About the same as I pushed the send button on that article, @cwebber was finishing https://octodon.social/@cwebber/102535750059287652

I do believe that datashards would be a superior alternative to lists of ad hoc identifiers for most uses. Datashards assumes contemporary computing resources and applies to the creation of new archives. My blog addresses managing content collections in situ and discussion of data recovery by future archivists is an anticipation of successful adoption of ideas like datashards

@marcus
I've seen this go around a few times...

OP's issue: Append only without curation is bad
Blockchain: Is actually curated
CW: We mostly want structures where information will go away if it's being ignored. That's garbage collection, an important characteristics alongside having a structure where the community can't be forced to delete things

Expand: You don't make data go away in a blockchain. Some immutable and append-only stores stop being replicated when there's nobody interested in replicating them. Participants in a blockchain are forced to replicate the entire chain regardless of its size or contents. No amount of venture capital can remove the revenge porn someone inserted with an xor mask or enable me to access a 3 year old ledger from a device with only 16 gb of storage
@cwebber @liw @emacsen

@freakazoid
We've created a situation where this is a necessity, so a little responsibility and empathy is in order. The principle that "many eyes make bugs shallow" doesn't apply to the threat model that's evolved. People have reasons for making bad decisions. We distance ourselves and fix blame, but we do the same things and we caused this. So that's not fair

Fair is a value served by justice, which means doing the thing in our power to change outcomes
@cwebber

I was already familiar with the actor model from working with Psyc. It was the Go-Fed tutorial from @cj that convinced me I might be able to put an implementation together
@cwebber