This is a point I keep coming back to again and again.

Hrefna (DHC)

AS2 and AP are focused around the question of "how do I describe the world." This is also what RDF and by extension JSON-LD are focused on. You see this assumption further embedded in how people talk about the "open world assumption" and needing essentially unlimited flexibility, because they are trying to _describe_ things.

But that's not how protocols work. That's not how _any_ protocols work outside of AP.
https://hachyderm.io/@jenniferplusplus/113415270719967178

Hrefna (DHC)

I don't _want_ to have a "rich vocabulary to describe taking a car trip with three friends between Colorado and Arizona for a war outside of Phoenix."

It's not even an advantage to have it unless the parts of that can be _understood_ by a listener. Which in my mind is what a protocol is for: not so that you and I can talk, in English, about our problems (when you may not even speak English), but rather so that we can have our computers talk to each other.

Hrefna (DHC)

If I can't do that—unambiguously—then it does. not. matter. what I can "describe."

So what are the primitive operations of a social network? What I need is, unambiguously, for those to be clear in a way that I can implement a social network safely on top of them.

If I cannot do that then it does. not. matter. that I can describe that I an tentatively accepting a set of concert(/ritual) tickets to Heilung at 3Arena as part of their Albion Eiru ok Erop tour.

Even then: what does that _mean_?

Risotto Voted

@hrefna ...then what are the primitive operations you'd like?

Hrefna (DHC)

A social network is _defined_ in many ways by the _side effects_ of the messages I receive.

If I don't know what the side effects are, unambiguously, it does. not. matter. that I can describe blocking an object because no one understands what to _do_ with that sentence.

OTOH, if I _do_ understand what to do with that information, then I have a solid base to build off of.

Cassandrich

@hrefna I so viscerally *dislike* the ideology behind RDF.

Hrefna (DHC)

@risottobias doesn't matter so much that they exist and there's a way to know what they are.

I can choose to implement them or not, I can choose to accept the tradeoffs or not, but if I don't have them then I don't have a protocol.

Haelwenn /элвэн/ :triskell:

@hrefna Which in effect means that nobody can fully implement ActivityPub, and there isn't even some kind of strictly defined set that everyone would have to implement, everyone kind of just wings it.

Alessandra Sierra

@hrefna I came to a similar conclusion about RDF, after a period of admiring it years ago. If you can’t narrow down the scope of possible ways to express something, you can’t write a program for it

smallcircles (Humanity Now 🕊)

@lanodan

They either wing it, or they start to collaborate with others in their specific domain to define an interoperable model and logic to stick to, possibly to be standardized.

An organic decentralized bottom-up ecosystem emerges. People modeling people's social interactions.

But as @hrefna says, it requires a rock-solid robust and well-documented protocol extension mechanism or things become a mess.

Noah Kennedy

@hrefna honest it's as though folks are saying we don't need http because we have tcp

sure, you can move any data over tcp that you could via http+tcp, but http provides a semantic model which you actually need to implement a web

ap doesn't give you a usable protocol because it lacks a complete semantic model for this use case - you need to either codify something else on top of it or get rid of it and use something else entirely