ActivityStreams enables the open world assumption, just like the full JSON-LD, at the cost of a slight inconvenience of the developers, who now have to namespace their vocabularies. See example 28 of https://www.w3.org/TR/activitystreams-core/#extensibility

Contrary to the RDF model, you can only discuss 1 thing per resource, and you cannot use named graphs for a full dataset. I don’t really know if that is a problem.

#rdf #jsonld #activitystreams

Did you notice? Json-LD is slightly incompatible with AcvitityStreams. The former lets you override term definitions such as "object", while the latter does not. If you parse the same document as json-ld or AS, you will get two different results, if the context is malicious.

#jsonld #ActivityStreams

Un grand merci aux auteurs de la spécification #ActivityStreams, grâce à qui on peut éviter d’implémenter #JsonLD sans sacrifier l’hypothèse du monde ouvert.

@trwnh I disagree.
Mention is well defined (for instance extending the Link object)
#ActivityStreams

Dang, looks like #Mastodon doesn't handle relative URIs when parsing #ActivityPub / #ActivityStreams data.

For example I can't set the `id` of a post to be a relative URL since it loads it as raw JSON.

https://github.com/mastodon/mastodon/blob/main/app/services/activitypub/fetch_remote_status_service.rb#L37

This makes it harder to publish static data to #p2p

"We can't do this feature because the protocol blocks it" is

1. an excuse
2. a set of choices
3. not even correct

Hell, one of my major criticisms of #ActivityPub and #ActivityStreams is how it's almost like they didn't want the protocol to tell anyone "no."

Sometimes it doesn't help you (which means more work), sometimes you'll break backwards compatibility without additional safeguards (that's a choice, and it is a choice not to implement safeguards), but the protocol isn't _stopping_ you.

@atomicpoet @isotopp

Yep :D

The #openEngiadina project switched from #ActivityPub to #XMPP / #RDF.

https://openengiadina.net/en/

From there it implements #ActivityStreams and intends to bridge back to the AP-based #Fediverse again.

And in this space we find #Libervia ..

https://libervia.org

Both projects are doing great research funded by @NGIZero

@helge @fediverseideas

Something related to this is the great work that @pukkamustard has done with #openEngiadina which started with direct #ActivityPub based research projects and currently is on a #XMPP / #RDF stack.

See:

https://openengiadina.net/en/

And #CPub, the last AP iteration:

https://codeberg.org/openEngiadina/cpub

Note that the new XMPP project implements the #ActivityStreams vocabulary.

Longer term I want to remove the need for the HTTP gateway and the AP server from the equation and have users talk directly to each other.

Even longer term it'd be good to remove the need for an inbox server and do something fully p2p.

Then we could focus on pure #ActivityStreams and #RDF data on the client side. (once we figure out indexing to make this stuff actually fast)

@tfardet Of course it happens, that is the whole point of boosting.

It works on all #ActivityStreams platforms, because it is part of the protocol. 'Boosting' something is exactly like reposting it yourself.

@seanpmassey The #ActivityStreams protocol certainly supports both. Here is the specification for Event:

https://www.w3.org/TR/activitystreams-vocabulary/#dfn-event

A group can be implemented as a bot which republishes, to its followers, any post shared with it by one of its followers.

The implementation of these features by software that implements ActivityStreams is a bit lacking so far, but it's there in the architecture.

@atomicpoet We cannot technically prevent Meta from reading #ActivityStreams data, including data from servers which do not directly federate with it. But we can legally prevent them from using that data for advertising, or selling that data or metadata on it to others; and it's quite probable that the EU could be persuaded to do this.

#activitypub #activitystreams are open standards. Loads of code to implement and interoperate is readily available as open source. If they really want to implement #activitypub they can just do it.

@mousebot I started a project called 'Dog and Duck' when I was first trying to figure out how #ActivityStreams worked, and it's what I'll take up again if I do try to hook up Cryogen. But my problem is I start too much stuff and don't finish any of it, and I'm off chasing other hares at present.

If people poke me enough I might get back to it!

Documentation:

https://simon-brooke.github.io/dog-and-duck/

Source code:

https://github.com/simon-brooke/dog-and-duck

Hi there @tastapod 👋

Something that may interest you..

At the #SocialHub developer community that evolves the #Fediverse and the #ActivityStreams / #ActivityPub open standards we are thinking of using #Gherkin and #BDD test suites to formally define the expected behaviour of the protocol and AP vocabulary extensions that various apps use.

Among others this will be part of Fediverse Enhancement Proposals or #FEP's. See:

https://codeberg.org/fediverse/fep

https://socialhub.activitypub.rocks/t/best-practices-for-ap-vocabulary-extensions/3162/12

@misc
@fediverseideas
I think in general it would be super sick to have more typed metadata in the fedi. I don't see anything in #ActivityStreams that would be limiting in that regard, in fact JSON-LD should facilitate it.

I really like the markup syntax in MyST that could be like {loc}`City` but maybe for the fedi more like #loc:City .

would also help for stuff like uspol and etc.

I'm pretty positive #ActivityStreams already has properties for indicating that a Note is part of some specific context, and if not, well that's one of the underrated reasons why building #ActivityPub on #LinkedData is so cool, bc you can just define a new term.

@Inquiry you're soaking in it! 💅🏼

#ActivityStreams is the structured representation of activities that's used by ActivityPub, the protocol that makes Mastodon work.

It's inspired by Activity Theory and many of the structures and properties come from that discipline.

https://www.w3.org/TR/activitystreams-core/

@evan

What are #ActivityStreams?

Does anyone know of any work in #ActivityTheory in #psychology or #HCI that uses #ActivityStreams as a structured modelling language?