Today's question for a resilient #Fediverse is whether various different initiatives are willing to collaborate and cross-pollinate, while keeping their independence.
There's great opportunity to increase the cohesion of the #GrassrootsFedi#ActivityPub developer community and creating strong joins:
@smallcircles@w3c@fedidevs@Codeberg@dansup one thing maybe is that implementing ActivityPub (or adding it to existing projects) is harder than it needs to be. There's a slew of specs (ActivityPub, ActivityStreams etc) that have to be stitched together, Mastodon kind of does its own thing that everyone has to acknowledge, there is a lack of maintained libraries in different languages - probably because of the first issue, it's too poorly defined to write something for.
@smallcircles@w3c@fedidevs@Codeberg@dansup compare to how easy it is to build RSS publishers and parsers, or even Oauth2 support. None of these are "easy" specs but at least I can pull in a library or two to add these things to my project.
You're right. That is what my toot above related to. Since #ActivityPub became a #W3C Recommendation there would've been ample time to iterate on specs, iron wrinkles out, extend, align, etc.
Yet it didn't happen. Instead in grassroots movement observation is that everyone cares for their own initiative, much less for the technology sustrate they rely upon. We have a Fragmentiverse.
Above I call for simple cross-linking of initiatives. Zero interest.
@smallcircles@w3c@fedidevs@Codeberg@dansup exactly. And the thing is, if people are happy with "Fragmentiverse", why bother with ActivityPub etc. at all? Just have a project-specific API for sending messages between instances of your project, on top of e.g. Oauth and JSON. Far less work with more payoff than trying to build on top of this mess.
I think people do care about #Fediverse#interoperability. Just not much about doing the boring chores of bringing everything and everyone together.
There's a lot of innovation funding provided by Horizon Europe via @EC_NGI and @EC_DIGIT that's been hugely valuable. But most of this money is focused on tech-oriented R&I 'funding the fragments', as it were.
Next to no support exists for volunteers glueing & knitting ecosystems / tech substrate together
@danjac@smallcircles@w3c@fedidevs@Codeberg@dansup@EC_NGI@EC_DIGIT It's what I've been saying since many years without being heard :( Every #OpenStandard MUST consist of the Standard, an Open Source reference implementation and an open source validator to test interoperability against. All three MUST be maintained by the Standard Setting Organisation. Without this it isn't an Open Standard.
@danjac@smallcircles@w3c@fedidevs@Codeberg@dansup@EC_NGI@EC_DIGIT The advantages are obvious. The reference implementation proves that the standard can be implemented (we all know how ambiguity in standards limits implementations), it creates a productive feedback loop between implementors and standards authors. The validator radically reduces the cost of interoperability and limits the well-known risks of different interpretations of the standard. Etc.
@smallcircles@jwildeboer@w3c@fedidevs@Codeberg@dansup@EC_NGI@EC_DIGIT in lieu of the big corps, we just have "biggest implementor in the room" i.e. Mastodon driving the standard. But that makes everything slewed towards Mastodon, which itself is built on a hodge-podge of standards with its own project-specific tweaks.
@jwildeboer@smallcircles@w3c@fedidevs@Codeberg@dansup@EC_NGI@EC_DIGIT a reference implementation need not necessarily be a full-blown app like Mastodon: for example, a simple web form I can enter a Webfinger address or inbox endpoint or whatever and see a list of validation errors, what gets parsed, etc.
> If the people responsible for the standard are willing to go into such discussions instead of just demanding that the standard is the only thing that counts, things get better!
Here's a flaw. There's no one in charge. We have a grassroots movement. Now.. #SocialCG might be revived.. there's activity in that direction. But that needs to happen first, and volunteers found to do the chores, put in time and energy.
If Meta or Tumblr or whatever corporation decides that #ActivityPub is worth adding, then likely they can put people behind the steering wheel of the SocialCG with a good salary that makes them do these chores.
That is why grassroots movements are always in underdog position, and R&I investments serves ultimately to enable those corporations.
Exactly. Yet that is the threat when failing to build and maintain a healthy technology substrate. If it doesn't exist the ecosystem can be taken over on a whim by any large corporation.
A megacorpo "standards" aren't a threat to us. They have an incentive to make profit quickly, and that's what they standardize for. We don't have such incentives, we can freely ignore any bad ideas megacorpos have.
The ecosystem can't be taken over on a whim. The best they can do is make a new ecosystem and lure people there. And that's fine, anyone who wishes for popularity and is willing to sacrifice freedom for it, is not someone I care to keep around here anyway.
What we have on the Fediverse right now is people who want to be here, because there's people here they want to talk to. No other walled garden has this, people use those because other people use them. They are riding solely on the argument of popularity. We are not. Let's not try to be like the corpos, that's exactly what everyone who stays here is trying to avoid.
@jwildeboer@smallcircles@danjac@w3c@fedidevs@Codeberg@dansup@EC_NGI@EC_DIGIT This position needs an advocate in Brussels who knows how to a) access and then b) dialogue with policymakers in the European Commission. It might require a new not-for-profit that would exclude large-platform members and financing.
@jwildeboer@danjac@smallcircles@w3c@fedidevs@Codeberg@dansup@EC_NGI@EC_DIGIT Re: "Let's declare Mastodon (or Friendica, Firefish) as the reference implementation for the ActivityPub standard." That could work for a microblogging-specific subset of AP, but #ActivityPub is much more general than that specific kind of application. That's both a blessing and a curse.
I notice that you tooted without any text to it, and it makes me curious. Did you do that to 'subscribe' as it were to this thread so you get notifications of further replies?
> The standard setter is going to become Meta before long anyway
@zayathejew how is that worse than the current standard setter Mastodon, which has an abysmal ActivityPub implementation and punishes every other service that sends them valid but incompatible data?
Just because it's developed in the open, it means little if there's nobody interested in adhering to the standard.
@mariusor Not saying it'll be worse. Ind have no way of knowing that. The most I could do is conjecture that there might be some tension between, on the one hand, #Meta's incentives and how these might influence how it implements #activitypub, and, on the other, what is best for the broader community of #fediverse citizens. But I'd also have to recognise that Meta has the resources and capability to do something really solid and consistent. So who knows? 🤷🏻♂️ I'm never ideological about anything.
If only the EU would ever have funded ActivityPub conformant software, anything would be fine.
But if you give the money just to you “biggest” friend, you are as inhuman as ever and destroying the Protocol (same with services e.g. only accepting "type": "Image").
The conformance section says “The Entirety” quite a few times.
NLnet obviously doesn't know anything about the protocol.
@smallcircles Wait really? I tried importing that post by pasting the link into the searchbar on my instance so I could reply to it over ActivityPub but it didn't seem to work. Am I missing something here?
Add comment