I wrote up some Opinions™ about iterating on the #ActivityPub protocol and how we can continue improving it without sacrificing compatibility or fracturing the network: https://jeremydormitzer.com/blog/activitypub-good-enough-for-jazz/. Comments welcome here or on the post.
@jdormit I think you should join the preparation discussion for AP round table at FOSDEM.
@how How do I do that?
@how So I just register for the forum? When is the round table, and is it okay if I participate if I'm not attending FOSDEM?
@jdormit Yes please, or simply pass me your email in a PM and I invite you. All the answers to your questions are in the welcome message :)
@jdormit your reasoning is quite sound and holds some merrit. it is a good base for discussion. And it solves the question of trow away and restart. It does not break current projects and they probably can easily implement future iterations.
If everyoen agrees, the next discussion point should be what features to include. There are many shortcomings to cover. And the first of all is make AP implementation more uniform. Litepub, pleroma are signs AP has a tendency to splinter
Sure, I agree with that. But I think there's a way to implement security features which preserve the existing network in such a way that the more secure nodes can still communicate with out-of-date ones, e.g. HTTPS is built on top of HTTP but some websites redirect HTTP requests to HTTPS.
So in this case, say you add a capability-based access system to ActivityPub, you could also define some default capabilities that existing nodes are understood to have, and then nodes that implement the capability system can choose the level of access they allow to the default capabilities.
I wouldn't recommend Pterotype yet either, unless you want to be a beta tester on an early-stage product 😅
I'm definitely not trying to undermine your proposal before I've even seen it - looking forward to taking a look once you've published that follow-up post. I just want to make sure that the Fediverse doesn't sabotage its future by fragmenting before it really gets off the ground (and yes, it could also sabotage its future by not taking security seriously enough)
I do see some benefits for that flag day idea. I look at this from a functional standpoint. The dualstack idea is about as good as the backwards compatibility thing. It does not break the current network and it gives projects the time to implement the new AP version. Over time it will move to have AP 2 supersede the previous version
The social network of the future: No ads, no corporate surveillance, ethical design, and decentralization! Own your data with Mastodon!