that’s fine if you want a “mastodon network”, less so if you want a healthy WWW standard. sure, they’re not “doing anything wrong”, but they are in effect either flaunting standards or setting their own. arguably, part of this is because there isn’t a real “standard” beyond “POST to inbox”. so i think part of the solution has to be creating a framework for conformance profiles. “ActivityPub” as a brand name doesn’t suggest any sort of interoperability if the payloads are divergent or extra requirements are placed on top. another part of the solution should be promoting these standardized profiles as their own separate specs, so that people don’t default to looking only at what Mastodon does. (arguably, this is a harder problem to solve – the “mindshare” of mastodon causes people to not pay attention to anything else.)
All I meant to say is that we can make it more attractive to Mastodon to take more active role in broader standardization efforts, when there’s a slightly more coordinated - yet still retaining its grassroots dynamics - ecosystem working in that direction.
I made some notes in response to Dennis Schubert of Diaspora* critique of AS/AP some time ago. He describes AS/AP et al as constituting more of a “framework”. If we say to anyone “Implement the AS spec and join the Fediverse” we are raising wrong expectations. But we might bring more clarity to the individual components in this framework to ease the path towards reasonably interoperable apps.
Nice! I added your wiki link to the summary wiki. And a TODO item under activitypub.rocks as we should confer with @cwebber whether to get the docs site running from a repo in Fediverse codeberg org.
We can have something similar to Open Source Ideas for starters, as a repo under Fediverse codeberg org. A KISS way of starting. But fedi.foundation would also be perfect for this kind of thing. At Social Coding, long-term, we have an idea for a federated IdeationHub (an app idea in itself).
The activitypub.rocks website may be the place that presents these two separate parallel tracks clearly, and hence serves as a drilldown into all relevant Fediverse-related activity for anyone visiting it.
Update: On the front of the Docs site, I created an issue in @gabek’s fedidocs repo, that would provide a great starting point:
But FEP process is a formal process too. In theory, FEP process and W3C workgroup could complement each other. Two tracks can be imagined as two stages of standardization. Once proposal is finalized and there is a consensus between implementors that it would be beneficial to make it part of the core specification, the W3C workgroup can gather to translate the proposal into a set of changes to the specification document. This way a proposal like FEP-7888 can become a part of ActivityStreams Vocabulary spec, for example.
Unfortunately, this is not going to happen. People who are now trying to restart SWICG have shown zero interest in the FEP process and other things you put in “grassroots” category. I think the Fediverse would be better off without such steering.
I am definitely thinking this way as well. On the mailing list I proposed that the SWICG should have an “official” version of the FEP process, since the idea can theoretically be applied to any git repo preview + PURL service. However, I also noted that failure to take initiative there will probably lead to the FEP process proceeding as-is (and likely reifying further on w3id.org/fep as a basis for an “extension registry”, pending helge’s FEP or my upcoming rewrite of it.)
I noticed the dynamic as well. Cross-pollination between the two tracks should be stimulated as much as possible. My own focus was mostly on ensuring the ‘grassroots parts’ keep going and hopefully growing, so there might be a proper counterweight and balance once corporations and such start to weigh in for real.
I agree. I see the biggest risk that a corporate player launches a FEP-equivalent initiative that blows FEP out of the water. Very well set up - scoped as Grassroots Evolution in diagram above - and with paid people maintaining things. Think similar to React or many other such places (e.g. language level can think of Golang or Typescript).