Notices by Steve Bate (steve@social.technoetic.com)

Scenario: Human Q&A site accumulates large quantity of useful information (Quora, StackOverflow, etc.). AI model ingests that data. Users start relying on AI based Q&A. Original Q&A site dies (no traffic). No more training data. What happens then? 🤷♂️

I don't remember. How long has the W3C SocialCG been trying to form a Working Group to at least correct the errors in #ActivityPub? Has been 6 months, a year, longer?

Just for fun... try this #ActivityPub quiz to test your knowledge of the protocol.

https://apquiz.stevebate.dev/

@silverpill @helge I think the issue is more the view that requests are *processed* by actors (versus a server). Do you know of any place in the specification that states that an actor processes the activities arriving in its inbox?

I believe some of the spec authors are confusing the issue by claiming AP is based on Hewitt’s actor model.

@pfefferle Just to confirm, the WordPress /api/v1/instance/peers endpoint is implemented by a different plugin than the AP plugin, right? Which plugin typically implements that? TIA

@silverpill AFAIK, AP doesn’t define any type-specific Update side-effects (actor vs non-actor).

For standard AS2 actor types, neither inbox nor outbox are needed for detection. For inferring AP actor extension types, I think the FEP should be conformant with the AP spec.

@silverpill @naturzukunft @julian @evanprodromou I agree “transient” not clear at all. The
errata you’re referring to is for C2S Partial Updates (Section 6.3.1) rather than anonymous objects, right?

@silverpill @mariusor I believe that removing the Actor type from the AS2 specification was a mistake (especially after looking at the rationale behind it).

That said, I'm not sure that "core type" is a meaningful label.

@silverpill You allow a missing outbox in your FEP although it's required by AP. I think inbox should also be optional. For S2S, if a server has a sharedInbox, the actor-specific inbox endpoint is not very useful. A shared inbox endpoint plus activity/object targeting properties is enough. For duck typing, an "Actor" is a thing that performs an Activity: it's the object (in an RDF sense) of some Activity `actor` relationship.

@silverpill @naturzukunft @julian The AP spec conflates the "anonymous object" and "transient object" concepts. To be precise, an anon object explicitly has a JSON `null` for an id rather than not having an id property at all. (In JSON-LD, it's the same thing, but.... AP should primarily be interpreted as JSON, per @evanprodromou). I've seen very few anonymous objects in the wild (given that definition).

The "Social Web" (#ActivityPub) Foundation's support from Meta and their links to Meta and X are even more dubious after Zuck's policy announcement. I think they should remove the links (and the associated accounts).

'A place of joy': why scientists are joining the rush to Bluesky...

https://www.nature.com/articles/d41586-024-03784-6

@cwebber Thanks for the response. Both the original paper and Wikipedia state: "Everything is an actor". Not in AP. In response to messages, actors can create other actors and only modify their own *internal* state. Not specified in AP. Another difference is that AP actors can communicate to other actors without actor addresses (using "as:Public"). Interestingly, an "inbox" (or message queue) is not required in the Actor Model of Computation (see paper). Too many differences to list here...

@cwebber I'd like to hear more about AP follows the (Hewitt) Actor Model of Computation, if that's the one you mean. Just having message passing and an inbox and a thing called an "Actor" doesn't make the thing a unit of computation. Given the stated importance to AP, I don't see Hewitt's actor model mentioned in the spec or in any of the WG transcripts, so I'm curious what I'm missing.

https://arxiv.org/abs/1008.1459

#ActivityPub

@evan @deadsuperhero The "simple truth" is that no changes have been made to the AP Rec since 2018 despite numerous bug reports and extensive community feedback. There's currently no W3C group with the authority to make any changes. For at least those reasons, I think it's fair to say it's currently unmaintained. I know there has been some talk about chartering a WG to do a minor update. 🤞 FEPs are not W3C, are not AP-specific and are a separate, informal process. But, of course, you know that.

@evan @deadsuperhero One alternative.. talk about how the problems with #ActivityPub will be addressed in a reasonable time frame given it’s not actively maintained and the evolution of it is “closed” (tightly controlled by the W3C). I think this would be more effective for attracting developers than FUD and misinformation about other protocols and attacks on those with more inclusive perspectives. That strategy doesn’t help ActivityPub, in either the short or the long term.

I've been developing an #ActivityPub C2S-based (with extensions) API facade/proxy proof-of-concept for Mastodon. It runs as a separate process that supports proxying the Masto operations but also adds a postable C2S outbox with support for AP C2S activities. These activities are converted into upstream Mastodon API calls. This extended C2S API also supports search, streaming events, managing bookmark collections, and retrieving timeline collections. 1/2

I submitted a draft FEP describing the Mastodon and LitePub/Pleroma #ActivityPub relay protocols. Comments and corrections are welcome.

https://codeberg.org/fediverse/fep/src/branch/main/fep/ae0c/fep-ae0c.md

@silverpill Is there a script for checking a new FEP before submitting a PR? (For content and metadata issues, for example)

@jan Does that list include the software implementation details about the relays? I didn’t see it.