And against a very specific context and background I outlined in my blog post. When I joined the fediverse, a @cwebber vision of a Fantasary was still in the cards, which I see less clearly today.
FYI I received some great review and feedback from @nicol on fedi timelines, where I reacted at length. 
What do you mean? I only said that there is more than one active maintainer, and this is true. I don’t understand why you’re trying to discard contributions of other maintainers.
Not all contributions are in the form of commits. Governance decisions are an important (and time-consuming) part of the facilitator’s work, and these require at least two participants. Commits are a wrong metric anyway, because, again, the vast majority of commits attributed to me are updates of my own FEPs, and CI commits.
Hopefully, all concerns are addressed by now.
Just to be clear, I’m not especially concerned about it, but I understand why somebody might be. The FEP repo has many forks, so even if there is one active maintainer of the official fork and they decide to move on without an active replacement, the content is not permanently lost and the same or modified process could be re-hosted by some other organization.
Yes, I think analyzing worst-case scenarios is an good approach to evaluating sustainability.
The scenario where the repository is completely abandoned doesn’t seem realistic because in order for that to happen all four facilitators must suddenly disappear.
What would happen if two most active facilitators suddenly disappear? The process will probably slow down significantly, because the remaining facilitators will not be merging PRs as quickly as they are merged now. But it will keep going: ultimately it’s a collection of markdown documents, there are no meta-level issues that need to be tackled ASAP and automation scripts ‘just work’.
If this repository has any value at all, eventually someone will volunteer to become a facilitator. Maybe it will be someone from Mastodon team - they are well-funded and have a number of FEPs that need to be updated.
So, in my opinion, there is nothing to worry about.
That wouldn’t be my preference. It seems like the SocialCG is publishing “reports” or “notes” (maybe someone can provide the correct term for this) that look very similar in intent to FEPs but may be considered more “official” than FEPs by some developers (and funding agencies).
For what it’s worth, if silverpill ever disappeared for any reason and I was still a facilitator, I would be able to merge PRs and so on. Right now, I would say that PRs get merged by silverpill before me, not necessarily because silverpill is paying more attention to the repository, but moreso because a submitted FEP generally only needs 1 person to address it. If silverpill merges a PR and I see it 15 minutes later, then there’s nothing for me to do. It’s not like the FEP process is being flooded with submissions, nor does it place any onerous requirements on facilitators. It’s a kind of minimal outreach for people to publish a document and invite comment from others. The FEP process has mostly shied away from any substantial governance role, perhaps simply because no one would want FEPs to be so authoritative. The work at the W3C is nominally more authoritative because the W3C as an organization commits itself to maintaining various specifications and documents, and people can generally expect the W3C to continue to exist and do that, but that’s really a different model than FEPs, mainly differing in formality than anything else. Anyone can submit a FEP, but the W3C places a higher level of requirements on anyone contributing to their documents (like patent releases and the whole CG/WG structure which are supposed to represent more than one interest and involve some kind of consensus). At the end of the day, anyone can choose to adopt or not adopt a FEP or a TR or whatever.
I don’t think there are that many funding agencies running around, though. Maybe a handful of them? And having something published via the W3C’s process doesn’t necessarily invite funding, but I guess it could appeal to authority for some people. The funny thing about authority, though, is that it depends almost entirely on other people. If people don’t accept a certain authority, then it isn’t worth much. Reaching consensus is also way more nebulous, since people can generally disagree with anything. This is why enforcing authority is often needed to get people to follow that authority, if they won’t follow it on their own. And then you get to talk about things like coercion and pressure and network effects and so on.
I’m not sure what else to say about this thread since I honestly am having a hard time following what it’s even about, but judging by the title “Grassroots open standards for fediverse evolution” and the article summary presented in the 1st post, I would say that:
- Protocol decay is real but neither FEP nor W3C can solve that. This is a problem of implementers and ecosystems.
- Standards and the standardization process are a social dynamic in their own right, since they involve getting people to agree to do things in a certain way. I’m not sure what “open” or “unfit” are supposed to mean here, since what’s happening right now is entirely the natural outcome of what came before it.
- “Recentralization risk” could be something to watch out for, but out of the 4 things mentioned (server instances, app platforms, the FEP process, and W3C SocialCG) I would really only be concerned about app platforms, since if any particular platform gets too big, then standards become less relevant since the dominant platform can simply declare “I am the spec”. Comparatively, I would not worry about FEP becoming a centralization point at all, and in fact I think FEP is least susceptible to this. W3C is only a concern if people explicitly give it central powers, for example by centralizing all extensibility under a central “extension process” that involves a central registry of allowed terms and their meanings; currently the activitystreams context is not subject to such a threat, although an “extension process” has been proposed at https://swicg.github.io/extensions-policy/ (but not meaningfully accepted).
- It is possible for governance groups to create their own context and define an extension process for themselves, but I think the activitystreams context should be “neutral ground” so to speak, as a way to prevent the semantic foundation from suddenly shifting from underneath anyone depending on it. (I am personally advocating for better support for versioning and profiling, to avoid ending up in a situation where “activitystreams” starts to centrally govern extensions.)
- Re: “top-down protocol design”, I think that the lack of a consensus governing body means that this mode of standardization is not likely to get very far. Re: “evolve from the bottom up”, this also won’t get very far if everyone operates under the shared delusion that we have all already reached consensus and everyone already agrees on everything.
I think Arnold alludes to this in the “trigger” discussion, although perhaps more pessimistically (and I assume this is borne out of frustration at the intractability of the problem more than anything). It’s a hard problem to get people to agree. I’m not entirely sure how silverpill saw it as “an attempt to discredit FEP process”, but this reaction seems to mostly be spurred by Arnold concluding in the blog post that it is “unsustainable”, using a hypothetical situation where all facilitators “burn out and leave” – which I personally don’t think is a real risk, as discussed ad nauseum by others above.
If there is a perceived threat of corporate capture by “professional” “business people”, then this is what we should combat. There’s a limit to what can be done as long as everyone uses a W3C spec, and this is where W3C becomes a concern if an attempt to subvert it ever succeeds, but the only way to preempt that is to develop a framework for protocol negotiation.
Again, as a trial run of this, fleshing out versioning and profiling support for AS2 as a content format is the first step IMO, because if you can break the delusion that everyone will be doing the same thing and doing it forever, only then you can start to evolve. Without that, you have to invent an entirely new media type just to change anything. Or you just make a lot of assumptions that can easily be (and often are) wrong. Last time a protocol stack was migrated, it was Mastodon going from OStatus to ActivityPub, and I fear that there is no recognition of any possible future where it may be necessary to migrate off of “ActivityPub”. Worse yet, there is no recognition of the possibility that “ActivityPub” can change, or that it can mean different things to different people at different points in time.
3 Likes
Herding of cats chaotic grassroots environment, with very specific social dynamics. Organizing things in this complex developer/fedizen ecosystem is the theme of my lengthy blog post. And, as I tried to explain above and on 🫂 Commons custodians. Help increase FEP Sustainability and progress, which I spun off from this topic, SX defines “sustainability” in utter holistic fashion. I am a bit disappointed in the impatient and defensive reaction of Silverpill, and the distrust in accepting me on my word. If anything that constitutes a sustainability risk, the ability to communicate well with ecosystem participants. But once again, the subject is not on operational things, but vision, strategy and future of fediverse and direction, and also perhaps divergence from original power + promise of AP. Only praise for Silverpill, zero blame.
And that is a concept of SX itself: You can’t blame people for their contribution and how they go about their work as volunteers, and should be thankful of each 2 cents of value someone adds. If shortcomings are perceived one can offer kind advice, help and support. And jump into the fray yourself and proactively be the change, by your own efforts.
Also: You can’t expect anything from FOSS other than software code and an open license. The rest is social coding and must be either mutually well-understood, or are implied and perhaps unwarranted expectations. The ones that lead to all the drama between devs and “users” re: privilege, entitlement, and (BDFL) governance + ownership.
Under SX there’s zero blame to Mastodon, for instance, for de-facto turning fediverse in a microblogging-based social network. Their FOSS project. Their app. Their choice. It is the role of a healthy Grassroots standardization process to avoid that from happening, if that is not aligned to shared (technology) vision.
Not at all. This too is explained in my blog post, which I can’t blame anyone from not reading in full, given its length 
I describe this at the start of the Fediverse tomorrow section.
Thanks for your feedback. It’s exactly all these things that SX is seeking solutions for, and come up in various ways throughout the article. I hope to be able to sustain myself, so I can continue to give more shape and form to my applied research, which now happens on Hobby track, lacking sustenance / income.
Related? charters/stage-process.md at 5c400eefdd4b1e899f142d6396699f7fe2882fef · swicg/charters · GitHub
1 Like
That is very interesting @stevebate and yes, I think totally related. A very good initiative by @bumblefudge and @samuelgoto that I heartily applaud. I will take a deeper dive in this, thank you!