All my toots gone

In case you missed it, as I wasn’t able to share it via Mastodon:


This is terrible news, @mathewlowry. I’m so sorry for you.

I think this is a discussion to have. Automated account backups, or sending backups of accounts to all users before shutting down would be a more acceptable way to remove this frankly awful experience.

1 Like

I’m attempting to solve data portability with FEP-ef61. With server-independent IDs, content can be replicated to backup servers, and if your primary instance suddenly disappears, you simply switch to another one and continue posting as usual.

This is not a small change though, so the process of implementing it across the Fediverse may take many years,


I’m not a developer, so I can’t really interpret the link to better understand the approach’s feasibility, but to me (perhaps understandably) this looks like an incredibly important idea if the fediverse’s is (a) going to take off, and (b) not be dominated by Meta and other deep pocketed corporations.

What needs to happen for this to happen?

1 Like

Currently FEP-ef61 is just an idea. First, someone needs to implement it and learn how good it works in practice. That will likely be me; several other developers are also interested.

If the mechanism works well, this proposal will go through the FEP process and eventually will be finalized. For this to happen, the Data Integrity specification has to become a W3C standard because the proposal heavily depends on it. And I think there should be at least 2 interoperable implementations.

Once proposal is peer-reviewed and finalized, we’ll try to convince everyone in the Fediverse to implement it.

Sorry for your bad experience! There needs to be a better way to shut down a Mastodon instance (probably other fediverse applications too).


And what needs to happen for that to happen? Is funding required? I’m trying to understand how the fediverse advances (or not), given the obvious importance of such a development.

It is already happening :slight_smile: People are working on it, and I expect first results to come in Q1 2024.

I don’t think funding would make a big difference at this stage (though donations to open-source software projects are always welcome).

1 Like