Email or username:

Password:

Forgot your password?
Robert W. Gehl

Latest #FOSSAcademic blog post: on #ActivityPub, the Non-Standard Standard.

fossacademic.tech/2023/10/15/A

This is based on a presentation I'll be giving at #AoIR2023. It discusses 4 key ways ActivityPub is not a standard standard.

Comments welcome! You can comment on it by (publicly) responding to this Mastodon post.

17 comments
Joel Chan

@rwg really fascinating work! The creation of 7 standards is wonderfully different from the classic xkcd “one ultimate standard” failure mode. The erasure of 1/2 of activitypub reminds me of the lost vision of hypertext and the Semantic Web.

jonny

@joelchan86
@rwg
Im curious what the other standards were?! And also I have wondered what it would take to retrofit the AP client -server API into masto, since it already has scoped API endpoints under /v1 and /v2, could there be a parallel /ap set of endpoints???

Robert W. Gehl

@jonny @joelchan86 They made Webmention, Websub, Linkded Data Notification, Micropub, Activity Stream 2.0 and Activity Vocabulary (these last two are important for ActivityPub).

As for your second question: I have no idea!

jonny

@rwg
@joelchan86
Oh dang I didnt know linked data notifications also came out of this group. I have also wanted to do a "where are they now" for webmention and websub since they seem interesting but I havent seen em used

Robert W. Gehl

@joelchan86 Someone in the SocialWG actually shared that exact comment during one of their meetings 😀

Robert W. Gehl

@joelchan86 As for Semantic Web, a big faction of the SocialWG were very invested in that — the Solid folks. Their influence does appear on ActivityPub in the form of JSON-LD instead of regular JSON.

smallcircles (Humanity Now 🕊)

@rwg @joelchan86

How to deal with #LinkedData for defining good #ActivityPub extensions is something that the dev community struggles with for years and led to endless discussions. Meanwhile in practice most impls follow the "just-JSON" path and throw in some @context just for good measure.

Last weeks these discussions raged on in @fedidevs matrix chatroom.. matrix.to/#/#fediverse-develop

Maybe spec can be reformulated as following a JSON-first approach with optional additional LinkedData-profile.

@rwg @joelchan86

How to deal with #LinkedData for defining good #ActivityPub extensions is something that the dev community struggles with for years and led to endless discussions. Meanwhile in practice most impls follow the "just-JSON" path and throw in some @context just for good measure.

Last weeks these discussions raged on in @fedidevs matrix chatroom.. matrix.to/#/#fediverse-develop

Karl Auerbach

@rwg Have you seen RFC 1055? It is expressly a "nonstandard".

A NONSTANDARD FOR TRANSMISSION OF IP DATAGRAMS OVER SERIAL LINES: SLIP

datatracker.ietf.org/doc/html/

Karl Auerbach

@rwg Back in the days when the Internet was still powered by steam engines operated by gnomes, RFCs were more fun - we were just writing notes to our friends.

This RFC was during the onslaught of ISO/OSI - massive specifications without a whit of explanation. It seemed that something so simple as SLIP just wasn't up to carrying the weight of the word "standard".

Robert W. Gehl

@karlauerbach ooooh, yeah, of course! It was nonstandard in that it wasn’t a Global Standard by a Proper Standards Body!

I recall Andrew L. Russell writing about this.

Jamie Clark

This is a useful and important critique of the #standards process, as applied to #ActivityPub and its precursors. A bit overly bleak, and a bit one-off, as if there are no similar stories like this within established standards. Personally, I take it as a beacon of light that there are successes that •don't• require intervention from large corporate sponsors or celebrity consortium figureheads.
For professional standards moderators like me, this is good discussion and feedback. Expect more.
I'm going to read the whole paper, and think a bit more, before commenting further. HT @rwg @evan

This is a useful and important critique of the #standards process, as applied to #ActivityPub and its precursors. A bit overly bleak, and a bit one-off, as if there are no similar stories like this within established standards. Personally, I take it as a beacon of light that there are successes that •don't• require intervention from large corporate sponsors or celebrity consortium figureheads.
For professional standards moderators like me, this is good discussion and feedback. Expect more.
I'm going...

Robert W. Gehl

@jamiexml @evan

I totally agree that making the standard in the absence of corporate oversight was beneficial.

The problem was a lack of consensus on how to proceed, but because the group just decided to make a lot of standards, they ended up doing some great work.

But it took a toll on people, too.

smallcircles (Humanity Now 🕊)

@rwg @jamiexml @evan

Thank you for this great article. The way that ActivityPub evolved is indeed rather unique and led to emergence of a #grassroots ecosystem.

After standardization it wasn't just Mastodon that introduced #ProtocolDecay. Everyone treading new areas did! Something important is still lacking: Robust way to extend #ActivityPub interoperably.

Providing proper guidance here (and in a 3-stage #StandardsProcess) imho is crucial for #Fediverse future. See:

socialhub.activitypub.rocks/t/

@rwg @jamiexml @evan

Thank you for this great article. The way that ActivityPub evolved is indeed rather unique and led to emergence of a #grassroots ecosystem.

After standardization it wasn't just Mastodon that introduced #ProtocolDecay. Everyone treading new areas did! Something important is still lacking: Robust way to extend #ActivityPub interoperably.

volkris

@rwg this is really interesting, and I hope you’re taking an objective, historical approach to the work.

You know, this is what happened, and it’s an interesting tale of human dynamics, conflicting interests, maybe some game theory, and it can teach lessons for other projects.

I wouldn’t want the work to veer into assuming the outcome was better or worse than counterfactuals. There are serious criticisms of AP, and the link touched on some issues with outcomes.

So rather than praising or criticizing, it strikes me that this is important for the lessons it can teach regardless of how it worked out.

@rwg this is really interesting, and I hope you’re taking an objective, historical approach to the work.

You know, this is what happened, and it’s an interesting tale of human dynamics, conflicting interests, maybe some game theory, and it can teach lessons for other projects.

I wouldn’t want the work to veer into assuming the outcome was better or worse than counterfactuals. There are serious criticisms of AP, and the link touched on some issues with outcomes.

Robert W. Gehl

@volkris Indeed. For the purposes of my book, ActivityPub is a central aspect, and so understanding its history helps me better understand the development of the fediverse.

Go Up