The #fediverse-we-have is predominantly #AppCentric, and that rigid perspective is limiting the promise and potential of the #ActivityPub#SocialWeb protocol.

What would it mean to offer #services on the fediverse, that people can discover, obtain and compose into solutions that satisfy their #social networking needs?

By focusing more on the #ServiceOriented message exchange #architecture, the future fediverse can be one of versatile and interoperable Apps & Services.

#SX

@smallcircles Is it safe to say that what you describe here as "app-centric" could also be called "platform-centric"?

Because I also think we have too much stuff that requires signing up for a new account on a new web app first thing, before you even know if you'll like it.

That's why I'm building @encyclia (incl. the stuff beyond its core ORCID function) as a collection of service accounts that you can interact with from wherever you already are. https://encyclia.pub/roadmap

@julian @encyclia

Yes. In general there are a whole range of abstractions that have emerged over time, which the conceptual architecture of ActivityPub does not prescribe. They were 'invented' along the ways, based on particular design choices.

I wrote more about this in a recent SocialHub topic relating to C2S part of the AP specs:

https://socialhub.activitypub.rocks/t/liberating-clients-from-servers-without-throwing-out-baby-with-bathwater/5416/15?u=aschrijver

@smallcircles The mode of interaction I chose (bot accounts that post live content or that can be summoned to perform a task) is the core differentiator between Encyclia and OSN (Bonfire). FAQ: https://encyclia.pub/faq#opensciencenetwork

It limits the range of viable interaction, of course. There's no real way to give people buttons to click, and I don't want to make them learn a text-based command language. So I try to stick to bots with a clear singular purpose, where tagging them can mean "do your thing, bot".

@julian @nicol @steltenpower

My primary concern relates to broad technology adoption. Despite having made a lot of progress on further fleshing out a range of linked data standards, there are still huge challenges to overcome for the LD ecosystem to see their technologies have widespread uptake.

Linked data has its niche areas and audiences where it sees more popular use, but for average devs to "bridge the gap" and go all-in on LD (as all standards are intertwined) poses a giant leap to take.

1+ more replies (not shown)
@nicol @steltenpower @julian @encyclia

The other day @steve @trwnh and I had a very interesting brainstorm discussion - a #SX solution mockup - about a #Protosocial ActivityPub protocol extension, that is both service-oriented and supportive of a full linked data profile.

For anyone interested, the top of this "Groundwork labs" matrix chatroom discussion can be found at:

https://discuss.coding.social/t/protosocial-activitypub-protocol/665