Convener: Ryan Barrett (@firstname.lastname@example.org)
Participants (add yourself here if you want to be listed, no obligation): Name and Fediverse handle
- Manton Reece (@email@example.com)
- Charles E. Lehner (@firstname.lastname@example.org)
- Marlene Breitenstein (@BreitensteinArt.mastodon.art)
Intro: Fediverse is currently ActivityPub, but there are lots of others! should the fediverse be multi-protocol? If so, why, and how?
manton: very interested! implemented Bluesky API recently, also interested in Nostr multi-protocol starts to get weird though, protocol features and behavior don’t map one to one, mismatches are visible
rabble: Nostr/AP bridge sees these mismatches too, both technical, features, and content. cultures and people are different, context collapse of unexpected/mismatched users and communities is a problem another example: deleting is important in AP and fediverse culture, but not possible in append-only protocols like SSB, Nostr
depends on the protocols and cultures
hamish: already have this problem in AP fediverse, since it’s a big tent protocol with different features on different instances!
aaron: looking at ontologies of data models, how to generate code to map btw them privacy features important!
ryan: three parts: technical protocols, data models, and cultures/communities. different problems
aaron: bridging seems to get around instance-level blocking sometimes, need to resort to blocking bridge users instead, which is awkward
ryan: is bridging the only tool? should we try to get protocols to merge instead?
hamish: we can’t really force them!
manton: can we standardize identities at least?
ryan: most common standard identity across protocols is a URL
hamish: open web! important!
cel: non-web protocols may not have URLs
hamish: Matthias, the WP AP plugin is a bridge!?
matthias: not really a bridge. similarly, DIDs, but they’re huge and heavyweight. also love past bridges! Atom, OStatus, etc not sure if Bluesky will gain traction
charles: multi-protocol clients?
bob: Jon Udell’s Steampipe (https://steampipe.io/) is one! right now we assume clients are “owned” by their servers
aaron: interested in peer to peer
ryan: p2p is great but hard to fight through NATs, firewalls, even with STUN/TURN/ICE, WebRTC notwithstanding
rabble: had lots of approaches to p2p, still difficult. new protocols like Lens, Farcaster use user keys but hub/relay/server-centric eg phones aren’t designed to be p2p nodes. also, if servers don’t do moderation, need another approach. Nostr/Farcaster don’t seem to have one yet
matthias: multi-protocol client would be a step back. want to unify discussion across protocols, not fragment it
hamish: I’m a filmmaker, I want data to be just data, not
charles: multi-protocol client: fragmentation is a problem, acknowledged. then local bridge
bob: been here before! mail protocol wars in ’80s. lots of incompatible systems. used standards processes to specify common denominator, now have just one protocol. paved the cowpaths. start with desired use cases, unify those.
ryan: yes! we have good experience with how to standardize. also thinking about when, ie is it now or too early? fundamental differences still: p2p vs federated, append only vs deletable, users manage keys or servers, etc
hamish: lots of similar history here! open web happened out of the blue, it was great!
matthias: really like how the indieweb is doing this! working code and interop over heavy standards processes. somewhat similar in fediverse/AP too. Mastodon has its dialect, it works, everyone else follows. also WHATWG living spec is great too
bob: would OpenAPI/Swagger help? would make test suites much easier to build. also, lots of people predicted/foresaw the open web. Vannevar Bush, Ted Nelson, etc. I pitched it to TimBL et al fediverse work should be similar, lays important groundwork, take it seriously!
aaron: need AS3! and maybe new AP work too? maybe add DIDs, privacy
There is a need to find a sizable audience who will listen (country-dependent), as well as a way to battle the very organized, concerted effort to undermine climate activists.