus, extending secrecy to any vote related to "people" would also be equally uncontroversial. Then, there's the secrecy for technical votes, which I think is where the push back might be coming from. There's been mentions of mailing lists being way more revealing than a vote in GR, and counters to that mentioning that you do not need to participate in mailing lists. Both true. The problem I think, is that to participate in Debian in any technical role, you most definitely need to eventually make your opinion on technical matters public, because we operate on the open. Be that on bug reports, on changelogs, on VCS commits, or even on mailing lists. It also feels like closing up technical votes would go counter to the general tenets of the project and how we operate. And then, there's the secrecy for "political" votes. I think this might also be problematic, depending on the subject at hand. Because as mentioned in the thread, it might make public positions that people otherwise would not need to make so on their daily routines in Debian. I think the RMS vote, was a mix of personal + political, which is what made people uncomfortable with. The problem I see is that this is now being lumped into a general direction to close everything up, which seems excessive, TBH. I also think the DPL votes are different to any other votes, because the DPL has limited power, and even though a DPL can certainly disrupt or damage the project, in the end it's bound by a time limit. Compared to a GR where the consequences might live long, and where once settled people do not tend to try to overturn these every subsequent year. I've also got concerns about batching up unrelated changes, with potentially controversial ones. And even if minor I'd prefer to see those debundled, even at the cost of additional GRs. Thanks, Guillem
Re: FYI, Secret Ballots Proposal is Likely to Die for Lack of Support
On Wed, 16 Feb 2022 15:34:09 -0700 Sam Hartman wrote: > I'd want to see several additional people express support on > debian-vote before I'd feel comfortable proposing a GR. Sam, I was unaware of the discussion taking place but very much support your efforts towards secret ballots. Every voter should feel free to express their opinion without fear of reprisals. In the real world that often requires secret ballots, as demonstrated by the election process in numerous democratic countries. Debian is no exception: for the RMS GR I ended up spending more time considering whether or not to actually cast my vote than it took me to make up my mind about the options on the ballot. pgpE25sg4MexE.pgp Description: OpenPGP digital signature