Welp that answers a lot of why all .ml are down

BarterClub@sh.itjust.works to Technology@lemmy.world – 1470 points –
498

You are viewing a single comment

Why? I don't quite see the relationship between losing access to a domain, and it being uniquely bad for activitypub.

Because it means losing access to the content unless the operators go through the ardous process of moving to another domain, whereas with the matrix protocol the content would remain perfectly available and the only thing that happens when a server has domain issues is that the accounts and specific room alias become unusable.

XMPP has the same issue because it also relies on one central server to host a room, whereas with matrix ALL involved servers replicate the room which means that there is no central server to go down, which is just objectively better for things like chats and forums.

With activitypub all involved servers also replicate the content so I'm not sure what distinction you're trying to make. That's why we can still see all the communities, posts, and comments on the servers that are still online.

that is not entirely true from a lemmy perspective. When an instance subscribes to a community, the remote instance gets the last 20 or so posts, as well as subscribes to all new posts from then on. IT has a local copy of that community. What it doesn't have is any of the embedded media.

AFAIK, this is similar to how matrix works too. I do not know if this is a lemmy implementation choice, or a AP standard?

Edit: haha, i just saw that i am the 3rd person to say the same thing. oops!

Technically instances do actually duplicate the communities. Right now I'm responding to a comment cached on the database of lemmy.thesanewriter.com, and I would retain that ability even if your instance defederated from mine or went down forever.