A rant about social protocols Introduction
Recently, I read an article that talked about that someone, tried to do a new platform called “Content Nation”. This is a German platform that allows people to write content (to be honest, I don’t really know what it does.) and publish it. And recently, the creators tried to implement the ActivityPub protocol. They did so by using the official documentation provided by @w3c.
The problem was that the last time the official documentation was updated, was in 23 January 2018. So, this means that a lot of new standards that other platforms like Mastodon, Misskey, etc... use are not written in there. But this isn’t the fault of the service developers, this is the fault of the W3C that hasn’t been an update to the protocol officially to support the new standards in the industry such as Webfinger, SharedInbox, Privacy Scopes, and Opt-Out for Search…
The thing, is that this led to a lot of people thinking that this site was some kind of scraper and started making the crawler crash or, even worse, someone tried to load CP inside the platform. BlueSky
Recently, BlueSky opened its AT protocol for everyone to use and federate, due to this, there has been a bit of a discussion inside these platforms. This made me think, why did BlueSky feel the necessity to make another protocol? If there is one already, why do we need another one that competes, wasn’t the objective of protocols to allow interoperability?
So, I did a bit of digging and I found two things. The first one is that they wanted so solve a few things that AP does not support officially (here are the main points, not all of them):
Account portability. A person’s online identity should not be owned by corporations with no accountability to their users. With the AT Protocol, you can move your account from one provider to another without losing any of your data or social graph.
Algorithmic choice. Algorithms dictate what we see and who we can reach. We must have control over our algorithms if we're going to trust in our online spaces. The AT Protocol includes an open algorithms mode so users have more control over their experience.
A lot of these problems are already present on ActivityPub for a long time. The account portability of ActivityPub let’s say it’s not intuitive. You have to do a lot of things and even then, there are some things like the posts that you make or the favourites that don’t transfer (in the case of favourites you need to transfer them manually, the same for blocks and mutes).
Also, right now 99% if not all software that uses ActivityPub, does not have an algorithm that orders content for you to see, but shows you everything in chronological order (I don’t know if its intentional or if it’s a limit of AP) and the only thing you have to discover topics is trough hashtags that maybe someone forgot to tag.
Furthermore, not to mention that on ActivityPub, you are at the mercy of the server moderators, so this means that if you know someone that is on an instance that is blocked by yours, you won’t be able to talk to them unless you change the instance, which in a way it’s not very decentralized. The other protocols
By doing research, I realized that there are a lot of other protocols (for example Nostr) that have its own implementation of things maybe there are some that are bridged and other not.
Such protocols have different features, for example Nostr allows you to suggest content edit to other people’s posts, move your content easily, etc. How can we solve this?
First, we have to know why all these other companies make their own. I must say, that most of them probably do because AP does not allow customization of posts or the adding of new features for everyone and the fact that it’s not been updated for 6 whole years makes matters worse.
What the developers want, is a protocol that lets them create wherever they want and add everything the want, for example the edit thing that I said the Nostr supports, the only way to add it to AP, would be or only on your software or find another software that is willing to implement that feature, the rest of the market is left behind as well as the users that depending on what it is, they don’t understand.
My solution to this problem would be to add some kind of per user plugin system directly to the AP that allows for devs to implement add-ons that do with the JSON strings that add buttons or scripts at least to send and receive data. As well as to add some kind of CSS support for the posts and profiles. Of course, the point of these is that if you make a platform, and you are the only one using these characteristics, well… but in case that everybody wants to use it and everybody makes their own plugins it would be chaos.
For this, the solution I proposed would be like something you add while the W3C updates the protocol to support a very popular feature. #socialprotocols#nostr#activitypub#W3C#ATprotocol#rant#blogpost#ContentNation
> this is the fault of the W3C that hasn’t been an update to the protocol officially to support the new standards.
There is no "THE" #W3C that maintains or is responsible for evolving the #ActivityPub related specs. There's the #SocialCG, a community group, with representatives (read: volunteers) from the same grassroots movement that is engaged with the #Fediverse, i.e. fedizens.
@smallcircles@dadifroggie@w3c The W3C effectively owns ActivityPub. If one tries to make a grassroots change to the Recommendation, they will learn that immediately. My understanding is that even the W3C Social CG can’t make changes. A new W3C working group must be created or rechartered to do that.
The w3c owns AP as a spec. Not the SWICG, who is extremely limited in what they can do, and the only way to make true changes to the spec is through the chartering of another working group… which again the w3c controls the process for and everything down to how that group can make recommendations.
Individual implementations have made diverging choices, and there's no clear way to get these pushed up to AP.
One can make the argument that AP is essentially rudderless here because of the way that the w3c works and I won't disagree, but in terms of responsibility and ownership, those fall squarely at the feet of the w3c.
Well, technically yes. Similarish to how notary 'owns' the contract of my home's mortgage loan as a custodian, and I can't just make changes to that.
I'm responsible to make my home livable, keep it tidy. I can refurbish and add value to it. And if I plan well I can convince my bank & notary to adjust my contract, if I stick to the process for that.
Back to W3C. AP is unique as until now there weren't corporate entities driving the standard to its Rec. status.
@smallcircles@hrefna@dadifroggie@w3c I see almost no similarity between a document like the AP Rec and a mortgage contract (other than both being formalized written documents). I think the issue isn’t ownership per se, but what entity is authorized to perform maintenance of the document. For the AP/AS2 Recs, that entity is W3C and only W3C (enforced via copyright and trademark laws). No grassroots entity has that right.
Who's hosting the spec and dictates the improvement process.
If you blame AP for shortcomings.. there ain't people in 2) who do spec improvement. There's no "fault" at W3C. Without WG/CG's they're a group of clerks/custodians only.
In 1) its where it has to come from us, our grassroots movement, and until now all that's there came from us (with blood, sweat and tears).
@smallcircles@hrefna@dadifroggie@w3c For 1), the W3C is the only organization that can charter a working group (WG) to maintain AP/AS2 Recs (hopefully based on community input that has accumulated for 6 years). They have not done so despite many reports of issues in the documents. I disagree that the lack of maintenance is not their “fault”. The 1) legally can’t come from the grassroots*. See 2) for why.
(FEPs and spicy blog articles/toots are not spec improvements)
Nevermind. I think we're talking past each other. Adopt any standards process you want, then adhere to its rules. Still needs individuals coordinating.
There's an amount of formality and rigour to W3C process, and possibly flaws to organization structure (membership model), sure.
But AP Rec didn't come about by W3C saying "Hey, we wanna have AP. Here's a WG. Who's interested?". The W3C only ever says "Here's a standards process. Follow it to create standards".
> W3C is the only organization that can charter a working group (WG)
Yes, it will honour any request that adheres to the process for doing so.
> They have not done so despite many reports of issues in the documents.
Because they aren't responsible for doing so. They manage the process. It is us who manage the evolution of the technology, and make steps through that process (if we choose to adhere to W3C rules set out for them).
I, like most, advocate for a 'technology substrate' with highest chance to a) unleash full potential of fedi/social web, and b) remains open and accessible to anyone. Whatever process that takes.
I feel the process should bottom-up, facilitating a decentralized ecosystem. Should address the tendency in grassroots movements that everything fragments, and fiefdoms form. So it should stimulate collab.
@smallcircles@hrefna@dadifroggie I like the substrate proposal, but it mistakenly puts responsibility for W3C Recommendation into the Social CG box. There should be a grayed-out Social WG box for that (since it has been inactive since 2018 and must be rechartered).
Yes, I did not delve much into the WG considerations and left it out of the diagram above. Until now talk on rechartering a WG took place on the SocialCG mailing list. If/when/how it is established.. 🤷
As a W3C body I suspect a close involvement of SocialCG, but idk how the organization of Stage 3 will evolve.
Same is true for Stage 2. I drew FEP Process in the diagram, because I am one of its representatives. But there can be any other body there, like FediForum.
Of course the appeal for broad collab goes against grassroots social dynamics. @hrefna mentioned something along these lines as well. One can only advocate collab, not expect it.
My "fiefdoms" should not be read negatively, per se. It is logical that in chaotic environments (for the most part well-intended) people want to create orderly spaces, where they feel more in control.
That the standards didn't didn't evolve and the ecosystem evolved chaotically is not directly anyone's fault, it is just that people find it really hard to collaborate within grassroots movements. Stuff happens organically, things are evolve by spontaneous emergence, and if you try to force collaborative action, you'll experience the same difficulty as trying to "herd cats", as the saying goes.
OK, let's say you are right (you have a few replies that don't say so), my point remains the same:
this is the fault of the SocialCG volunteers that hadn’t been an update to the protocol officially to support the new standards.
And to be honest, 6 years for people to collaborate on a protocol is a lot of time. If the people haven't been able to collaborate to add simple things like WebFinger into the standards, even with things like #fediforum, maybe we need a reorganization of how things work.
Yes, I agree with that. It's what has my fascination and attention for a while now.
The social dynamics in a grassroots movement are fundamentally different than what exists in corporate settings (with paid employees set to a task to do what's needed).
Such movement needs a different Technology adoption model (TAM), than what you usually have. Like HMSAM [0] or MISC [1] or new variations thereof. And organize accordingly.
Btw, I think a bottom-up standards process might work. One where not just the technology is decentralized, but things emerge organically in a decentralized ecosystem with all kinds of different independent devhubs and projects.
I sketched that to be a possible 3-stage process of: grassroots ecosystem --> FEP process --> W3C SocialCG/WG.
Where increasing formality and rigour is needed going up to open standard level. See:
Add comment