!VyoUhyWvlhSpFWWxHL:matrix.org

Zulip setup coordination

92 Members
Coordination to setup https://nixpkgs.zulipchat.com/, see https://github.com/NixOS/foundation/issues/14335 Servers

Load older messages


SenderMessageTime
20 Feb 2024
@apcodes:matrix.org@apcodes:matrix.org * but this room is called "Nix Platform Governance" so... 21:48:11
@7c6f434c:nitro.chat7c6f434cAnd project-scale actually-shared goals are exactly what we have always failed at.21:49:14
@7c6f434c:nitro.chat7c6f434c

Which raises the question what causes them to dis-align in some cases so that these interactions become a "search for a least-disliked outcome".

And also this — this is not disalignment, RFCs figure out time and time again that not everyone sees the proposed changes as useful, so the changes need to be made less harmful at least.

21:50:45
@7c6f434c:nitro.chat7c6f434cIt's not dysfunction, it's actually acknowledging and solving the real question.21:51:16
@apcodes:matrix.org@apcodes:matrix.org but if my goals and your goals are fully contradictory, then making things less and less harmful to both will end up doing nothing at all and no one gets anything. So complete inaction will be the result. However if our goals are actually contradictory, that we can only know if we acknowledge what our goals are. 21:52:57
@apcodes:matrix.org@apcodes:matrix.organyways, have a good night everyone!21:53:04
21 Feb 2024
@tomberek:matrix.orgtomberekI'm surprised by this. We are discussing governance, a thing that only makes sense when trying to manage limited resources; maintainer attention, money, contributors time, official status, infrastructure, so forth. Having a guiding principle helps when tradeoffs need to be judged. I agree that we have often failed at project-scale goals. Perhaps exactly because we've not set any, or talked about them. I dont mean that if we establish such goals that they are to the exclusion of everything else. My optimism for the benefit of being more explicit about vision+goals is that it seems to help get things done and aid in predictability. Already a few teams have started to write down what they are trying to accomplish and publish them, should they stop? Perhaps they are focusing on the wrong things, how would we know? Jonas has established and communicated a personal goal of unblocking efforts, that is way more effective due to knowing about it. Knowing the NAT was going to work on one project at a time managed expectations. Knowing docs would be overhauled was beneficial. Knowing that the S3 cache needed dedicated time to fund, coordinate, and work on was useful. Marketing got contributors by communicating a goal of overhauling the website implementation. Nixpkgs commits to two NixOS releases each year. If we explicitly want to set no direction for the ecosystem, even that statement would help make it clear for people to not expect project-level things to get done/prioritized - help manage those expectations.07:05:05
@7c6f434c:nitro.chat7c6f434c

Perhaps they are focusing on the wrong things, how would we know?

They are. And they are not. Whatever they are doing are right things for them and wrong things for someone else. Just not wrong enough to actively oppose.

Nixpkgs commits to two NixOS releases each year.

And just to make a point, the releases are a small pure negative in terms of what I use from the project. This is sincere. I just know people actually care about releases existing, and not always mistakenly, so whatever.

not expect project-level things to get done/prioritized

They can get done — if a committed enough team of people who do wnat them can convince the others these things are not too harmful, and then carry out the task.

07:19:50
@apcodes:matrix.org@apcodes:matrix.org

They can get done — if a committed enough team of people who do wnat them can convince the others these things are not too harmful

Which, to reiterate my earlier point, in practice might just mean that nothing can get done because of a lack of higher level goals and a lack of decision making processes that would implement them in case of lack of consensus. So concretely, if you find enough people to make a decent initiative appear very harmful, everything just stalls. This approach appears to me to have a strong tendency to lead to extreme risk aversion since people have a tendency to confuse apparent risk with harm sadly.

Which I think is a shame.

12:34:49
@zimbatm:numtide.com@zimbatm:numtide.comGovernance test; in your view, who can make the decision whenever we allow unfree packages to be built on Hydra or not? It's not urgent, but I think is a good example of something that has been lingering and nobody feels empowered to make a decision ( see https://github.com/NixOS/nixpkgs/issues/83884 ). My interest is not to debate on the question, but what mechanism should be used to reach a conclusion.13:38:12
@apcodes:matrix.org@apcodes:matrix.orgI think that is a really interesting question and a really good example I feel. Do you have some already existing mechanism in mind that could lead to a conclusion? In the debate on GitHub it appears that Graham was of the opinion that "We'd need sign-off from the NixOS Foundation which is responsible for supporting the infrastructure". So the NixOS Foundation would be empowered to decide. However you could make the case that that is maybe not the best arrangement. In particular if the foundation doesn't really seek an active role in micromanaging things. My impression so far was that it doesn't seek such a role. However if we agree that the foundation has the ability to make such decisions, maybe it also has the ability to delegate it's authority to someone else? For instance the infrastructure team.14:42:07
@zimbatm:numtide.com@zimbatm:numtide.comThis idea of local vs global vision is interesting. Even if it's hard to build a globally shared vision, we can at least work to make local visions a bit more coherent. Part of the issue with the RFC process how it's currently structured is that it tends to destroy visions. If you get shepherd involved, then by nature they will want to give their inputs. That's just the nature of humans. And if you end up paired with people with different views, then it creates a weird, incoherent, design by committee, but kinda agreed on output. For smaller things, I would much prefer for people to write up their vision, submit it for approval/rejection, than going trough that grueling process.15:02:19
@k900:0upti.meK900FWIW I don't think the "should we build unfree on Hydra" thing is a good example of a decision that needs governance process to decide15:02:45
@k900:0upti.meK900What it needs is a lawyer to say "it's OK to do this"15:02:57
@k900:0upti.meK900And the foundation is realistically the only entity with the authority to get counsel from a lawyer15:03:36
@zimbatm:numtide.com@zimbatm:numtide.comunfree+redistributable are legal, assuming that the classification is done properly. there is a practical issue of; how do we purge the cache if we discover that the classification is incorrect. this can already happen today if people put the wrong license on a project, but is more likely to happen if we start building unfree+redistributable packages.15:06:59
@k900:0upti.meK900Yeah, see, this is exactly the kind of thing that requires a lawyer to know how to handle correctly15:08:08
@zimbatm:numtide.com@zimbatm:numtide.comEven if you are right, I don't think this is why this issue is being blocked. The reason is that nobody feels empowered to drive this to a conclusion.15:09:14
@apcodes:matrix.org@apcodes:matrix.org

nobody feels empowered to drive this to a conclusion

That makes a lot of sense to me. Which is unfortunate seeing that it is already lingering for 4 years now.

15:11:12
@k900:0upti.meK900 I'd argue that having sign-off from a lawyer would also empower people to drive it to the conclusion 15:12:13
@delroth:delroth.net@delroth:delroth.netif the foundation accepts the legal risks then they don't need a lawyer's signoff :)15:14:27
@delroth:delroth.net@delroth:delroth.net (and I'm saying this half-jokingly but also half-seriously, I don't think this needs a lawyer to be consulted) 15:15:03
@zimbatm:numtide.com@zimbatm:numtide.comAnother blocker is that it's unclear if this would be aligned with the project's values.15:15:30
@delroth:delroth.net@delroth:delroth.netimo there are multiple stakeholders here for actual technical problems (e.g. legal, infra, etc.) but this is also a project direction / ethics / values problem, and that's where we have a decision making gap15:15:50
@apcodes:matrix.org@apcodes:matrix.orgYes, me not being around for that long I'd have some trouble actually explaining what these values are that would decide here. I always had the impression the Nix community is much more pragmatic on open source and free software than others. But what are the ethics / values in this case then? Where can I find them?15:18:56
@zimbatm:numtide.com@zimbatm:numtide.com

The foundation is here to unblock things, so they can be invoked if necessary. Generally speaking, it's good if the community can self-organize and if we can keep the foundation's involvement as an exception.

Do you have some already existing mechanism in mind that could lead to a conclusion?

Some simple decision making framework based on criteria. Like: is the decision local or global? Who needs to be consulted? Who can make a decision?

15:28:23
@zimbatm:numtide.com@zimbatm:numtide.comTo make things simple, we don't have officially stated values :)15:32:23
@zimbatm:numtide.com@zimbatm:numtide.comIn general I would say the considerations are mostly pragmatic15:32:43
@zimbatm:numtide.com@zimbatm:numtide.com K900 ⚡️: what would be a better example in your view? 15:34:47
@k900:0upti.meK900I think two most contentious things I can think of are the code of conduct RFC and the formatting RFC15:35:54

Show newer messages


Back to Room ListRoom Version: 10