J12t,
@J12t@social.coop avatar

Very much agreeing with @evan here on the strategy for and related.

https://evanp.me/2023/10/06/activitypub-the-socialcg-and-the-social-web/

markstos,
@markstos@urbanists.social avatar

@J12t @evan “Everything needs to be W3C- and SocialCG-led”

Nope. Sometimes governance fails on effectively govern.

I’m grateful for all the successful projects that exist only because they forked.

evan,
@evan@cosocial.ca avatar

@markstos @J12t you make the mistake of confusing Open Source software and open standards. They are not the same thing.

In particular, a formal standards body is much more important than formal organisations in Open Source.

ActivityPub has been successful because we did the hard work of getting it through a standards process.

If you prefer informal social networking protocols, though, I have good news. There are literally dozens to choose from.

librenews,
@librenews@mastodon.social avatar

@evan @markstos @J12t Like how SOAP won because it was the standard

evan,
@evan@cosocial.ca avatar

@librenews @markstos @J12t there are plenty of counterexamples, but if you'd like we can enumerate all the standards we're using to communicate here.

librenews,
@librenews@mastodon.social avatar

@evan @markstos @J12t Just to clarify my position. I 100% agree forking AP would be the wrong thing. I just don't think that counting competing protocols out is prudent. Hard to see, the future is, and I don't believe in a winner take all scenario. AP is so broad, it is a strength and a weakness.

evan,
@evan@cosocial.ca avatar

@librenews @markstos @J12t

I am trying to remain open to letting people teach me about ActivityPub.

Metcalfe's law says the value of a communications network goes up proportional to N^2.

If you split the network in two equal unconnected halves of size N/2, they each have value of N^2/4, for a total value of N^2/2.

Dividing networks makes them less valuable.

evan,
@evan@cosocial.ca avatar

@librenews @markstos @J12t "winner take all" is a misleading phrase here you should probably avoid.

It's not a monopoly by a single software company.

It's a single, extensible protocol everyone can use and improve on.

We all win when we use standards.

librenews,
@librenews@mastodon.social avatar

@evan @markstos @J12t I agree. It was in response to you saying Nostr and AT were competition that hurts AP. And you are right, that can be the case. But I'm not sure GraphQL (always) hurts REST, though they compete at some level. We may find them living together happily with different use cases. Or not. I can't say. Who really can?

evan,
@evan@cosocial.ca avatar

@librenews @markstos @J12t I can.

Communications protocols are different from API formats.

librenews,
@librenews@mastodon.social avatar

@evan @markstos @J12t Well I think that's ridiculous. I appreciate everything you do/done but I think you have an issue here.

markstos,
@markstos@urbanists.social avatar

@librenews @evan @J12t Right, SOAP, ugh.

Yes, the period where XML dominated web APIs, before JSON appeared and there was mass adoption.

Even the XHTML working group charter was cancelled and a different group, WHATWG took over HTML spec work.

HTML briefly went to formal XML and then loosened again in HTML5.

evan,
@evan@cosocial.ca avatar

@markstos @librenews @J12t HTML is the other classic example.

paninid,
@paninid@mastodon.world avatar

@J12t @evan
Sometimes, I wonder if society could be backward-compatible.

smallcircles,
@smallcircles@social.coop avatar

@paninid @J12t @evan

One question is: What is the meaning of compatibility in a protocol that currently is so tremendously flexible with ill-defined extension mechanism, that each app is its own flavor?

I previously spoke about Standards Process not being afraid of sketching its own path forwards, and that that not be wholly in line with The Mastodon Way™. And I think such approach is good. A quality spec provides incentives for Mastodon to be compliant, and it is not the other way around.

smallcircles,
@smallcircles@social.coop avatar

@paninid @J12t @evan

As for the Standards Process I think it is necessarily a 3-stage one, in the way I explained in the topic:

https://socialhub.activitypub.rocks/t/fep-process-guaranteeing-an-open-and-decentralized-ecosystem/3602

Anything less than those 3 stages will only serve to hamper evolution.

Priority in the Standards Process, besides filling in 'core protocol' gaps & holes is providing rock-solid Integration Guidance with best-practices for creating domain-specific extensions.

evan,
@evan@cosocial.ca avatar

@smallcircles @J12t

Forking the spec is a bad idea. It will squander the goodwill you have built up.

smallcircles,
@smallcircles@social.coop avatar

@evan @J12t

I do not understand. This is the second time you mention forking the spec to me. I am not talking about forking the spec at all though.

smallcircles,
@smallcircles@social.coop avatar

@evan @J12t

Just to make it very clear: I do NOT advocate forking the spec. On the contrary, I would like to streamline the 3-step process from people doing stuff in the ecosystem based on best-practices that give most interop guarantees, to AP extensions / protocol mechanisms that need further SocialHub/FEP standardization, up-to full consideration within the W3C CG/WG context. Final and most formal stage.

This I advocated for years, and still do in same manner. See: https://socialhub.activitypub.rocks/t/fep-process-guaranteeing-an-open-and-decentralized-ecosystem/3602

smallcircles,
@smallcircles@social.coop avatar

@evan @J12t

What I do advocate however is that it shouldn't be the task of SocialHub/SocialCG to reverse-engineer the full tech debt out of the existing ecosystem. That would give a horrible standard anyway.

Bad practices exist now in codebases, and the Standards Process needs to hand guidelines and incentive to correct those, by providing quality specs and docs.

There is a Mastodontoverse / Microbloggoverse that may or may not contain some of those practices, where standards can help them.

evan,
@evan@cosocial.ca avatar

@smallcircles @J12t I guess we have different ideas of what it means!

If you make a separate copy of the AS2 and AP documents, make changes, and publish it elsewhere, I call that a fork.

You call it a "living spec", which is inaccurate, since there is already a published spec.

I like the idea of building extensions through FEPs. I like your idea of publishing profiles (hopefully also through FEPs). We are aligned there.

smallcircles,
@smallcircles@social.coop avatar

@evan @J12t

Ah, there's a misunderstanding then, as this:

> If you make a separate copy of the AS2 and AP documents, make changes, and publish it elsewhere, I call that a fork.

is not what I am calling for. The specs are sole domain of the W3C. No forks, no copies.

FEP's are less formal. They are not AS/AP, but the 3 --> 2 --> 1 Standards Process they can be handed over to W3C for further standardization in whatever form CG/WG thinks is best.

Specs are "living" if this 3-stage rocket works.

sl007,
@sl007@digitalcourage.social avatar

@evan @J12t

Ok, listen @smallcircles :

Please stop lying and your wars.
Many people complaint about you after you banned official people from socialhub, tried to split the community, tried to split the specs.
You were the reason that we are on a mailing list and changed forums.
You answered none of my posts or hamish nor did you answer W3C. You only fought against Social CG and maybe 100 people reported.
After I was elected, you began your war against me. Would you now please stop?
People cry cause of you and I want that to end !!! (!)

evan,
@evan@cosocial.ca avatar

@J12t thank you!

J12t,
@J12t@social.coop avatar

@evan as my grandmother said, what’s true needs to be said. :-)

  • All
  • Subscribed
  • Moderated
  • Favorites
  • fediverse
  • kavyap
  • thenastyranch
  • GTA5RPClips
  • tester
  • InstantRegret
  • DreamBathrooms
  • ngwrru68w68
  • magazineikmin
  • everett
  • Youngstown
  • mdbf
  • slotface
  • rosin
  • cisconetworking
  • megavids
  • khanakhh
  • normalnudes
  • osvaldo12
  • cubers
  • tacticalgear
  • Durango
  • ethstaker
  • modclub
  • anitta
  • provamag3
  • Leos
  • JUstTest
  • lostlight
  • All magazines