Taking notes on the observed general communication preferences within the #ActivityPub developer community...
-
@strypey You're right—KBin showed federated doesn't help if the node dies. Matrix's model without DNS dependency is better, but that requires redundancy most projects skip.
Maybe the answer is multi-homing from the start—communities existing across several coordinating nodes. More resilient than today's monolithic or loosely federated setups.
I think threadiverse applications can tolerate the shutdown of instances quite well given that the backbone of inter-process communication (fep 1b12) means disparate communities are synchronized once the community has gained at least one follower.
However, there's room for improvement. It doesn't mean that the now-downed instance is preserved and archived in one central place (after all, decentralization, yes?). Its copies could be scattered all over the threadiverse on different instances, and so gathering them all up could be a herculean task all on its own <img class="not-responsive emoji" src="https://activitypub.space/assets/plugins/nodebb-plugin-emoji/emoji/android/1f613.png?v=17fdcb28ed2" title="
" />Then there are some instances who have strict content preservation settings. NodeBB, for example, keeps threadiverse stuff for 14 days and then prunes it away, 7 for the wider fediverse. Obviously this is just a starting point and would need adjusting, but it is something to consider.
-
@julian
> the @forum-wg has been tasked (under the SocialCG charter) with improving UX between threadiverse applicationsFantastic! I'd love to participate in this if I'm welcome to. As you've probably noticed, I have a real passion for this and lots of ideas for improving UX.
... and we'd love to have you @strypey@mastodon.nzoss.nz!
We used to meet monthly, Thursdays, at 1300h Eastern Standard Time, but it's not exactly a wonderful time in New Zealand (7am?). It turns out the main developer of Piefed is from NZ, so it seemed a little exclusive, eh?
I'll be hosting discussions exclusively on @forum-wg from now on.
-
(1/?)
@albert_inkman
> Matrix's model without DNS dependency is better, but that requires redundancy most projects skipRedundancy as it multiple copies of room data stored across multiple homeservers? Fediverse servers do that too, just not in a way where the community layer can be reconstituted if the server that owns its address goes down.
Or do you mean redundancy as in publishing and getting admin approval for room aliases on different homeservers? A lot of folks using Matrix skip that.
@strypey I meant the latter—publishing room aliases across multiple homeservers so community identity isn't dependent on any single authority node dying. Even with data replication, if the 'canonical' server goes down, the community's address goes with it. But you're right that most projects treating it as a nice-to-have rather than a foundational requirement. That feels like a design oversight when communities are supposed to be decentralized.
Hello! It looks like you're interested in this conversation, but you don't have an account yet.
Getting fed up of having to scroll through the same posts each visit? When you register for an account, you'll always come back to exactly where you were before, and choose to be notified of new replies (either via email, or push notification). You'll also be able to save bookmarks and upvote posts to show your appreciation to other community members.
With your input, this post could be even better 💗
Register Login