Convener: Kaliya Young

Participants who chose to record their names here:

• Andy Piper (@andypiper@macaw.social)
• Bob Wyman (@bobwyman@mastodon.social)
• Lisa Dusseault (@lisarue@mastodon.geekery.org)
• Ryan Barrett (@snarfed.org@snarfed.org)
• drimplausible (@drimplausible@mastodon.online)
• charles blass (@lovevolv@mastodon.social)
• Paul Fuxjaeger (@cypherhippie@chaos.social)

Notes

Kaliya’s grant from Summer of Protocols was used to study IETF under the lenses of 3 different pattern languages. Many parts of the IETF and its processe and organization were matched to different patterns in those pattern languages. Often IETF diagrams show different entities “floating in space” rather than under a hierarchy. The RFC process set itself apart from other SDOs from the very beginning, being less formal and hierarchical, and also requiring a larger amount of practical experience before formalizing something.

In 1992 the “protocol wars” of the time inovlved international telcos and OSI. From then til now, IETF rejected formal voting and the participation of state actors. At the time the IAB (Internet Architecture Board) had spun up the ISOC (Internet Socity) to be the legal entity for all this work. But they also stated that they were moving IAB towards OSI endorsement and the membership blew up. In the aftermath of this they agreed that all subgroups in IETF should either be open to anyone, or have a well-documented restricted membership in which voting members are elected or nominated through an open process. Kaliya showed the IETF mission statement and discussed its strengths.

The openness of the IETF has made it strong, at the same period in which the W3C has grown weak because of the lack of openness in how its participants are selected and how payment is involved.

Were fediverse specs to be brought to the IETF to become a WG, it could become one. We saw a great deal of listening. Polling (what used to be humming) is used not to decide finally, but to organize and note agreement going forward.

Anon: It would be pretty disruptive to move all the people already engaged in W3C mailing lists, rooms, processes and so on. Nevertheless we can think about it. IETF has a lot of work that ActivityPub uses, including DIDs, WebFinger. I’d be pretty supportive of IETF processes.

Lisa: I don’t know what should happen. There are STRONG arguments for keeping momentum where it is. But there are also strong arguments for an open forum. As soon as I showed up [in SWICG] and said “hey we’re making a WG right?”, six to eight people already working in the W3C context immediately objected and said “no no, we can’t make a WG because we can’t have open participation.” Those folks push us in both directions.

Bob: IETF has its own standards for things like encryption, and there will be pressure to adopt those, right?

Anon: Is it possible to do such high level stuff in the IETF? Not only ActivityPub but also stuff on top of AP?

Bob: IETF traditionally has focused on interop, If you try to specify how an implementation behaves rather than how it interoperates, you’d be guided towards interop. W3C is willing to say more about how implementations behave.

Lisa: IETF areas are fungible. no technical reason they can’t have higher-level protocols, not just AP but above AP. That said there is a very pragmatic concern where IETF.

Anon: There is not the same big-org control over social activity work that there is over browser tech behind closed doors so I’m not as concerned about that.

Bob: Sometimes it’s not worth bringing work to W3C without a big co behind the work, according to the conventional wisdom. Unusually, the ActivityPub work is happening all through smaller orgs.

Anon: What area would this take place in?

Lisa: Apps area could always be resurrected; it’s an intersection of identity and social graphs but without trying to tackle all things that affect those both.

K: In a way, the W3C spun out of IETF but because TBL showed up at a time of IETF chaos before it settled down again - and then TBL set up the W3C with a BDFL and stayed in that role for decades before figuring out another structure for it.

Anon: Currently the SWICG is operating with open participation. Are there things it could import from IETF?

Kaliya: Possible yes, like the IETF hackathon. To Bob’s point, not that you have to show up and do the work, there’s no way to get stamped by the IETF.

Bumblefudge: there’s not just the normative tracks, there’s also the independent track and experimental track which can be used to get RFCs.

Kaliya and iIsa: not so sure… the independent track may have been throttled.

Anon: there’s still more than one track to RFCness.

Kaliya: those are often hard in their own way and require relationships.

Evan: For people here who participate in the W3C, if we wanted to use the IETF, what would be a current project that would make sense to take to the IETF?

Lisa: LOLA would totally fit as an IETF Internet-draft, on RFC track; we’d want consensus around doing that and to do it in a regular WG, to continue to have the legitimacy of community involvement. And technically speaking, an RFC in the IETF can build upon a W3C TR.

Kaliya: I feel like the FEP energy and activity in the Fediverse is a good fit for IETF processes, where you can write internet-drafts and WGs can adopt those and move fast at times.

Anon: it’s also possible to have feet in both camps.