> We are limited (to a degree) by some design choices made in ActivityPub
I mention not addressing "Misconceptions" early as most costly mistake of AP ecosystem development. Your sentence above will be interpreted differently by different people based on their Perspective what fediverse means to them.
W3C ActivityPub spec is powerful, based on granular message exchange between addressable actors on the social graph.
"Likes", "Boosts", "Replies" are underspecified, mechanisms are fedi-specific, follow interpretation of AP that has become common on fedi. A big misconception imho is that its unclear what is 'core protocol' and what is extension with solutions on top of it. Like a "Microblogging" application domain, or "Software development" business domain.
A Like is but an ActivityStreams social primitive, a building block. The other 2 are domain-specific. Leaked abstractions, now.
Honestly, I'm not super familiar with Tumlbr. Could you point me to an example of this in action?
We are limited (to a degree) by some design choices made in ActivityPub long ago. "Likes", "Boosts", and "Replies" are all separate things that are hard (but perhaps not impossible) to combine into a single unit.
@benpate@bengo@smallcircles@strypey@EUCommission@nlnet literally just re-implement tumblr, with the exact same reblog chain UX, and the same fusion of likes+reblogs into "notes" (before Matt Mullenweg ruined that, which was the subject of much caustic backlash from the userbase, many of whom have been there for ≥ 15 years) … and see what happens.
and no, Wafrn hasn't done this. I did have a good chat with the Wafrn guy about why the OG tumblr UX is important (its emergent effects) though.
When it comes to Conway's Law then, what do we have today in terms of alternative #SocialNetworking environment, here on #ActivityPub fediverse?
If I squint my eyes so the details become vague, I see more or less a copy/paste of existing #SocialMedia that we are all familiar with, and as #BigTech forces it through our throat. BUT! Decentralized.. a great achievement. We can now build our own roads, instead of being forced to take the highway.
The observation that we "copy/pasted" may or may not be an indicator of the risk that Conway's Law does its work. I leave that as part of my call-for-reflection. Same holds for the risk of corporate capture, who can quickly pave over with asphalt any 'desire path' that became popular, and perhaps make it a toll road.
@strypey just mentions Conway's Law, and how it shapes and affects all that we do. The driving force is #Emergence, which over time also shaped modern global #society as it stands today.
Question for grassroots environments that are able to healthily evolve and naturally grow into long-term sustainable ecosystems - in case of the #fediverse able to support diverse and vibrant online culture, where people cocreate and participate in a value-based collaborative economy - is how #ActivityPub based enabling #technology can be designed to foster the right social dynamics that influence this emergence.
Or else we get US road network, emerged by the lobbying powers of Big Oil. Corporate capture in case of #fedi. Or traffic chaos and road jams, stifling #innovation.
My #SX blog addresses how @EUCommission#funding (via the great @nlnet ) encourages - in traffic terms - creation of infra building blocks. But not road vision, policies, enforcement. Lacks socio-cultural care.
Interesting paper, thanks. I just made an analogy to how road networks evolve over time, and what that means for the #social environment. This against the backdrop of my long blog article about #ActivityPub fediverse #evolution.
Social experience design examines the #SocioCultural ecosystem that emerges by the #technology landscape and is determined by the shape of our #tech that it must grown on. Think like organic moss, that is able to take a foothold in the nooks and crannies of slick aluminium roofs. #SocialWeb is a forest.
An observation is that we generally severely underestimate the impact of "adding an extra online channel, so now we can be social remotely". This way of perceiving social totally misses how everything is different online, and at the same time that many things should / can be very similar to how we do offline #SocialNetworking for ages. Increasing social bandwidth on the wire.
We are in a new editorial cycle for ActivityPub, so if anyone thinks these or other parts of the spec are underspecified, they should be opening issues on the ActivityPub repo right now!
Yeah, the new work on ActivityPub is encouraging. Will this be a version 2.1? Or something bigger like a 3.0?
For me, the biggest opportunities might be to pull some of the FEPs into the spec, for instance, blessing some of the Threadiverse work as the official way to handle groups and discussions.
Is is possible that some FEPs might make their ways into the specs?
FEPs by default don't have the IP statements that make them useful for incorporating into a W3C recommendation. They need to be submitted under the SocialCG CLA. I plan to submit all the ones I've authored.
There are a million Fediverses. ActivityPub is already big and diverse with many different extensions and applications.
Nobody has to push their ActivityPub profiles or extensions to the WG if they don't want.
Anyone can make extensions to ActivityPub under whatever license they want. The onus is on extension creators to make their work compatible on the wire with ActivityPub core.
Nobody who works on ActivityPub core wants to break well-defined extensions on the Fediverse.
@evan from what i understand, some have raised concerns that certain FEPs wont be able to be merged into SocialCG because of the CLA/some creators wanting to be anonymous. How are you planning on handling that?
Will we end up with two fediverses? One that is defined by whatever SocialCG cooks up, and another that is based on all the different FEPs?
- Use fallbacks. - Support multiple profiles -- people are doing great with draft-cavage-12 and RFC 9421, for instance. - Think about how your software will interact with software written by someone who never read your FEP. Will it fail gracefully, fallback to default behaviour, or crash and leak private data? - Be compatible with the spec, and not just compatible with Mastodon.
@valorzard personally, I think a very tight core, with a clear extensibility mechanism and lot of extensions for different applications, is the right model for the Fediverse.
@evan do you think that the portable objects/content signing/PDS stuff can be an extension, or is it so much of a difference it needs to be added to the spec