Discussion · bonfire.cafe

If I was to join fedi development, I'd start with teaching devs and FEP authors that a dictionary isn't a valid representation of HTTP headers 🤷

silverpill

@silverpill@mitra.social · last week

@mattesilver Why not? 🤔

I know that a header may appear more than once, but this is not incompatible with dictionary (the data type).

Zło To

@mattesilver@101010.pl · last week

@silverpill @weekinfediverse

If I was to join fedi development, I'd start with teaching devs and FEP authors that a dictionary isn't a valid representation of HTTP headers 🤷

julian

@julian@activitypub.space · last week

@silverpill@mitra.social I'm sorry you feel that way. I agree that I suppose there hasn't been a lot of true R&D lately, though a lot of that could be we're just shipping with little fanfare.

silverpill

@silverpill@mitra.social · last week

Often fanfare is the only thing that is being shipped

Evan Prodromou

@evan@cosocial.ca · last week

@julian @silverpill Are you guys even on the same Fediverse as I am?

- Dozens of FEPs under development
- 10+ task forces in the CG
- New work like quote posts from Mastodon
- Ghost.org went live in 2025
- E2EE from Emissary and Bonfire this year
- New Working Group to revise ActivityPub

There is a tonne going on. We just had a FOSDEM Social Web Devroom with 20+ presenters. FediMTL.ca is next week. Fediforum is the week after that.

silverpill

@silverpill@mitra.social · last week

Yeah that's pretty much what I meant.

Jupiter Rowland

@jupiter_rowland@hub.netzgemeinde.eu · last week

@ Evan Prodromou

- Dozens of FEPs under development

...and they're usually only implemented by those who maintain them. These and their works, in turn, mostly stay in the shadows while the bigger players in the Fediverse remain blissfully unaware of them.

Pretty much exactly what @silverpill has been doing for a while now.

- New work like quote posts from Mastodon

Technology taken over from GoToSocial.

The principle of how these quote-posts are done taken over from Misskey.

And quote-posts themselves were introduced to the Fediverse as early as May, 2010, by what's known as Friendica today.

This is what @silverpill meant with:

Features that already exist somewhere in Fediverse presented as new inventions.

Mastodon presented their quotes as if they had just introduced an all-new, totally revolutionary feature to the Fediverse. In fact, however, the Fediverse had had quote-posts for a decade and a half at that point.

Let's all face it: The main hindrance in Fediverse development is Mastodon. It's a painfully incomplete implementation of a hopelessly outdated version of the ActivityPub spec. At the same time, its cult-like followers see it as the one and only gold standard and the reference implementation of ActivityPub. They believe that Gargron has invented both ActivityPub and the Fediverse because, frankly, that's what the Mastodon folks want everyone to believe.

If the Fediverse had some fair competition, then Mastodon would have to catch up with software like Misskey or Akkoma or Mitra and eventually powerhouses like (streams) or Forte. But what Mastodon lacks in features, it makes up for with sheer market power. The loudest voices that promote the Fediverse don't actually promote the Fediverse; they only promote Mastodon.

The features of the non-Mastodon Fediverse are unknown and outright unimaginable both on Mastodon and outside the Fediverse. And just about everyone believes that Mastodon is as perfect and fully-featured Fediverse software as it ever comes. Pretty much the only ones who don't are those who daily-drive non-Mastodon Fediverse software.

This actually goes as far as Mastodon users trying their hardest to force e.g. Friendica users to throw away their own culture, switch to Mastodon's culture and abstain from using some 80% of Friendica's features because they aren't covered by Mastodon's culture. At the same time, Mastodon users staunchly refuse to adopt any part of any non-Mastodon Fediverse culture. This means that even if non-Mastodon Fediverse devs introduce new features, they'd better not let Mastodon users know.

And so Mastodon can get away both with a laughable set of features, with completely ignoring not only FEPs, but also large parts of the ActivityPub spec, with advertising features which just about the whole rest of the Fediverse has had for years as completely new original inventions of their own, and with implementing non-standard stuff and forcing the rest of the Fediverse to implement "proprietary", non-standard Mastodon developments.

CC: @ julian

# Long # LongPost # CWLong # CWLongPost # FediMeta # FediverseMeta # CWFediMeta # CWFediverseMeta # MastodonCentricity # MastodonNormativity

🫧 socialcoding..

@smallcircles@social.coop · last week

@jupiter_rowland @silverpill

> The main hindrance in Fediverse development is Mastodon. It's a painfully incomplete implementation of a hopelessly outdated version of the ActivityPub spec.

I have a different take on this. As long as we have an app-centric development model that drives fediverse evolution we'll have these problems. I agree that all the Mastodon quirks that have crept on-the-wire is severely holding back the fedi, but I'd rather lay blame by the ecoystem as a whole who continues to follow a post-facto interoperability leader.

Mastodon is an app. If you take on their #ActivityPub namespace extensions you take on a dependency, put them in charge. But they are an app, a product, a single #FOSS project. And they do product development for their user base first and foremost. There's nothing to criticize on that, and also adopting an open standard doesn't come with obligation to maintain/evolve it. Though it would be smart to do so.

The ecosystem should set its own course.

🫧 socialcoding..

@smallcircles@social.coop · last week

@jupiter_rowland @silverpill

Saying "outdated version of the #ActivityPub spec" is inaccurate. Mastodon isn't a version of the spec, but a app domain-specific implementation on top of an incomplete spec and then subsequently many adopting Mastodon-specific stuff out of convenience. After all they tackled these problems already, so very pragmatic to copy that. But that pragmatism comes at high cost, and here we are today still complaining about it.

The AS/AP ecosystem should be able to stand on its own feet.

What we have achieved with the #FEP is great, but it is still a bandaid. The best we can do in an ecosystem of fiercely independent people intent to go their own way and who shun the design-by-consensus parts (which can be very frustrating indeed) that interoperability requires. It is a way to deal with protocol decay and tech debt after-the-fact, when it has already been created and adopted, part of installed bases and hard to turn towards commonly adopted best-practices.

silverpill

@silverpill@mitra.social · last week

@jupiter_rowland You're right on point, and Mastodon quotes is an excellent example

@julian @evan @weekinfediverse

julian

@julian@activitypub.space · last week

@evan@cosocial.ca maybe there just hasn't been a lot of true R&D coming from me then 😩

Evan Prodromou

@evan@cosocial.ca · last week

@julian you've been doing great work.

I don't think @liaizon meant the original post to be a bummer for the Fediverse, but a wake-up call for engaging developers.

I agree that client-first development in ATProto is extremely exciting. We need to do better on the Fediverse side. Fortunately, our social API is great; we just need to start implementing it and using it.

wakest ⁂ FOSDEM

@liaizon@social.wake.st · last week

@evan @julian
I was just ranting about what I see going on. There are some really promising user facing integrations between separate projects happening on the other side and it pains me to see it cause I know we can also have that here but there are some serious things lacking here and I desperately want to see more movement in exciting directions. The work @silverpill has been doing around portability & DID stuff I have been pointing out to people repeatedly cause its so important.

julian

@julian@activitypub.space · last week

@evan@cosocial.ca I agree, and hopefully I'll have some bandwidth in the coming months to put a PoC together for using the social API!

End of the day I think AP is on more structurally sound software foundations. There's absolutely merit in re-using boring tech (HTTP, JSON, et al.) to make something new and interesting. ATProto makes my spidey sense tingle, in a code smell sort of way. Like Angular did, back in the day.

That's why I'm bullish on AP.

Evan Prodromou

@evan@cosocial.ca · last week

@julian Maybe we should try to implement a forum using just the API.

1 more replies (not shown)

julian

@julian@activitypub.space · last week

@evan@cosocial.ca wait no, I have a better answer.

I must be on one of the other million fediverses @jaz@toot.wales talked about where nothing of note happens!

Evan Prodromou

@evan@cosocial.ca · last week

@julian @jaz lol

silverpill

@silverpill@mitra.social · last week

@ex_06 I think @jupiter_rowland is on Hubzilla, perhaps Mastodon doesn't like it.

>atproto

It's true that funding and centralized architecture makes development easier, but these are not the only things that matter. You also need a culture where people are rewarded for building cool stuff and solving problems.

>Mastodon moves slowly for new big features but everytime I try another software I notice the lack of simple stuff

This is simply because Mastodon has more people working on it. Sooner or later other micro-blogging software will catch up if developers don't burn out.