JESUS when even the stochastic parrot throws shade at Msatodon implementation of Activity Pub.

[maikel@vmst.io]

JESUS when even the stochastic parrot throws shade at Msatodon implementation of Activity Pub.

I'm honestly starting to wonder what is the point of people putting years in creating a standard if the main platform using such standard is spitting on it.

#activitypub #mastodon

[benpate@mastodon.social]

@maikel I also wish Mastodon followed the specs closer.

At the risk of sounding like an apologist, real-world constraints and deadlines were probably a factor. Shipping stuff is *hard* and there are always difficult tradeoffs to be made.

C2S API is a good example. They NEEDED a mobile client, but the standard wasn’t enough.

However we got here, it’s on the standards community to build a spec that will work in the real world, then make incentives for the market leader to follow.

[mat@friendica.exon.name]

@benpate @maikel That doesn't sound right. Mastodon came first, ActivityPub came later. Surely that's enough of an explanation? AP was a retro-fit to an existing social media site. Of course it was a bodge job.

To me, that's the true purpose of AP. It works best for connecting existing communities. We should reject the idea of tearing apart and rebuilding what exists today. Instead we should grow the network by meeting people where they are. Talk of "purity" can only be destructive.

[maikel@vmst.io]

@mat talk of purity can only make all other platforms connect better among themselves. Yet every time someone implements Activity Pub the first burden they find is "ooops, this gives an error from Mastodon side".

Standards exist for this very reason.

I agree with you with meeting in the middle, but Msatodon has been up since 2016, Activity Pub since 2018, we're on 2026. Mastodon migrated in late 2017 early 2018 before the spec was even fully ratified. It's been more than 8 years.

If at-the-very-least it didn't have the "huge undocumented behaviours", it wouldn't be stalling all other AP implementers.

@benpate

[julian@activitypub.space]

@maikel@vmst.io not surprising in the slightest considering ChatGPT is trained on public discourse, and most AP development discussions involve how much Mastodon fails at x y and z. Being the big fish means everybody takes pot shots at you.

So it's just a reflection of the training data. Us.

@benpate@mastodon.social you're just feeding the AI machine! 😱

[eyeinthesky@mastodon.social]

@benpate @maikel IMO, the "shipping stuff is hard", while true, isn't the issue. If you had created Emissary 8 years ago and stopped updating it after the first release (even with massive user feedback about UX and technical issues), your project would either be dead or massively forked (if there was enough interest). Even if you had a small release to fix a few serious bugs after eight years, that wouldn't be sufficient. If it weren't for Mastodon, AP would effectively be a dead protocol.

[benpate@mastodon.social]

@eyeinthesky @maikel You're right about that. Add in the lack of tools for building and testing ActivityPub, JSON-LD baggage, and HTTP signatures insanity, and it's amazing that anything gets built at all :)

Fortunately, it looks like some real progress is being made, so hopefully 2026, 2027, and 2028 will bring the community together with better tools and specs, and give devs (even the big one) good reasons to improve compatability.